SuperUROP All SuperUROP Projects
spacing
SuperUROP
 
Home
Letters of Recommendation
Login

SuperUROP Projects

"IIoT: Intermittent Internet-of-Things"
Faculty Advisor: Michael Carbin
Mentor(s):
Contact e-mail: mcarbin@csail.mit.edu
Research Area(s): Computer Systems
Energy-harvesting devices are small computational devices that mate sensors (e.g., accelerometers, magnetometers, audio, and gyroscopes) with processors to produce small lightweight devices can be deployed in the field and powered solely via energy harvested from the atmosphere (e.g., RF, motion, and solar). Applications for these devices include emerging domains such as astroid mining and ingestible robots. A major challenge with these devices, however, is that each individual device can typically only harvest and store enough energy to execute for several thousands of processor cycles at time. The software we’d like run on this platforms, however, typically requires more than several thousand cycles to execute. In this project, you will build a software platform for distributing small computations across a fabric of energy-harvesting devices. A single computation may therefore migrate from one device to another device until the computation can be successfully completed.
"DeepCode"
Faculty Advisor: Michael Carbin
Mentor(s):
Contact e-mail: mcarbin@csail.mit.edu
Research Area(s): Artificial Intelligence, Computer Systems
Researchers in computer vision and text understanding have recently developed a variety of new techniques that leverage deep learning methods to classify and reason about the content and structure of images and text. However, yet to date, there has been limited exploration into how these techniques can enable automated reasoning about program source code. In this project you will use Tensorflow to build models of program source code towards two goals: 1) deep embeddings — embedding snippets of code into a high-dimensional vector space on which one can perform clustering to identify similar snippets of code in a large repository and 2) compilation — teach a neural how to translate high-level programs (e.g., C) to low-level machine code.
"Resilient networking over unreliable substrate"
Faculty Advisor: vincent chan
Mentor(s): Shane Fink
Contact e-mail: chan@mit.edu,sfink@mit.edu
Research Area(s): Computer Systems
The Internet was designed and matured under a non-adversarial paradigm for networking and transport. While networks today are well-equipped to mitigate benign failures, many open problems exist in addressing adversarial attacks (including so-called Byzantine failures). Some phenomena to be addressed include: modeling failures, fundamental network limits, correlated failures under an attack model and the applicability of learning models. Potential projects include network modeling, algorithm design for adaptive network reaction to attacks, and network simulations. A strong background in 6.041 or 6.042, as well as an interest in network security, is a plus.
"Managing and controling networks under uncertainties and imperfect state knowledge"
Faculty Advisor: Vincent Chan
Mentor(s): Arman Rezaee
Contact e-mail: chan@mit.edu,armanr@mit.edu
Research Area(s): Communications
Managing and controling networks under uncertainties and imperfect state knowledge

The proliferation of applications for Big Data Analytics and the ever-growing interest in interactive video applications have resulted in a drastic increase in the number of dynamic sessions in today’s communication networks. These dynamic sessions can lead to frequent and bursty changes of the network state, and present a new challenge for network management and control (NMC) systems.

The goal of this project is to design, develop and analyze innovative NMC systems that allow for fast and agile reconfiguration of the network resources without the need for collection of excessive network state information. The student is expected to design, analyze and/or simulate networking protocols and conditions to verify that the proposed network architecture and algorithms work well under nominal conditions and can remain stable and efficient under severve congestions or other extreme (e.g. black swan) events.
"Building a Principled Understanding of Training Deep Neural Networks"
Faculty Advisor: Aleksander Madry
Mentor(s):
Contact e-mail: madry@mit.edu
Research Area(s): Artificial Intelligence, Theoretical Computer Science
The project aims to develop a principled understanding of deep learning from a continuous optimization perspective. Specifically, its main focus will be on an experimental study of deep neural network training process. This study will then guide development of theoretical models and, in turn, lead to devising improved training algorithms.

Fair warning: This project is somewhat speculative and might be challenging. It will also require good implementation skills as well as strong algorithmic background and mathematical maturity. Basic knowledge of continuous optimization is a plus.
"Robust Perception (for Robots)"
Faculty Advisor: Russ Tedrake
Mentor(s):
Contact e-mail: russt@mit.edu
Research Area(s): Artificial Intelligence, Control, Signals and Systems
Autonomous cars will soon be driving my kids to school. Robots will be moving through are homes picking things up. And they will be using perception algorithms based on deep learning. Perception algorithms that work really well a lot of the time, but we know they also make mistakes. That's terrifying!

The Robot Locomotion Group at CSAIL is trying to understand the problem of how to formalize concepts of robustness for robots/ feedback systems with complex perception systems in the loop. Understanding where these systems makes mistakes is a good start. Proving that they don't make horrible mistakes is better. Generating fundamentally new perception systems that give stronger robustness guarantees is the home run.

We are looking for SuperUROPs to help with this process, which could include experiments on our robot hardware (up to an including a 400 lb humanoid) to investigating novel optimization/verification algorithms.
Robust Perception (for Robots)
"Robust Manipulation"
Faculty Advisor: Russ Tedrake
Mentor(s):
Contact e-mail: russt@mit.edu
Research Area(s): Artificial Intelligence, Control
Despite the amazing youtube videos, the truth is that most robots still fail to pick up most objects, most of the time. Manipulation in unstructured environments will be one of the key technology break-throughs in robotics over the next few years. The Robot Locomotion Group in CSAIL is attempted to address this problem by understanding the mathematical structure of the decision making problem using combinatorial and continuous (ideally convex) optimization. To make these optimizations tractable, we need to leverage the structure in the equations of motion of the mechanical system.

We are looking for SuperUROPs to help with this process, which could include experiments on our robot hardware (up to an including a 400 lb humanoid), to investigating optimization algorithms, to focusing on potential insights from mechanics/control.
Robust Manipulation
"Open Data for MIT Students"
Faculty Advisor: David Karger
Mentor(s):
Contact e-mail: karger@mit.edu
Research Area(s): Graphics and Human-Computer Interfaces
This project aims at improving the ability of MIT students to access and share data that can help them at MIT. On the academic stide, starting from a project like the MIT course picker (picker.mit.edu), we envision creating broader tools that enable students to share information about their overall academic goals and progress (four-year plans), share experience about courses to take (or not), accurately assess workloads, and coordinate curricula with friends. On the recreational side, we'll consider tools that make it easier to locate and use resource at MIT---such as improvements in event planning and room scheduling. Some of this work will involve coordinating with the MIT registrar and data warehouse to open up access to data that can be of value to students; another part will be to build applications that can collect new kinds of data from students and make good use of it.
"Information Scraps, Quick Notetaking, and Personal Information Organization"
Faculty Advisor: David Karger
Mentor(s):
Contact e-mail: karger@mit.edu
Research Area(s): Graphics and Human-Computer Interfaces
Our lives are filled with small, random scraps of information that seem to have no natural home. Where do we put them, and how do we find them later? We've created List.it (Link a fast, lightweight browser extension for capturing and organizing such scraps. Listit has over 25,000 active users who have recorded more than 100,000 scraps. Analyzing them we've discovered important subpopulations such as packrats, minimalists, and spring cleaners. To advance our study of personal information organization, we want to study the activity of our current users (both the content they create and their interaction with it), add useful functionality to the tool (such as sharing, reminding, and context sensitive retrieval) and study the way users react to the new functionality. Information Scraps, Quick Notetaking, and Personal Information Organization
"Interactive Data Visualization for Everyone the Web"
Faculty Advisor: David Karger
Mentor(s):
Contact e-mail: karger@mit.edu
Research Area(s): Graphics and Human-Computer Interfaces
Exhibit (Link is an open source Javascript library that helps non-programmers author and publish rich interactive data visualizations on the web. We use Exhibit to push the boundaries of web authoring without programming, with our ultimate goal being to enable end-users to WYSIWYG-author complete web applications. Exhibit has been adopted on over a thousand web sites by hobbyists, scientists, merchants, and journalists and has served several million page views. Opportunities to advance Exhibit include (i) incorporating new types of visualizations such as heat maps or network layouts, or entire visualization frameworks such as theJit or dojo GFX into Exhibit, (ii) incorporating Exhibit into common web platforms such as Mediawiki (see Link or Wordpress (see Link (iii) enhancing performance using powerful Javascript libraries such as datavore, and (iv) studying Exhibit's thousands of uses on the web to learn more about how people manage information and what could make Exhibit more useful. Interactive Data Visualization for Everyone the Web
"Transparent Web Browsing"
Faculty Advisor: David Karger
Mentor(s):
Contact e-mail: karger@mit.edu
Research Area(s): Graphics and Human-Computer Interfaces
Nowadays, all sorts of shady companies are collecting information about your browsing activities and using it for their own mysterious purposes. How could that information be used to your benefit? We propose to build Eyebrowse, a web browser extension that gathers information about your web browsing activities and shares that information (under your control) with others to mutual benefit. Potential applications include discovering interesting new web sites based on the browsing activity of your friends, improving web navigation by blazing trails to the important parts of web sites, supporting chance encounters when you and your friends are visiting the same web site, collaboratively browsing the web, identifying links between pages that ought to exist but don't, reporting on global web activity trends, tagging sites and pages according to the interests of people who visit them, and other exciting applications that you will come up with. Experience with Javascript, Django and general web application design is a plus, as is good performance in 6.813/831.
"The Future Textbook"
Faculty Advisor: David Karger
Mentor(s):
Contact e-mail: karger@mit.edu
Research Area(s): Graphics and Human-Computer Interfaces
Now that we can put textbooks on the web, how can we change them to make them better? How can we make them more dynamic, more adaptable to individual students, more sociable, or more informative? We've tackled some of these questions with Nb (Link a tool that lets students hold forum-type discussions in the margins of their online reading material. Nb is currently in use in roughly 25 classes at 6 universities. We have a long list of improvements to implement and assess in Nb, including social moderation, key-question highlighting, organization via tagging, chat and wiki functionality, support for sketching diagrams and other non-text annotations, hot-spot mapping for faculty, html and video annotation, and many others that students like you think of. Experience with Javascript and Django is a plus, as is good performance in 6.813/831. The Future Textbook
"Interactive Data Visualization for Journalists using Wordpress"
Faculty Advisor: David Karger
Mentor(s):
Contact e-mail: karger@mit.edu
Research Area(s): Graphics and Human-Computer Interfaces
There's a new movement in journalism to incorporate rich data visualization in news stories, but many journalists lack that skills to create their own "news apps" for this purpose. We've prototyped a data visualization framework, Datapress (Link to support authoring (not programming) such visualizations in Wordpress, a popular platform for journalism. Your job is to flesh out this prototype. This will involve a cyclic process of contacting and working with journalists to support their use of Datapress, discovering what improvements need to be made to enhance the value of the tool, and implementing those improvements to the Datapress platform. Interactive Data Visualization for Journalists using Wordpress
"From Mockup to Web App: Building the Next-Generation Web Template Language"
Faculty Advisor: David Karger
Mentor(s): Lea Verou
Contact e-mail: karger@mit.edu
Research Area(s): Graphics and Human-Computer Interfaces
Think web frameworks like Node and Backbone are cool? Then help us develop the future of web templates. We are working on a web template language that continues to have benefits long after the page is rendered, including: rich copy-and-paste of data between websites, in-browser WYSIWYG editing, automatically-generated APIs, and site themes that are trivially transportable from site to site. We aim to empower casual web users with the ability to make professional web sites: from just a mockup, we hope to infer the data-backend and editing interface; by just pointing at another site, we hope to import that site's style for reuse on one's own. Experience with Javascript (or Coffeescript) and web development is a plus, as is good performance in 6.813/831. From Mockup to Web App: Building the Next-Generation Web Template Language
"Investigating Privacy Preferences, Expectations and behaviors of voice activated devices"
Faculty Advisor: Ilaria Liccardi
Mentor(s):
Contact e-mail: ilaria@csail.mit.edu
Research Area(s): Graphics and Human-Computer Interfaces
This project will focus on voice-activated devices, both physical (Google Home and Alexa) and mobile-based (Google Now and Siri). A student will be charged with looking into these devices and evaluating their technical ability to capture and retain data, by using traffic analysis and/or inspecting the actual hardware, as well as looking into the existing privacy practices of the companies that deploy these devices. The student will compare and evaluate them.

Most importantly, he or she will grapple with the more fundamental human behavioral questions related to:

- What kind of privacy perceptions and expectations have people considered and what are they concerned about when using these devices?
- What sorts of privacy and transparency controls are desirable for these types of device? In particular, how would people react if these indicators were added?
- How can we ensure that consumers fully understand the privacy trade-offs that come with the use of these technologies?

The right student will have a strong interest in privacy and human-computer interaction research; he or she will possess the technical knowledge or background needed to perform traffic analysis, to develop and deploy web-based probes (preferably Chrome extensions written in JavaScript) and mobile (Android or iOS) apps.
"Mechanistic modeling of exhaled CO2 concentration profiles associated with various respiratory conditions, for application to quantitative capnography"
Faculty Advisor: George Verghese
Mentor(s): Dr. Baruch Krauss, Boston Children's Hospital; Abubakar Abid, PhD student, Stanford EE
Contact e-mail: verghese@mit.edu
Research Area(s): BioEECS, Signals and Systems
Capnography is the measurement of CO2 concentration in exhaled breath, and is a widely used monitoring modality in various clinical settings. Previous work in our group has shown how the diagnostic power of capnography can be greatly extended by a more systematic and quantitative analysis than is typically performed. We have had some success with both machine learning (ML) as well as mechanistic modeling (MM) approaches (the figure depicts the latter). This project will focus on extending and refining the MM approach, in order to enable application to: (a) distinguishing patients with congestive heart failure (CHF) from those with chronic obstructive pulmonary disease (COPD); and (b) monitoring severity and response to treatment in patients with asthma. Ideal background would be 6.011 + 6.022 + a strong interest in clinical application of signals and systems approaches. Mechanistic modeling of exhaled CO2 concentration profiles associated with various respiratory conditions, for application to quantitative capnography
"Safety Verification for Autonomous Driving"
Faculty Advisor: Daniela Rus
Mentor(s):
Contact e-mail: rus@csail.mit.edu
Research Area(s): Artificial Intelligence, Control
Autonomous car research has seen tremendous progress over the last decade and many systems have shown great potential for application in the real world. More recently, more effort has been put into certifying the performance of such systems so that they can be deemed safe in any situation. The common approach to this - also known as the one-trillion mile challenge - is to test against as much data as possible to ensure the system acts safely at all times. However, this approach suffers from the exponentially large quantities of data that are required to cover all driving scenarios that could potentially arise, making it computationally intractable.

In this project, we have developed an alternative approach that allows to verify the safety of autonomous systems in a wide variety of road situations by relying on tools and methods drawn from the area of formal verification and reachability analysis. The underlying foundation is to capture a large variety of behaviors in one single set, which is then used to verify an entire set of driving scenarios, instead of single instances. This way, we avoid the scalability issue from the common approach.

Current limitation of the project, however, include a rather limited choice of road networks to choose from. The goal of this particular aspect of the project is to extend the variety of road networks that can be verified. A larger variety would imply that we are able to verify the safety on networks as large as Cambridge. To cope with the computational complexity that arises from the problem computations should be carried out on a cluster.
"Road Understanding with Deep Learning"
Faculty Advisor: Daniela Rus
Mentor(s):
Contact e-mail: rus@csail.mit.edu
Research Area(s): Artificial Intelligence, Control
We are investigating the generation of probabilistic path envelopes with the goal of autonomous driving in complex and cluttered urban environments without the use of a prior map. Furthermore, we will be automatically labeling data containing proposed paths and obstacles without requiring manual annotation from recorded routes driven by our data collection vehicle. Therefore, we are able to train a deep segmentation network from vast amounts of labeled data.

Current autonomous driving approaches heavily rely on highly detailed prior maps. Unfortunately, scaling the creation and maintenance of maps spanning vast regions remains an unsolved problem. Mapping solutions so far are unsuccessful in reflecting changes of the road network such as road works in an appropriate time frame. Our method solely relies on direct sensory inputs such as vision and LIDAR without the use of prior maps.

Machine learning methods capable of effectively labeling image regions as drivable space in a deterministic fashion exist (eg. SegNet).
Nonetheless, neglecting the estimation of uncertainty of learned models is reducing robustness and causing potentially hazardous behavior in the case of false or inaccurate detections. We intend to closely couple perception, and motion planning and decision making by estimating and appropriately propagating uncertainties.

Current work packages include the implementation of a automatic labeling pipeline for vision (OpenCV) and LIDAR (PCL), implementation of the deep-learning framework including custom layers (TensorFlow) and testing.

We welcome applications from motivated and talented students for this project and are looking forward to hearing from you.
"Development of Photonic Nanocavities for Extreme Light-Matter Interaction"
Faculty Advisor: Dirk R Englund
Mentor(s): Hyeongrak Choi (Chuck)
Contact e-mail: choihr@mit.edu
Research Area(s): Applied Physics, Numerical Methods, Signals and Systems, Theoretical Computer Science
The development of dielectric nanocavities has enabled a range of major breakthroughs in cavity quantum electrodynamics over the past decade, including the demonstration of strong coupling of single quantum emitters with electromagnetic modes, single-emitter lasing, and efficient classical and non-classical light sources.
A common assumption has been that the mode volume of dielectric cavities is ultimately governed by the diffraction limit. To achieve nanocavities with higher light-matter interaction, the common method is to design higher quality factors (Qs) to boost Q/Veff where Veff is the cavity mode volume. Although various cavity designs have been published, that theoretically predict Q factors of billions, practical limitations such as fabrication imperfections and material losses limit Qs to the order of a few million. In practice, high quality factors are also often problematic as they introduce instabilities, e.g., in temperature.
We reconsider whether the mode volume is, in fact, limited. Surprisingly, we were able to show that it is possible to strongly concentrate the electromagnetic field inside the cavity, which allows dramatic reduction of the mode volume. In particular, we showed that two design mechanisms that allow us to recursively reduce Veff to the point that it is limited only by fabrication. Remarkably, our semi-analytical design reveals the origin of recently reported blind computational mode-minimization results. Because we now understand the essential features that reduce mode volume, we can design for them directly, while also retaining high quality factors. Indeed, our approach allows orders of magnitude higher Q factor and smaller mode volume than the best blind optimization approaches, despite being much faster, accounting for the full 3D problem, and making no unrealistic assumptions about continuously varying dielectric materials.
The extreme field concentration of these nanocavities opens up surprising new possibilities in a range of applications. In the paper, we focus on the “extreme” area of quantum nonlinear optics at the single photon level - a topic of intense research today (Chang et al. 2014). An open question in this field is whether nonlinear optics at the single photon level is possible without the use of an atomic medium or atom-like emitters, which are often difficult to fabricate and control. Using the nonlinearity of dielectric materials would have many practical advantages, as it could be incorporated monolithically into photonic circuits even at room temperature.
Your roles: The SuperUROP will work in a team of graduate students with professor Dirk Englund, on designing the cavity for beating the conventional limits. This will include the development of step-by-step design process, as well as study on simulational optimization methods. You will learn about all aspects of the process flow, including numerical device simulations, device fabrication, and device characterization in the laboratory.

Qualifications: We are looking for a student with a background in coding (matlab) and a strong motivation to learn about quantum technologies, fabrication, optics, etc. The student will need to participate in weekly meetings at least twice per week, which include updates from all team members. This is a visionary project and we are looking in particular for students who will be self-motivated, who have leadership abilities, and who can work well in a hard-charging team.
Development of Photonic Nanocavities for Extreme Light-Matter Interaction
"High-fidelity microwave electronics and machine learning for high-fidelity quantum gates on semiconductor spins"
Faculty Advisor: Dirk R Englund
Mentor(s): Chris Foy, Donggyu Kim
Contact e-mail: cfoy3@mit.edu
Research Area(s): Applied Physics, Artificial Intelligence, Circuits, Control, Nanotechnology, Signals and Systems, Theoretical Computer Science
The last decade’s progress in manipulating quantum systems has lead to steadily increasing interest in developing a quantum computer. One possible qubit system is the nitrogen-vacancy (NV) center in diamond. NVs have spin dependent fluorescence, second-scale electron spin coherence times, and can be manipulated at room temperature, making them ideal candidates for quantum information applications. Key to using the NV center as a qubit is the development of scalable modular components which can be assembled into a quantum computer. In this project, we will design the modular component for microwave control of the electron and nuclear spins of the NV center. We require that this component allows for individual control of the NV’s spin state and low crosstalk between adjacent components.

Your role: You will design the microwave components so that it satisfies the design constraints given by a modern nanofabrication process. Combining with electromagnetic characterization of the microwave components and its action on quantum two-level spin qubit, you will develop and optimize the control microwave pulses to suppress coherent error of the spin qubits in general. You will also use a spin microscope setup in the lab to characterize these components.

Another interesting question we will investigate in this project concerns the development of quantum control on multiple spins. Quantum control techniques generally are built from first principles to protect against certain types of errors and decoherence. However, here we will try to use machine learning methods trained on an experimental system to develop optimal control techniques. The artificial neural network will be able to detect and compensate against errors and decoherence, without the need to first develop first-principles models to describe the noise and errors. This approach could be especially advantageous as one scales to many qubits with possibly difficult underlying noise models.

Qualifications: We are looking for an undergraduate who has a background in electricity and magnetism and, ideally, some familiarity with quantum mechanics, especially spin-½ systems. The student should be highly motivated and welcome challenges in scientific computing and laboratory experiments. Previous lab experience will be useful in transitioning from initial simulations to experimental work. Above all, we are looking for a student who is proactive, ready to learn, and excited about designing devices and quantum systems.
High-fidelity microwave electronics and machine learning for high-fidelity quantum gates on semiconductor spins
"Linguistic Analysis of Wikipedia for Question Answering"
Faculty Advisor: Boris Katz
Mentor(s): Sue Felshin
Contact e-mail: boris@csail.mit.edu
Research Area(s): Artificial Intelligence
We study how natural language processing and information access methods can improve human-computer interaction. Wikipedia, the world’s largest crowdsourced encyclopedia, is a great source of knowledge for question answering systems. Using linguistic analysis of information in Wikipedia, we hope to considerably increase the precision and coverage of our START Natural Language Question Answering System.

Good candidates are passionate about well-designed code and have strong experience with Python, SQL, large codebases, and Test Driven Development; exposure to linguistics is a plus.
"Developing a QubitFoundry"
Faculty Advisor: Dirk R Englund
Mentor(s): Michael Walsh, Noel Wan
Contact e-mail: mpwalsh@mit.edu
Research Area(s): Applied Physics, Artificial Intelligence, Computer Systems, Graphics and Human-Computer Interfaces, Nanotechnology
Project description: The past decade has seen tremendous advances in quantum technologies. Quantum computing is now becoming a reality, pushed by governments, venture capital, and established companies including Google, IBM, and Microsoft1. Quantum measurement tools are now, in many fields, the most precise instruments available, and are enabling new technologies, such as magnetic resonance on single atoms, that were impossible just a few years ago. Quantum communications allows for unconditionally secure communications, and governments and countries are setting up large-scale terrestrial and satellite-based quantum communications systems. A recent report summarizes the advances and future prospects of quantum technologies: Link

Underlying all of these quantum technologies are quantum memories, usually in the form of quantum bits, or `qubits.’ A particularly promising type of qubit consists of atom-like color centers in diamond2,3, which incorporate electron and nuclear spin qubits coupled to photonic qubits. A major bottleneck in advancing quantum technologies based on such solid-state spin systems is the fabrication and processing of high-performance qubit devices. The Quantum Photonics group at MIT has developed new processing methods that now produce world-leading solid-state qubit devices. Because these tools now enable the production of thousands of qubit devices in one production run, which is far more than any one team requires, there is now the opportunity to open up these process runs to other users. Much like multi-project-wafer runs have revolutionized the way that semiconductor electronics research was done in the 1980s, this “qubit foundry” process has the potential to dramatically accelerate research and development in semiconductor quantum technologies.

Your roles: The SuperUROP will work in a team of graduate students and postdocs, with professor Dirk Englund, on establishing a framework for the world’s first “QubitFoundry.” This will include the development of a process design kit (PDK) consisting of essential qubit devices, as well as a web-based interface for multi-project-wafer device design and process flow. You will learn about all aspects of the process flow, including numerical device simulations, device fabrication, and device characterization in the laboratory using robotic microscopes.

Qualifications: We are looking for a student with a background in coding (python, matlab) and a strong motivation to learn about quantum technologies, fabrication, optics, etc. The student will need to participate in weekly meetings at least twice per week, which include updates from all team members. This is a visionary project and we are looking in particular for students who will be self-motivated, who have leadership abilities, and who can work well in a hard-charging team.

References:
1. Mohseni, M. et al. Commercialize quantum technologies in five years. Nature 543, 171–174 (2017).
2. Childress, L., Walsworth, R. & Lukin, M. Atom-like crystal defects: From quantum computers to biological sensors. Phys. Today 67, 38–43 (2014).
3. Aharonovich, I., Englund, D. & Toth, M. Solid-state single-photon emitters. Nat. Photonics 10, 631–641 (2016).
Developing a QubitFoundry
"Distortion-minimizing bijective surface correspondence"
Faculty Advisor: Justin Solomon
Mentor(s):
Contact e-mail: jsolomon@mit.edu
Research Area(s): Graphics and Human-Computer Interfaces, Numerical Methods, Theoretical Computer Science
The problem of *correspondence*, or mapping from one surface into another, finds application in medical imaging, computer graphics, CAD, and other disciplines.

State-of-the-art bijective correspondence involves mapping two surfaces into a common domain like the plane and then composing these maps. This guarantees that every point gets mapped, but it does *not* minimize distortion (stretch/shear) of one surface mapped onto the other. In this project, you will develop tools for distortion-minimizing correspondence using these methods as a starting point.
Distortion-minimizing bijective surface correspondence
"Regularized vector fields and optimal transport"
Faculty Advisor: Justin Solomon
Mentor(s):
Contact e-mail: jsolomon@mit.edu
Research Area(s): Graphics and Human-Computer Interfaces, Numerical Methods, Theoretical Computer Science
"Optimal transport" is a popular mathematical tool for matching between geometric features and probability distributions, with applications in computer vision, graphics, and other disciplines. Recently, new algorithms have been formulated for optimal transport that make use of vector fields, popular for triangle meshes in the computer graphics world.

In this mathematically-oriented project, you will implement a state-of-the-art algorithm for optimal transport and formulate/implement/analyze an extension of it designed to improve stability and efficiency.

This project is math-heavy, and students with background in numerical methods or geometry will be preferred.
Regularized vector fields and optimal transport
"Kernel Functional Maps"
Faculty Advisor: Justin Solomon
Mentor(s):
Contact e-mail: jsolomon@mit.edu
Research Area(s): Artificial Intelligence, Graphics and Human-Computer Interfaces, Numerical Methods
"Functional maps," introduced in 2012 by Ovsjanikov et al. to the computer graphics community, are popular tools for expressing correspondences between shapes. These make use of linear algebra to represent a map from one surface into another.

In this project, you will formulate and implement a *kernelized* version of functional maps, using ideas from kernel methods in machine learning. The hope is to improve the quality and reliability of correspondence algorithms.

This project will make use of numerical linear algebra and geometry, so students with background in these areas are preferred.
Kernel Functional Maps
"THz Image Sensors with Electronic Beam Forming"
Faculty Advisor: Ruonan Han
Mentor(s):
Contact e-mail: ruonan@mit.edu
Research Area(s): Circuits, Signals and Systems
Terahertz wave is a safe (compared to X-ray) and high-resolution (compared to microwave) tool for non-invasive imaging in security and industrial quality control. We have developed focal-plane detector arrays for THz imaging. However, due to the large wavelength (~0.5mm) that determines the on-chip antenna size, it is impossible to integrate a large number of THz pixels on a single chip. Because of this, THz imagers often use mechanical scanner which is bulky and slow. In the next generation of THz image sensors, the pixels detect not only the intensity of the incident THz wave, but also its phase. It is then possible to achieve electronic scanning inside a 2D space; hence the mechanical scanner is eliminated. The SuperUROP in this project will have the opportunity to (1) build the test board for beam steering and data sampling, (2) develop signal processing algorithms that reconstruct the images from the raw data of our chips. THz Image Sensors with Electronic Beam Forming
"Highly-Stable On-Chip Clock: Characterization and Electronic Design"
Faculty Advisor: Ruonan Han
Mentor(s):
Contact e-mail: ruonan@mit.edu
Research Area(s): Applied Physics, Circuits, Materials and Devices, Signals and Systems
Polar gas molecule has very narrow and stable transition lines in the sub-THz range. By probing the exact frequencies of these lines using a CMOS silicon chip, we can build a very small “atomic clock” that provides highly stable output frequency (<10-10 error). Students in this project will be in charge of (1) testing the stability of the clock prototype and its sensitivity to temperature change, and (2) build a microcontroller circuit that senses the environment temperature and adjusts the clock’s output frequency accordingly to eliminate its temperature dependency (10-11~10-12/degree). The student also has the chance to participate in the on-chip electronic design of this clock.
Prerequisite: analog and digital circuits (6.002).
Highly-Stable On-Chip Clock: Characterization and Electronic Design
"Democratizing Deep Learning"
Faculty Advisor: Lalana Kagal
Mentor(s):
Contact e-mail: lkagal@csail.mit.edu
Research Area(s): Computer Systems
The application of machine learning to vision recognition problems usually requires (i) the collection of pictures for a training set, (ii) the labeling of these pictures, (iii) a machine learning model to be training on this data, and (iv) the model to be deployed in an accessible manner. In order to ensure accuracy, this entire process of model development is iterative, and thus time-consuming. It requires both machine learning expertise and domain expertise. We are developing a framework, which automates the model development life cycle from data collection to final deployment. Utilizing the pervasiveness of mobile devices, this framework will be deployed as an app, allowing users to participate in the model development cycle via their phones. It will enable non-experts both to crowdsource data collection and to harness the labeling expertise of domain experts, and it will not require in-depth knowledge of machine learning.

For this project, we are looking to integrate our deep learning framework with App Inventor (Link that provides a block-based interface through which individuals can easily create mobile apps. The goal is to democratize deep learning in order to enable non-technical end users (including children) to develop mobile apps for all kinds of domains, from telling apart butterfly species to identifying sneaker brands.

This project is being carried out by PIs Kalyan Veeramachaneni and Lalana Kagal.

Ideal candidates will have taken a machine learning class and will have some knowledge of Web and mobile development.
Democratizing Deep Learning
"THz Spectroscopy for Molecule Sensing and Identification"
Faculty Advisor: Ruonan Han
Mentor(s):
Contact e-mail: ruonan@mit.edu
Research Area(s): Applied Physics, BioEECS, Circuits, Materials and Devices, Nanotechnology, Signals and Systems
A wide range of gas molecules have rotational states that can be excited by terahertz (100GHz~10THz) waves. Based on this principle, spectroscopy using continuous-wave, phase-locked THz signal is able to provide very high sensitivity and specificity. Potential applications of this technology include airborne-agent identification, medical diagnosis (e.g. breath analysis for blood glucose level monitoring), and fundamental applied science. The Terahertz Integrated Electronics Group has developed a fully-integrated THz spectrometer on CMOS chip. The research for the SuperUROP student in this project includes (but not limited to): (1) the design of high-resolution frequency synthesizer at the PC board and chip levels. (2) Gas sensing experiments and breath analysis demonstration, and (3) data-processing algorithms and software that help resolving the absorption spectrum when significant overlaps between the “fingerprints” of different molecules exist due to spectral broadening. This project is perfect for the students who wish to apply frontier electronic technologies into the study of chemistry, biology and applied physics. THz Spectroscopy for Molecule Sensing and Identification
"GaN probes for opto-genetics"
Faculty Advisor: Tomas Palacios
Mentor(s):
Contact e-mail: tpalacios@mit.edu
Research Area(s): BioEECS, Materials and Devices, Nanotechnology
Optogenetics allows us to study the brain by firing specific neurons though exposure to blue light. In this project, we will design a new optogenetic probe made of GaN that integrated micro-LEDs, as well as chemical sensors to detect the concentration of different neuro-transmitters. This will allow unprecedented resolution in the study of the chemical operation of the brain. Micro-fabrication experience or 6.152J is useful but not required.
"Wearable biochemical sensors based on graphene"
Faculty Advisor: Tomas Palacios
Mentor(s):
Contact e-mail: tpalacios@mit.edu
Research Area(s): Materials and Devices, Nanotechnology
This project focuses on the development of a wearable sensor for glucose and hydration monitoring. In collaboration with graduate students, the Super-UROP student will develop and test the control electronics for the wearable device and interface it with graphene-based sensors of different kinds. The student will also lead the system-level integration and testing of the entire wearable system. A machine learning algorithm will also need to be developed to extract sensor information in a noisy environment.
"New approaches for thermal management in GaN electronics"
Faculty Advisor: Tomas Palacios
Mentor(s): Tomas Palacios
Contact e-mail: tpalacios@mit.edu
Research Area(s): Materials and Devices
Power dissipation significantly limits the performance of most electronic systems. In this project, we will be experimentally studying new technologies to improve the thermal management in GaN-based electronics. Specifically, the GaN wafers will be bond to diamond, AlN and SiC wafers to help reduce the operating temperature of the GaN electronics. The work will mainly be conducted in the cleanroom facility at the Microsystems Technology Laboratories. Micro-fabrication experience or 6.152J is useful but not required.
"Between-ride routing algorithms for improving the efficiency of private transportation services"
Faculty Advisor: Mardavij Roozbehani and Munther Dahleh
Mentor(s): Mardavij Roozbehani and Ian Schneider
Contact e-mail: mardavij@mit.edu
Research Area(s): Control, Numerical Methods, Signals and Systems, Theoretical Computer Science
The market for mobile-based ride services is growing rapidly, especially in urban areas. Mobile applications like Lyft and Uber represent substantial value to consumers, but they also have problematic side effects, including potential increases in urban traffic and increased greenhouse gas emissions.
When a driver drops off a passenger, it is difficult for them to decide what to do next. For instance, even if a driver is aware of peak prices in a certain part of the city, they have a difficult time deciding if it is worth the fuel and time costs to drive to that part of the city in order to increase the expected value of their next ride. This calculation can be much more effectively performed by software.
More effective between-ride routing algorithms could represent substantial benefits to drivers of taxis and application-based ride services. They also could help reduce urban traffic and greenhouse gas emissions, by making existing transportation services more efficient.
Our group has developed a simple between-ride routing algorithm as the solution to a dynamic programming (DP) algorithm. We hope to find as student who is excited to develop and implement the algorithm in order to test its properties, work on effective ways to reduce time-complexity, and explore the interactions of different drivers in a competitive environment.
Between-ride routing algorithms for improving the efficiency of private transportation services
"Competitive pricing to maximize efficiency in electricity demand response programs"
Faculty Advisor: Mardavij Roozbehani
Mentor(s): Mardavij Roozbehani and Ian Schneider
Contact e-mail: mardavij@mit.edu
Research Area(s): Control, Energy, Numerical Methods, Signals and Systems
The variability of renewable resources poses new challenges for electricity systems. If unmanaged, the growth of renewable resources will strain existing electricity generation and increase the likelihood of power failures. Innovative methods must be developed to improve the flexibility of the power grid; otherwise, its current capabilities will limit the deployment of renewable resources.
Demand response—the ability to modulate consumer electricity demand—is a promising opportunity for providing flexibility to the electricity grid. Demand response aggregators contract with homes, businesses, and factories by installing control devices on their electric loads and varying consumption based on market price signals and individual consumer requirements.
For a demand response aggregator to stay in business, it needs to successfully estimate the flexibility of the loads under its control and to intelligently offer demand response services at prices that maximize its revenues. We are looking for an undergraduate who is excited about optimization, game theory, and electricity systems to help develop bidding procedures to maximize demand response revenues and to implement the resulting algorithms using Python or Julia.
Competitive pricing to maximize efficiency in electricity demand response programs
"Long-Distance Quantum Communication with Diamond Spin Qubits"
Faculty Advisor: Dirk Englund
Mentor(s): Eric Bersin
Contact e-mail: ebersin@mit.edu
Research Area(s): Applied Physics, Control, Materials and Devices, Nanotechnology, Numerical Methods
The exceptional optical and spin properties of nitrogen vacancy (NV) centers in diamond [1] have led to the demonstration of a wide range of quantum technologies including quantum entanglement [2,3], quantum teleportation [4], and precision sensing [5-7]. Key to developing large-scale quantum communication and computation networks is the ability to transport flying qubits efficiently. However, current applications have been limited by high absorption losses of the near-infrared emission in optical fiber [8]. While red frequency down-conversion to the telecom band has been achieved, demonstrations at NV wavelengths have thus far been limited to 44% efficiency due to fluorescence noise from the above-band pump lasers used to perform the conversion [9]. We are developing a system for converting the 637 nm NV emission to the near-telecom band, where we can use below-band pumping to mitigate this background. Doing so will enable us to use a recently established fiber link between MIT Building 36 and MIT Lincoln Lab to entangle NV centers over a distance of 20 km. Once established, this link will be key in performing long-distance quantum communication, quantum key distribution, distributed computation, and developing quantum repeater technologies.

The student will be expected to focus on the following tasks:
Build and calibrate the frequency conversion system using periodically poled nonlinear crystals and 637 nm lasers
Learn how to coherently manipulate and readout an NV spin qubit
Use the system to coherently convert NV emissions into the near-telecom band for long-distance entanglement protocols

Prerequisites: Highly motivated students should have a course in quantum mechanics and be capable of programming in MATLAB or Python.

[1] Doherty, M. W. et al, Phys. Rep., 528, 1–45 (2013)
[2] F. Dolde et al, Nature Phys., 8, 1–5 (2013)
[3] H. Bernien et al, Nature, 497, 86–90 (2013)
[4] W. Pfaff et al, Science, 345, 532–535 (2014)
[5] J. Taylor et al, Nature Phys., 4, 810–816 (2008)
[6] F. Dolde et al, Nature Phys., 7, 459–463 (2011)
[7] G. Kucsko et al, Nature, 500, 54–58 (2013)
[8] H. Bernien et al, Nature, 526, 682-686 (2015)
[9] R. Ikuta et al, Optics Express, 9, 11205-11214 (2014)
Long-Distance Quantum Communication with Diamond Spin Qubits
"Developing High-Speed FPGA Architecture for Quantum Cryptography"
Faculty Advisor: Dirk Englund
Mentor(s): Darius Bunandar
Contact e-mail: dariusb@mit.edu
Research Area(s): Applied Physics, Circuits, Communications, Control, Nanotechnology, Numerical Methods, Signals and Systems
Project description: Quantum cryptography is the art of encrypting and transmitting information with absolute security that is guaranteed by quantum mechanics, rather than the difficulty of certain mathematical problems. We have developed nanophotonic chips that are able to perform quantum cryptography using quantum light. In this particular project, we are attempting to integrate the electronics and the photonics aspects of nanoscale quantum cryptographic systems to make them practical for everyday use.

Your roles: You will develop an FPGA signal generator capable of generating high-speed analog signals for driving quantum cryptographic systems. Eventually, you will deploy the systems and perform quantum cryptography experiments between MIT and MIT Lincoln Laboratory.

Qualifications: We are looking for a student who has a strong background in FPGA development as well as digital and analog signal processing. Knowledge of circuit design and/or quantum mechanics are a plus, but not necessary. More importantly, the student should be pro-active and excited about designing opto-electrical quantum systems.
Developing High-Speed FPGA Architecture for Quantum Cryptography
"Quantum Plasmonic Interface Between Photons and Spins"
Faculty Advisor: Dirk R Englund
Mentor(s): Matt Trusheim
Contact e-mail: mtrush@mit.edu
Research Area(s): Applied Physics, Nanotechnology
Project description:

In recent decades, researchers have recognized that the world of quantum mechanics holds enormous potential for a new generation of applications that address unsolved problems in communications, computation, and high-precision measurements. The goal of this project is to engineer strong interactions between light (in the form of a surface plasmon polariton) and matter (in the form of atomic defect centers in diamond), to achieve a link between the two different qubits. The key element in this link is a metallic nanophotonic interfaces, or a plasmonic antenna, that allows efficient coupling between photons and these atomic quantum memories. This project will focus on designing and testing such metallic quantum interfaces. The goals will be to demonstrate efficient initialization and readout of the electronic and nuclear spins associated with the memory, to develop optical nonlinearities at the single-photon level, and to realize plasmon-mediated spin-spin interactions.

Your roles: You will design plasmonic nanostructures that maximize the coupling between light and matter, through the use of numerical simulation software. Later on in the project you will characterize your quantum devices through a variety of measurements including optical and spin-based spectroscopy. These skills form a base of knowledge that will prepare you for future work in optics, photonics, and quantum information.

Qualifications: We are looking for an undergraduate who has a strong background in electricity and magnetism, with quantum mechanics being a plus but not necessary. The student should be highly motivated and welcome challenges in scientific computing and laboratory experiments. Previous lab experience will be useful in transitioning from initial simulations to experimental work. Above all, we are looking for a student who is pro-active, ready to learn, and excited about designing optical devices and quantum systems.

Contact: Please contact Matt Trusheim (mtrush@mit.edu) if interested.
Quantum Plasmonic Interface Between Photons and Spins
"Scalable integration of high-performance lasers for on-chip quantum information processors"
Faculty Advisor: Dirk Englund
Mentor(s): Hyeongrak Choi
Contact e-mail: choihr@mit.edu
Research Area(s): Applied Physics, Nanotechnology
The past two decades have witnessed rapid advances towards building quantum information processors that promise to be exponentially more powerful than classical computers at certain tasks. Research groups, governments, and industry -- including Google, IBM, Microsoft, Raytheon, D-Wave etc -- are just a few examples of companies investing heavily into quantum computing. Among solid-state qubit systems, the nitrogen-vacancy center in diamond has emerged as an excellent optically addressable memory with second-scale electron spin coherence times. Recently, quantum entanglement and teleportation have been shown between two nitrogen-vacancy memories, but scaling to larger networks requires more efficient spin-photon interfaces and on-chip integration of quantum memories with high-performance optoelectronic control. One important problem is to find a way to provide optical pumping to hundreds of quantum registers, ideally with hundreds of individually tunable lasers. Developing such an architecture of chip-integrated lasers is the goal of this SuperUROP project. If successful, the project woudl have impact beyond quantum computing in fields such as sensing and high-speed optical interconnects in multiprocessor computing.



2. Description on the project
The specific goal of this project is to develop a process to integrate many low-power, tunable on-chip light sources for manipulating atomic quantum memories -- in particular here the Nitrogen Vacancy center in diamond [2], as illustrated in the figure. To achieve this, we will use photonic crystal nanocavities with a gain medium of quantum dots [3] or 2D materials. The student will learn these concepts and will dropcast
Aluminium Nitride ring resonators currently developed in the Quantum Photonics Groups group have quality factors Q ~10^5 that correspond to sub- GHz linewidth. The gain materials will be pumped using green laser light distributed throughout the photonic integrated circuit.
The project end goals are:

Development of process to integrate lasers on-chip, working with graduate students who already produce the chips using nanofabrication
Confirmation of lasing operation by light-in, light-out measurements
Power dependence of spectrum to confim lasing operation
Operation of lasers at 3K in a cryostat
(Optional) Investigation of high-speed modulation and tuning schemes

Students should have some background in E&M and should be enthusiastic to work in an optics laboratory in a dynamic, hard-working team of graduate students, postdoc, and scientists. The project also has components of theory and numerical modeling.

3. Reference
[1] T. Monz et al., Science 351, 1068-1070 (2016).
[2] H. Bernien et al., Nature 497, 86-90 (2013)
[3] Y. Wang et al., Nano reviews 2 (2011)
Scalable integration of high-performance lasers for on-chip quantum information processors
"Designing Quantum Plasmonic Devices"
Faculty Advisor: Dirk Englund
Mentor(s): Matt Trusheim
Contact e-mail: mtrush@mit.edu
Research Area(s): Applied Physics, Materials and Devices, Nanotechnology, Numerical Methods, Theoretical Computer Science
Project description: Our lab works in the field of experimental quantum information processing. We use various different physical systems, including color centers in diamond, quantum dots, and photons, to implement schemes ranging from quantum metrology and sensing to quantum computation. In this particular project, we are attempting to engineer strong interactions between light (in the form of a surface plasmon polariton) and matter (in the form of a solid-state quantum system), to achieve a link between two different forms of quantum bits. The key element in this link is a plasmonic nanostructure, made of metal patterned at scales well below the wavelength of light, whose design must be careful optimized.
Your roles: You will design plasmonic nanostructures that maximize the coupling between light and matter, through the use of numerical simulation software. Later on in the project you will characterize your quantum devices through a variety of measurements including optical and spin-based spectroscopy. These skills form a base of knowledge that will prepare you for future work in optics, photonics, and quantum information.

Qualifications: We are looking for an undergraduate who has a strong background in electricity and magnetism, with quantum mechanics being a plus but not necessary. The student should be undaunted when working with unfamiliar simulation software: programming experience is helpful. Above all, we are looking for a student who is pro-active, ready to learn, and excited about designing optical devices and quantum systems.
"Hydration monitoring in a wearable wristband"
Faculty Advisor: Luca Daniel
Mentor(s): Ian Butterworth
Contact e-mail: dluca@mit.edu
Research Area(s): Applied Physics, BioEECS, Circuits, Materials and Devices, Numerical Methods
Our hydration project (hydration.mit.edu) is focused on developing the technology missing for reliable and physiologically meaningful hydration tracking, with an aim of optimising hydration and avoiding dehydration. We created the project at MIT with a primary focus on the substantial need for improved hydration management in the elderly, but with a view to the needs in many sectors and the broader population.
We are currently carrying out clinical testing in Madrid and Boston of two novel non-invasive technologies in a non-miniaturised prototype, yet have designed both technologies to function on miniaturised architecture for a 24/7 wearable format, focusing on comfort and reliability of the underlying measurement.
Our approaches comprise of:
A non-contact RF approach that tracks bulk water volume fluctuations in the wrist by monitoring 
 water-dominated attenuation in the low GHz range through a proprietary approach that aims to make the bulk 
attenuation measurement insensitive to the spatial and rotational variations present in a wearable format.
Hydration monitoring in a wearable wristband
"Predictive Model of Urban e-Mobility integrated with Renewable Energy Sources"
Faculty Advisor: Luca Daniel
Mentor(s): Michela Longo
Contact e-mail: dluca@mit.edu
Research Area(s): Energy, Graphics and Human-Computer Interfaces, Numerical Methods
Latest technological developments and renewed attention to eco-sustainability have fostered the vision of smart power grid interacting with smart cities. This revolution is now quickly spreading also to the field of mobility (e.g. human transportation), originally sustained exclusively by fossil fuels. This project aims at developing a predictive model and a software tool that will enable the increase adoption of electrical vehicles while leveraging renewable energy sources to help maintain the power grid stable. The model will be developed and calibrated using available data (e.g. the electric infrastructure, the current number of vehicles, population density, wealth and attitude towards technological innovation). The tool will be able to predict the impact of electric vehicles usage and charging station locations on electric power distribution network. Hence users will be able to explore tradeoffs and “what if” scenarios for instance with the objective of optimal placement of private and public charging stations. More specifically, our model will exploit the fact that electric vehicles may work either as loads (requiring energy from the grid and possibly absorbing surplus from local renewables), or as generators (feeding energy back into the grid, thus supporting the supply network in case of micro-shortages, especially at the local level). Predictive Model of Urban e-Mobility integrated with Renewable Energy Sources
"Active caches: Guiding processors over the memory wall"
Faculty Advisor: Daniel Sanchez
Mentor(s): Anurag Mukkara, Guowei Zhang
Contact e-mail: anurag_m@csail.mit.edu, zhanggw@csail.mit.edu, sanchez@csail.mit.edu
Research Area(s): Computer Systems
Data movement is a key bottleneck in modern computer systems: memory accesses and communication are orders of magnitude more expensive than compute operations. Unfortunately, current systems suffer from a processor-centric design that causes much more data movement than needed. Processors treat sophisticated cache hierarchies as passive storage elements, and programs have no visibility into cache contents, so they cannot shape their behavior to reduce data movement.

In this project, you will work with a team of students to investigate and design a new memory system organization that turns on-chip caches into active agents, letting them guide what, where, and when computation is done. Caches are in an ideal position to reshape computation because they know what data can be accessed efficiently (i.e., the data already in the cache), and much of their logic can be reused for this purpose. You will focus on one of several applications of active caches, such as (i) eliminating redundant computation through pervasive memoization, (ii) scheduling and reorder tasks in graph-processing and other irregular algorithms to reduce their data movement, and (iii) performing transparent data transformations like compression and encryption.
Active caches: Guiding processors over the memory wall
"A memory hierarchy for modern programming languages"
Faculty Advisor: Daniel Sanchez
Mentor(s): Po-An Tsai
Contact e-mail: poantsai@csail.mit.edu, sanchez@csail.mit.edu
Research Area(s): Computer Systems
Current computer system abstractions were designed for unmanaged programming languages, like C and Fortran. In particular, memory hierarchies expose a flat address space and rely on expensive virtual memory mechanisms to provide protection and isolation among programs. But modern languages, like Go, Java, and Javascript, rely on automatic memory management and disallow raw memory accesses to eliminate many types of memory-access bugs. While productive, these languages are mismatched with current abstractions: they need expensive runtime support to manage their flat memory address space (e.g., garbage collection), and virtual memory is overkill for them and incurs unnecessary overhead.

In this project, you will investigate a new memory hierarchy and hardware/software interface that are specifically designed for modern languages with automatic memory management. The system will perform hardware-accelerated memory management, will leverage the rich object-level information available in modern languages to better place data along the memory hierarchy, and will use lightweight protection mechanisms that retain safety while incurring negligible overhead.
"Unlocking the Potential of Multicore Systems"
Faculty Advisor: Daniel Sanchez
Mentor(s): Mark Jeffrey
Contact e-mail: mcj@csail.mit.edu, sanchez@csail.mit.edu
Research Area(s): Computer Systems
Current multicores suffer from two main limitations: they can only exploit a fraction of the parallelism available in applications, and they are very hard to program. We are designing a new type of multicore architecture that tackles both problems. In this architecture, called Swarm, programs consist of very short tasks, as small as tens of instructions each. Hardware queues and distributes tasks among cores, reducing the overheads of fine-grain parallelism and allowing many more applications to be parallelized. Moreover, parallelism is implicit: instead of using locks, semaphores, or other error-prone explicit synchronization techniques, programmers simply define an order among tasks. Under the covers, Swarm hardware figures out what order constraints are superfluous and elides them, running most tasks in parallel. As a result, Swarm programs are almost as simple as their sequential counterparts, and at the same time outperform the best parallel programs.

We have already demonstrated Swarm's benefits on challenging applications, like graph analytics and databases, achieving speedups of 50-300x on several parallel algorithms where no prior parallel implementations could beat the sequential ones. But there is much more to do! Given the broad scope of this project, there are many areas where you can contribute, depending on your interests. For example, you can:

- Port challenging parallel applications to Swarm (e.g., circuit and network simulators, SAT solvers, or a program of your choice).

- Develop new profiling and visualization tools that help programmers understand performance and parallelism bottlenecks.

- Design new techniques to parallelize sequential programs almost automatically.

- Help build an FPGA implementation of Swarm.
Unlocking the Potential of Multicore Systems
"Low Cost Micromanipulator"
Faculty Advisor: Joe Steinmeyer
Mentor(s):
Contact e-mail: jodalyst@mit.edu
Research Area(s): BioEECS, Circuits, Control
Automated micromanipulators are electromechanical devices enabling very fine positioning in 3 dimensions (often far less than one micron xyz spatial resolution). They are used widely in biology and neuroscience for sample manipulation and interfacing, and have emerging roles in high-resolution 3D printing and other applications. Current commercial solutions are very expensive (>$5K a piece for low-end models and significantly more for higher-end ones). This project is investigating how to design and manufacture a prototype micromanipulator with similar specs to commercial models but at a lower cost, thus allowing wider deployment in high-throughput biology and 3D printing applications.
"Learning Causal Graphs and Applications to Gene Regulation"
Faculty Advisor: Caroline Uhler
Mentor(s):
Contact e-mail: cuhler@mit.edu
Research Area(s): Artificial Intelligence, BioEECS, Control, Numerical Methods, Signals and Systems, Theoretical Computer Science
Causal inference is a cornerstone of scientific discovery because it asks “why?”. Most methods for learning causal directed graphs assume that the underlying graph is a DAG, i.e., that it does not contain any directed cycles. However, feedback loops in biological networks are not only common but also crucial features. In addition, many causal inference algorithms do not allow for imposing prior knowledge on the directed graph or cannot be applied to large networks. In this project, the goal is to develop causal inference algorithms that can overcome these limitations and can be applied to infer gene regulatory networks. These networks have about 20'000 nodes, but there is a lot of prior information on the network coming from knock-out experiments. In order to increase the power of the methodology, it is important to be able to use this prior information. What statistical guarantees can be obtained and what is the computational trade-off? How well does the algorithm perform on simulations? Does it provide meaningful gene regulatory networks when applied to real biological data?
"Learning Brownian Motion Trees and Applications to Cell Differentiation"
Faculty Advisor: Caroline Uhler
Mentor(s):
Contact e-mail: cuhler@mit.edu
Research Area(s): BioEECS, Numerical Methods, Signals and Systems, Theoretical Computer Science
Every cell in our body contains the same genetic information. However, we have many different cell types and they all show different gene expression patterns. In order to understand how this is possible, we will study how the different gene expression patterns develop, starting from a stem cell and differentiating into the different cell types. We will assume that gene expression develops along the tree according to a multivariate Brownian motion with correlation among the genes. Given gene expression data from various points along the unknown differentiation tree, the goal of this project is to develop algorithms that can simultaneously map the cells to the differentiation tree and learn the underlying tree topology. What statistical guarantees can be obtained? How well does the algorithm perform on simulations? Does it provide meaningful differentiation trees when applied to real biological data?
"Ellipsoid Packing and Applications to Chromosome Organization"
Faculty Advisor: Caroline Uhler
Mentor(s):
Contact e-mail: cuhler@mit.edu
Research Area(s): Applied Physics, Artificial Intelligence, BioEECS, Control, Numerical Methods, Signals and Systems, Theoretical Computer Science
The spatial organization of the genetic material in the cell nucleus is known to be important for gene regulation. During most of the cell cycle each chromosome occupies a roughly ellipsoidal domain in the cell nucleus. Hence, the spatial organization of chromosomes can be modeled as an ellipsoid packing problem: 46 ellipsoids of a given size and shape (the chromosomes) should be packed into an ellipsoidal container (the cell nucleus) so as to minimize their overlap under various constraints. The goal of this project is to use optimization theory to develop algorithms that find locally optimal minimal overlap configurations of ellipsoids under various constraints. Biologically relevant constraints include distance to the boundary of the container, overlap between specific ellipsoids, or changing container shapes. The obtained ellipsoid configurations will be compared to experimental data and used to predict the new chromosome neighborhoods when altering the shape of the nucleus.
"Decoding visual information from the human brain"
Faculty Advisor: Aude Oliva
Mentor(s): Dimitrios Pantazis
Contact e-mail: pantazis@mit.edu
Research Area(s): Applied Physics, Artificial Intelligence, BioEECS, Circuits, Signals and Systems
The human brain can rapidly and effortlessly recognize complex visual information within only a couple hundred milliseconds. To understand this remarkable behavior, we use magnetic sensors surrounding the head to measure neuronal signals as the brain transforms low level visual information into semantic content. We are recruiting motivated undergrads interested in collecting and analyzing brain activation data, and developing methods to decode information from brain magnetic sensors. Our goal is to systematically characterize the spatiotemporal and representational space of brain activity in different cognitive tasks, which can eventually offer insights in designing new computer vision models. Decoding visual information from the human brain
"Computer vision for infographics and visualizations"
Faculty Advisor: Aude Oliva
Mentor(s): Zoya Bylinskii
Contact e-mail: zoya@mit.edu
Research Area(s): Artificial Intelligence, Graphics and Human-Computer Interfaces, Signals and Systems
Given a graphic design or information visualization (graph, chart, infographic) as input, we are working on neural network models that automatically discover the most relevant regions of these designs, predict where people look, and recognize the important text. We are exploring automatic design and inference tools built on top of such predictions, including (1) visual thumbnailing and text summarization, (2) interactive feedback in design tools and optimization of element layout, and (3) automatic question-answering for visualization understanding. We are recruiting motivated undergrads interested in working together in a team with graduate students, with a focus on the computer vision, natural language processing, or user interface/design elements of this project. Computer vision for infographics and visualizations
"Accessibility of Mobile Apps"
Faculty Advisor: Lalana Kagal
Mentor(s):
Contact e-mail: lkagal@csail.mit.edu
Research Area(s): Computer Systems, Graphics and Human-Computer Interfaces
Handheld devices including mobile phones, tablets, and smartwatches have replaced desktops and laptops as the primary communication and computation platform for many individuals. With almost one billion people with disabilities in the world, ensuring that apps for these handheld devices are accessible by people with disabilities is important and becomes highly critical during emergency situations, when dissemination of evacuation information and coordination of relief efforts are essential. Though mobile apps are partially covered by the same accessibility standards as Web applications, there aren’t sufficient tools for developers who want to build accessible mobile apps. As a result, most developers do not take the initiative to make their apps accessible, and disabled users are left behind with respect to mobile device usage. In this project, we are interested in facilitating the development of accessible mobile applications to open up the world of mobile computing to more users with disabilities.

Currently we have an initial library and testing framework for Android Studio. We're looking to expand this work, conduct user testing, continue engaging with the accessibility community, and develop support for iOS. Ideal candidates should have some mobile app development experience and be passionate about accessibility.
Accessibility of Mobile Apps
"Building a full-stack system and machine learning algorithms for smart home sensing"
Faculty Advisor: Dina Katabi
Mentor(s): Chen-Yu Hsu, Rumen Hristov
Contact e-mail: cyhsu@mit.edu, rhristov@mit.edu
Research Area(s): Artificial Intelligence, Computer Systems
Our group has built a device that can localize people without requiring the user to carry any sensor on their body and it works through walls Link We have expanded the system to capture human figures and recognize the people Link and we used machine learning to even recognize the emotions of a person using wireless signals Link

We are looking to bring these technologies into people’s homes to enable smart homes that react to people habits and liking. For example, the home can detect when the user sits on the TV couch and tune to his favorite channel. It can also collect analytics on how much time the user sleeps, watches TV, or works at his desk. We have deployed more than 20 devices in users’ homes. We have also built a full stack software system to support our current deployment. The system includes a backend Django server, Cassandra instances for storing data and a web page for visualizing the analytics.

We are looking for students who are interested in developing higher level analytics and making the system scalable and reliable as we deploy more devices. We have projects ranging from machine learning and signal processing to web server infrastructure and web programming

Prerequisites: programming experience with Python
Building a full-stack system and machine learning algorithms for smart home sensing
"Grounding Language and Actions in Perception"
Faculty Advisor: Boris Katz
Mentor(s): Andrei Barbu
Contact e-mail: boris@csail.mit.edu
Research Area(s): Artificial Intelligence
Are you interested in getting robots to follow commands, understand natural language, and engage in dialog with humans? We are developing such robots along with a range of other vision and language applications.

Our project aims to develop new approaches to grounding language and
actions in perception combined with the ability to reference the past and remember key facts. We are looking to expand our work to more challenging robotic domains, complex multi-step plans, partially-observed worlds, collaboration with other agents, and dialog with humans. We're combining together many strands of research: natural language processing for grounding language; vision to understand the environment and to disambiguate commands; planning to guide robots, and more.

Great candidates will be interested in how these areas combine together to lead to more intelligent machines that can be controlled by humans and integrated into society.
"Building Apps from Concept Cliches"
Faculty Advisor: Daniel Jackson
Mentor(s): Santiago Perez De Rosso
Contact e-mail: dnj@mit.edu
Research Area(s): Computer Systems
Déjà Vu is a new platform for end-user development of apps with rich functionality. It features a novel theory of modularity for binding concepts; an extensive library of reusable concepts; and a WYSIWYG tool for specifying bindings and customizing visual layout.

As a user you might have noticed the fundamental similarities between the many applications you use daily. Maybe it was the day you were scrolling through your Facebook news feed and then through your Twitter feed? Or when you gave a 5-star review to a restaurant in Yelp, and then to a book in Amazon? Or that time when you replied to a tweet and found yourself later replying to a comment on Reddit?

Now picture the many software engineers developing web applications, ranging from internal business applications to those used by millions of users. How many of these engineers are, at this moment, working on implementing a password recovery mechanism? How many are adding some kind of news feed to their application? What about adding chat functionality? A shopping cart? Adding star ratings or likes? Letting users write comments? In each of these instances, developers are not all doing the exact same thing. In some cases, the feed is listing posts authored by users, in other cases it’s showing shopping products, or books. Some developers need the feature to be tweaked in a unique way, or are using different languages and frameworks. Fundamentally, they are all doing the same thing: combining pre-existing concepts in novel ways. If we could successfully exploit this fact, applications could be built much faster than how they are built today. Déjà Vu does exactly this.

The current state (March 2017) is that we have a prototype and are building a suite of sample applications. By next fall we expect to have advanced the project considerably, and to have a variety of new opportunities in extending and applying the framework.
"A New Computational Engine for Apps"
Faculty Advisor: Daniel Jackson
Mentor(s): Matt McCutchen
Contact e-mail: dnj@mit.edu
Research Area(s): Computer Systems
We are developing a new platform for developing client-server apps with an order of magnitude less code. The core of the platform is a new kind of spreadsheet that replaces the traditional flat spreadsheet with a hierarchical sheet that offers the expressive power of a relational database, but with a simpler and more succinct formula language. A variety of projects are available in developing the computational model, building a declarative security mechanism, filling out challenging aspects of the platform design, and in applying the platform to new application areas such as IoT.
"Soft Robots"
Faculty Advisor: Daniela Rus
Mentor(s):
Contact e-mail: rus@csail.mit.edu
Research Area(s): Artificial Intelligence, Materials and Devices
Soft-bodied robots, running appropriate control algorithms, promise to be safe for interaction with humans and push the envelope on machine capabilities. Soft robots have bodies made out of intrinsically soft and/or extensible materials (e.g. silicone rubbers) that can deform and absorb much of the energy arising from a collision. The nature of the soft body offers the potential for unprecedented adaptation, sensitivity, and agility, such as (1) Moving in a way that allows the robot to bend and twist with high curvatures and in turn this allows the robot to adapt to, work in, and travel through, confined spaces; (2) Deforming the robot body in a continuous way and which leads to movements that emulate the motion of biological systems; 3) Adapting their shape of the body to the environment and relying on compliant motion to grasp and manipulate un-modeled objects, move on rough terrain, and exhibit resilience, and (4) Executing rapid, agile maneuvers, such as the juggling, reaching under and over an object in a continuous motion, or imitating the escape maneuver in fish. Our project has opportunities for each of these directions.
"Printable Robots"
Faculty Advisor: Daniela Rus
Mentor(s):
Contact e-mail: rus@csail.mit.edu
Research Area(s): Artificial Intelligence, Graphics and Human-Computer Interfaces, Materials and Devices
Designing and fabricating new robotic systems is typically limited to experts, requiring engineering background, expensive tools, and considerable time. In contrast, to facilitate everyday users in developing custom robots for personal use, this project aims to create algorithms, systems, and tools to easily create printable robots from high-level structural specifications. From that, the system generates complete mechanical drawings suitable for fabrication, instructions for the assembly of electronics, and software to control and drive the final robot.This project aims to develop steps towards creating a hardware compiler. Printable Robots
"iDiary"
Faculty Advisor: Daniela Rus
Mentor(s):
Contact e-mail: rus@csail.mit.edu
Research Area(s): Artificial Intelligence, Computer Systems, Theoretical Computer Science
Information extraction from mobile phones, smart glasses, or robots sensors such as GPS and video over long periods of time enable the autonomous analysis of the activity stream for the mobile agent that generated these capabilities. The collected data is valuable for mapping, situation awareness, and modeling behaviors, but requires efficient tools that can extract the right information at the right time efficiently. In this project we develop algorithm and system for autonomously extracting activities from GPS and video streams. iDiary
"Quantum CMOS Design"
Faculty Advisor: Rajeev Ram
Mentor(s):
Contact e-mail: rajeev@mit.edu
Research Area(s): Applied Physics, Materials and Devices, Nanotechnology, Numerical Methods
This project involves the design of single photon counting detectors in VLSI CMOS. These devices are critical building blocks for a host of quantum information processing devices and molecular sensors. The project involves simulation, design, and testing of photon counting devices fabricated in deep sub-micron CMOS. Design will involve numerical simulation using Synopsys Sentaurus Device. These models will be benchmarked against preliminary detectors fabricated in 65nm and 90nm CMOS processes. The validated model will be used to design a next generation device. This optimized photon counter will be used as a critical building block for quantum information processing and molecular sequencing systems being developed in the group. Quantum CMOS Design
"Computational Spectroscopy"
Faculty Advisor: Rajeev Ram
Mentor(s):
Contact e-mail: rajeev@mit.edu
Research Area(s): Applied Physics, Signals and Systems
Recently our lab has developed wearable spectrometers that are capable of achieving resolution comparable to lab systems. These miniature spectrometers use a combination of CMOS imaging, optical design, and machine learning to realize chip-scale spectrometers that operate over a broad wavelength range. This project is working to explore the ultimate performance limits for these devices. In particular, we will explore the sensitivity limits for computational spectrometers with particular focus on fluorescence spectroscopy. Fluorescence sensing is used for applications ranging from chemical sensing, medical diagnostics, to the characterization of new materials. Computational Spectroscopy
"Rapid Blood Analysis"
Faculty Advisor: Rajeev Ram
Mentor(s):
Contact e-mail: rajeev@mit.edu
Research Area(s): Applied Physics, BioEECS, Signals and Systems
Blood analysis is core to medical diagnostics and disease management. This project explores rapid optical spectroscopy on microscopic droplets of whole blood for on-site, clinical analysis. The hope is that we can perform a whole blood work up and identify biomarkers in the doctor's office or the patients home without the expense and time delay of lab testing. Students will learn to perform optical spectroscopy, use BL2 protocols, and perform chemical analysis. Chemometric techniques that combine machine learning, optics, and physical chemistry will be used to extract diagnostic information from the spectra. Rapid Blood Analysis
"Leveraging Clinical Data Sets to Optimize Oxygen Delivery to Newborns"
Faculty Advisor: Thomas Heldt
Mentor(s): Dr. Wendy Timpson
Contact e-mail: thomas@mit.edu
Research Area(s): Artificial Intelligence, BioEECS
Tight oxygen titration in the preterm neonate is a key aspect of neonatal intensive care due to the mortality associated with hypoxia (low oxygen saturation) and the morbidity associated with hyperoxia (high oxygen saturation) in this vulnerable population. Despite these known complications of sustained oxygenation outside target ranges, most Neonatal Intensive Care Units (NICUs) fail to reliably maintain infants’ oxygenation saturations within target range. The primary goal of this project is to leverage large volumes of physiological data streams collected in the NICU to identify clinical, demographic, physiological and workflow factors that place preterm infants at risk for hypoxia and hyperoxia.

This project offers an opportunity to partner with neonatologist and clinical staff at Beth Israel Deaconess Medical Center and actively participate in data analytics to directly improve the care of the tiniest patients.
"Secure IoT"
Faculty Advisor: Prof. Anantha Chandrakasan
Mentor(s): Dr. Rabia Tugce Yazicigil
Contact e-mail: rty@mit.edu
Research Area(s): Communications, Computer Systems
Security is the most important consideration in future low-power wireless networks focused on connecting edge devices. Internet of Things devices such as building sensors, health monitors, and industrial equipment often communicate using the Bluetooth Low Energy (BLE) protocol.

We propose to develop secure radio systems that offer new security paradigms through non-conventional frequency hopping techniques for Internet of Things.

The project will involve the following opportunities:

* Developing systematic methods to analyze the security attacks to IoT devices communicating wirelessly. These methods are expected to capture the effects of personalized attacks.

* Investigating possible solutions to overcome personalized attacks due to increasing connectivity.

* Building an off-the-shelf radio and FPGA platform as a demonstration of connected sensors using BLE. First part of the demonstration will focus on showing how the BLE protocol with traditional packet-level frequency hopping is vulnerable to selective jamming attacks.

* Building a compact demo platform for a custom-designed chip + FPGA. This platform will be replicated for a multi-transmitter network that uses a novel secure wireless protocol. Includes developing a GUI and real-time processing on the FPGA.

Experience or related coursework (e.g. 6.033, 6.857 and 6.02 background) is a plus.
Secure IoT
"Machine learning of disease progression in multiple myeloma"
Faculty Advisor: David Sontag
Mentor(s):
Contact e-mail: dsontag@csail.mit.edu
Research Area(s): Artificial Intelligence, BioEECS
Multiple myeloma, a rare blood cancer, is believed to be one of the most promising targets for precision medicine. Over the last decade there has been enormous progress toward developing novel treatment therapies, with over 10 new drugs on the market and many more in clinical trials. However, little is known yet about which drugs work best for whom. Evaluating this from data is challenging because no two patients are identical, with each person having different gene expression, mutations, biomarker levels, treatment strategies, and comorbidities. This project will work with patient data from the landmark CoMMpass study by the Multiple Myeloma Research Foundation, an ongoing clinical trial that already has several years' data from nearly 1000 patients. We have multiple goals, from discovering the genetic drivers of the disease, to using machine learning to predict when a patient will progress to the next stage of the disease, and building in-silico models that can predict which treatment will work best for a patient. Machine learning of disease progression in multiple myeloma
"Machine learning to predict pregnancy complications"
Faculty Advisor: David Sontag
Mentor(s):
Contact e-mail: dsontag@csail.mit.edu
Research Area(s): Artificial Intelligence, BioEECS
Can we figure out which pregnancies will result in complications such as preterm birth, miscarriage, or preeclampsia? This project uses machine learning on health data from over a hundred thousand pregnancies to learn a model which would allow us to predict early in the pregnancy whether a complication is likely. We seek to make these predictions using deep learning on high-dimensional time-series data consisting of patients' past diagnoses, procedures, medications, and laboratory test results. Other directions we may explore include learning interpretable models, seeking to understand the causes of pregnancy complications using causal inference, and considering other modalities such as fetal heart monitoring or ultrasound. Machine learning to predict pregnancy complications
"Trajectories Like Mine: Machine Learning in Healthcare:"
Faculty Advisor: Una-May O'Reilly
Mentor(s):
Contact e-mail: alfa-apply@csail.mit.edu
Research Area(s): Artificial Intelligence
The machine learning problem of “trajectories like mine” is to efficiently find patients with physiological waveforms similar to a reference waveform. Once a similarity set is found, it can be exploited for future or diagnostic extrapolations to the patient of reference without model-based learning. One ML approach for retrieving “trajectories like mine” is locality sensitive hashing. We are interested in practical implementations of LSH extensions for prediction problems in EEG, ECG or arterial blood pressure (ABP). Can different hashing families be combined? Is there different hashing families for different types of predictions or data? Trajectories Like Mine: Machine Learning in Healthcare:
"Adversarial Neural Networks for Cyber Security"
Faculty Advisor: Una-May O'Reilly
Mentor(s): Stjepan Picek
Contact e-mail: alfa-apply@csail.mit.edu
Research Area(s): Artificial Intelligence
Computer systems are easy to attack if considered in a static scenario. The adversary has the advantage in time to study the system, find its vulnerabilities and choose the place to attack. To counter that, one can use the concept of moving target defense (MTD) by making the system dynamic and consequently more difficult for attacker to exploit since he also then has to deal with a great deal of uncertainty just like defenders do. This project aims at using adversarial neural networks concept in order to model the dynamics between the defender and attacker. There, both defender and attacker would be represented with a neural network that learn how to perform tasks of defense and attack, respectively. Adversarial Neural Networks for Cyber Security
"Coding the Tax Code: Regulation to Formalism"
Faculty Advisor: Una-May O'Reilly
Mentor(s): Erik Hemberg
Contact e-mail: alfa-apply@csail.mit.edu
Research Area(s): Artificial Intelligence
AI techniques exist that translate case law into software and that support intelligent reasoning and argumentation around it. This project focuses alternatively on the regulatory form of law, e.g. tax law. It will involve developing an automatic parsing system for translating regulations into a formalism. It is part of the larger STEALTH project (Link Link ) This project will appeal to students interested in programming languages and/or natural language text understanding and representation techniques. Coding the Tax Code: Regulation to Formalism
"Machine Learning and Cyber Security in Peer-to-Peer Networks"
Faculty Advisor: Una-May O'Reilly
Mentor(s): Erik Hemberg
Contact e-mail: alfa-apply@csail.mit.edu
Research Area(s): Artificial Intelligence, Computer Systems
Denial of Service (DoS) Cyber attacks continue to increase and cause numerous disruptions in both industry and politics. With more and more critical information moving through networks, it is important to keep these networks available. A Peer-to-Peer network can be utilized against DoS attacks due to its centralized nature, but the separation between physical and logical layer is still challenging. The project will involve applying machine learning to investigate how to secure Peer-to-Peer networks against autonomous and adaptive adversaries. It is ideal for students planning on taking or who have taken 8.857 and/or 6.858 Machine Learning and Cyber Security in Peer-to-Peer Networks
"Using Machine Learning to Reduce False Arrhythmia Alarms in the Intensive Care Units"
Faculty Advisor: Roger Mark
Mentor(s): Li-wei Lehman
Contact e-mail: rgmark@mit.edu
Research Area(s): BioEECS
The Laboratory for Computational Physiology (lcp.mit.edu) at MIT is currently engaged in NIH-funded projects focused on the study of complex biomedical and physiologic signals. The aim of this project is to explore the use of machine learning techniques, including neural networks, to address the problem of reducing false arrhythmia alarms (e.g., asystole, extreme bradycardia, extreme tachycardia, ventricular tachycardia, and ventricular flutter/fibrillation) in intensive care units (ICUs) and distinguishing clinically important events from noise and artifact. ICU false alarm rates have been reported to be as high as 86%, and can lead to disruption of care and slow response time due to desensitization of clinical staff. We aim to develop techniques based on multiple physiological waveforms measured during routine clinical monitoring (ECG, arterial blood pressure, photoplethysmogram, and respiration). Faculty Advisor: Professor Roger Mark. Mentor: Dr. Li-wei Lehman (lilehman@mit.edu) Using Machine Learning to Reduce False Arrhythmia Alarms in the Intensive Care Units
"Pattern Analysis of Large-Scale Health Data to Improve Patient Outcomes"
Faculty Advisor: Roger Mark
Mentor(s): Dr. Li-wei Lehman
Contact e-mail: rgmark@mit.edu
Research Area(s): BioEECS
Large-scale health records are a vital resource in medical research, and provide an opportunity to better understand the associations between complex disease processes and patient outcomes. The Laboratory for Computational Physiology (lcp.mit.edu) at MIT is currently engaged in NIH-funded research to develop and study a major research database of physiologic and clinical data from intensive care patients that will support the design and evaluation of advanced patient monitoring algorithms. This project aims to apply machine learning techniques to analyze large volumes of physiological and clinical data from an Intensive Care Unit (ICU) database for patient phenotyping, risk stratification, and outcome prediction. The goal of this project is to develop an early warning system to alert clinicians of patients at high-risk of developing adverse events so as to enable timely interventions for improved patient outcomes. This project requires familiarity with basic machine learning concepts and techniques. Faculty Advisor: Professor Roger Mark. Mentor: Dr. Li-wei Lehman (lilehman@mit.edu)
"An Energy-Efficient FPGA Computer Vision platform for Real-time Object Detection"
Faculty Advisor: Vivienne Sze
Mentor(s):
Contact e-mail: sze@mit.edu
Research Area(s): Artificial Intelligence, Circuits, Computer Systems, Signals and Systems
Object detection is needed for many embedded vision applications including surveillance, advanced driver assistance systems (ADAS), portable electronics and robotics. For these applications, it is desirable for object detection to be real-time, robust and energy-efficient. Real-time processing is necessary for applications such as ADAS, and autonomous control in unmanned aircraft vehicles (UAV), where the vehicle needs to react quickly to changing environments. High frame rate enables faster detection to allow more time for course correction. High-resolution images enable early detection by having enough pixels to identify objects at far distances. Finally, in both UAV and portable electronics, the battery, whose weight and size must be kept to a minimum, limits the available energy. For ADAS on the other hand, the power consumption is limited by the heat dissipation. Thus, energy-efficient object detection is also desirable.

The goal of this project is to create an energy-efficient FPGA platform for vision processing to support real-time high definition object detection. There are several opportunities to get involved in this project ranging from:
** RTL design for the FPGA
** PCB design to build a compact FPGA + sensor system
** Embedded software development for an ARM core
** GUI to support real-time annotated display and system configuration

Experience or related coursework (e.g. 6.111) is a plus.
An Energy-Efficient FPGA Computer Vision platform for Real-time Object Detection
"Constellation"
Faculty Advisor: Max Goldman
Mentor(s):
Contact e-mail: maxg@mit.edu
Research Area(s): Graphics and Human-Computer Interfaces
Constellation enables collaborative programming in the Eclipse IDE -- think Google Docs for Eclipse. It is designed for active learning in the classroom, with students working in pairs on small exercises. We use it nearly every class in 6.031 Software Construction. In addition to collaborative editing for students, Constellation allows course staff to review and assess students' work.

See: Link

Some possible research projects are listed below. On any project, you will review related work; design and implement new capabilities; evaluate them either in deployment to 6.031 or in a lab study; and report on the results.

- Improved collaborative editing: add missing collaboration features that help students focus on learning rather than logistics
- Visualizing single pairs: better visualizations that capture more of each pair's process and results
- Visualizing all pairs: visualizations that help staff identify struggling students, see who is done, or gauge the progress of an entire multi-hundred-student class
- Capture and use more data: these visualizations might benefit from more data about what students do in Eclipse; or new data might power algorithms that assist students or staff
- Beyond the classroom: e.g. Constellation could enable on-line office hours where staff help remotely

Prerequisites: 6.005/6.031 and at least one of 6.813 or 6.170.
Constellation
"Recoloring 3D Printed Objects using a Projector and Bi-stable Photochromic Inks"
Faculty Advisor: Stefanie Mueller
Mentor(s):
Contact e-mail: stefanie.mueller@mit.edu
Research Area(s): Graphics and Human-Computer Interfaces, Materials and Devices
The goal of this project is to build a system that can recolor existing objects on the fly. As an example, imagine having a pair of shoes for which on any given day you can decide which color or pattern they should show ('dynamic product design').

We enable this through the use of bi-stable photochromic inks. Photochromic inks change color when exposed to light of a specific frequency and if they are bi-stable they will keep this color even when removed from the activation light source (think of 'flipping' in e-ink displays).

For instance, a photochromic pixel can flip from transparent to green when exposed to UV light, and keep this color until exposed to IR light which flips the pixel back from green to transparent. To enable this workflow, we will equip a projector with the right light sources for activation and deactivation.

The software you will build will enable the following workflow: Users put the object in front of the projector. Your software will projection map the desired design onto the object. The user can remove the object from the projector and the pattern will be permanent until the user puts it back and the projector erases the pattern.

My group works at the intersection of software / hardware / materials. For this project you should have a strong background in software to develop the described tool, but you should be comfortable getting your hands off the keyboard and onto some real physical prototypes.
Recoloring 3D Printed Objects using a Projector and Bi-stable Photochromic Inks
"3D Printing Material to Fabrication Pipeline"
Faculty Advisor: Stefanie Mueller
Mentor(s):
Contact e-mail: stefanie.mueller@mit.edu
Research Area(s): Graphics and Human-Computer Interfaces, Materials and Devices
The goal of this project is to implement an automated pipeline for 3D printing with novel materials.

When trying to print with a novel material, users encounter many problems, e.g. when mixing the 'base ingredients' for the material the ratios depend heavily on the fabrication process used (e.g. inkjet 3D printing vs. syringe based extrusion), and thus the mixing is typical done by a material scientist or chemicist.

To enable ordinary users to create their own 3D printing materials, we will build a software tool that abstracts this expert knowledge and automatically generates the right 'recipe' and fabrication parameters. Having such a tool will rapidly advance the range of objects that can be made with 3D printing.

My group works at the intersection of software / hardware / materials. For this project you should have a strong background in software to develop the described tool, but be open to get your hands onto some material datasheets as well.

You will work closely with me and my postdocs.

Link
3D Printing Material to Fabrication Pipeline
"Fabricating 3D Displays"
Faculty Advisor: Stefanie Mueller
Mentor(s):
Contact e-mail: stefanie.mueller@mit.edu
Research Area(s): Graphics and Human-Computer Interfaces, Materials and Devices
The goal of this project is to free displays from their rectangular form factors and to enable fabrication of freeform displays of any shape ('everything can be a display').

We will explore different fabrication methods to build such a display (e.g., 3D printing, thermoforming, hydrographics, robotic fabrication, and pad printing).

We will also implement a design software that allows product designers to easily create content for this new type of display medium.

I'm looking for strong students who are interested in working at the intersection of software / hardware / materials. You don't have to be an expert in all three, but you should be comfortable getting your hands off the keyboard and onto some real physical prototypes.

You will work closely with me and my postdocs.

Link
Fabricating 3D Displays
"ECG Pattern Matching"
Faculty Advisor: Roger Mark
Mentor(s): Alistair Johnson
Contact e-mail: rgmark@mit.edu
Research Area(s): BioEECS, Signals and Systems
Our lab is collecting physiologic waveforms including electrocardiograms and photoplethysmograms from patients in the emergency department (ED) of the Beth Israel Deaconess Medical Center. The waveforms are continuous whenever the patient remains in the same room and is attached to the same monitor. However, in the ED patients are frequently moved in order to get imaging studies or other tests, and when they return they are often attached to a different monitor. We collect all waveforms from all monitors. Typically, segments of a given patient’s data may be found among the outputs of several monitors. The goal of the project is to identify all segments of a given patient’s waveforms and reassemble them into a continuous stream. The pattern matching problem could be stated: given a set of 2-lead ECG segments (minutes to hours long), identify which ECGs belong to a given patient. ECG Pattern Matching
"Billion-fold biomolecule preconcentrator for TB diagnostics"
Faculty Advisor: Jongyoon Han
Mentor(s): Wei Ouyang (EECS Ph.D. candidate)
Contact e-mail: jyhan@mit.edu
Research Area(s): Applied Physics, BioEECS, Materials and Devices
Detection of ultra low-abundance biomarkers is critical for the early diagnosis of diseases. Currently there does not exist a technique for protein amplification, and the PCR technique for DNA amplification is error-prone and expensive. We propose to develop a billion-fold microfluidic electrokinetic biomolecule concentrator, which will dramatically increase the concentration of target biomarkers by squeezing them from a 1 mL sample into a volume of 1 pL. We have demonstrated a single-stage concentrator with a concentration capacity of 0.1 million fold within 15 min. Students will be participating in the design of a highly multiplexed, multi-stage, three-dimensional microfluidic concentrator to further enhance the concentration capacity. Upon the completion of this work, we are able to support the student for a trip to Peru to validate the device in field for TB and other disease diagnostics, under a MIT-Peru MISTI project grant. (possibly in Jan 2018) Billion-fold biomolecule preconcentrator for TB diagnostics
"Understanding language to see the world, and seeing the world to understand language"
Faculty Advisor: Boris Katz
Mentor(s): Andrei Barbu
Contact e-mail: boris@csail.mit.edu
Research Area(s): Artificial Intelligence
Why is human vision so much better than machine vision? How do we describe what we see? How do you recognize an event that someone is describing? How do you learn language when all you hear are sentences with words you don't understand?

We are interested in understanding these questions by developing models that jointly address vision and language. These models should allow you to perform tasks that come naturally to children but are difficult for machines: describing what you see, seeing what is being described, asking questions about what you see, determining if what you are seeing is reasonable, using knowledge from language to change your mind about what you are seeing, etc.

This project aims to develop new models, extend current models to novel environments, or answer novel questions about language and vision.
"Using natural language processing to identify advance care planning among critically ill patients"
Faculty Advisor: Roger Mark
Mentor(s): Leo Celi
Contact e-mail: lceli@mit.edu
Research Area(s): BioEECS
Project description: Advance Care Planning (ACP) involves documentation of patients’ goals and priorities, and has been associated with improved healthcare quality and reduced cost. New quality measures and aspects of the Medicare payment reform focus on ACP. However, the inability to efficiently identify ACP in medical records represents a critical obstacle. Currently methods rely on manual chart review, which is expensive and not scalable. The goal of this project is to develop natural language processing to identify documentation of ACP in the electronic health record of seriously ill patients. A second aim is to study outcomes associated with ACP documentation. There is tremendous interest from the medical community because more efficient identification of ACP offers promise to improve healthcare value. Results from this study will be published in a peer-reviewed journal.
Database: MIMIC is an openly available dataset developed by the MIT Laboratory of Computational Physiology, comprising deidentified health data associated with >40,000 critical care patients, including medical text notes.
Requirements: Knowledge in natural language processing. We primarily use Python programming.
Mentors: Drs. Charlotta Lindvall (Dana-Farber Cancer Institute) and Leo Anthony Celi (MIT Laboratory for Computational Physiology and Beth Israel Deaconess Medical Center)
"Second Generation Physiologic Sensors for Low-Income Countries"
Faculty Advisor: Roger Mark
Mentor(s): Leo Celi
Contact e-mail: lceli@mit.edu
Research Area(s): BioEECS, Materials and Devices, Signals and Systems
Project description: Unexpected cardiac arrest, in which unmonitored patients have cardiac arrest while awaiting care, is a frequent, preventable cause of death in low-resource hospitals. The standard of care, monitoring by nursing staff, is challenging with strained nurse to patient ratios. These unexpected cardiac arrests are a frequent occurrence within overcrowded public hospitals in low-income countries.

Sana and the Makers Laboratory are partnering to create an inexpensive, rugged, wireless monitor in order to detect deterioration among inpatients in understaffed healthcare facilities. The gist is to combine a number of sensors but rather than work on each sensor's performance in comparison to the gold standard, the goal is to use the combined information to detect deterioration. The other functionality that needs to be built in is allowing the sensor to "calibrate" at the start of measurement when the patient is relatively normal from a physiologic point of view.

The device would allow hospital staff to extend limited telemetry resources to patients who would otherwise be unmonitored due to equipment rationing or perceived low risk. Should this be successful, we hope to offer the technical details of this system as a free, open-source, toolkit for low resource emergency departments around the world.
"Website for a speech analysis toolkit"
Faculty Advisor: Stefanie Shattuck-Hufnagel
Mentor(s): Jeung-Yoon Choi
Contact e-mail: sshuf@mit.edu
Research Area(s): Communications, Signals and Systems
This project involves creating a semi-interactive website for an ongoing project on the modeling of human speech perception with implications for automatic speech recognition systems. The website will include an existing tutorial on labelling the linguistically significant information in the speech signal, labelling tools for displaying and annotating the signal and analyzing the results, and several labeled speech databases, along with a prototype speech analysis system and relevant publications and theses. Contact Dr. Stefanie Shattuck-Hufnagel, Speech Communication Group, sshuf@mit.edu.
"Acoustic phonetic variation in speech recognition"
Faculty Advisor: Stefanie Shattuck-Hufnagel
Mentor(s): Jeung-Yoon Choi
Contact e-mail: sshuf@mit.edu
Research Area(s): Communications, Signals and Systems
Words are pronounced very differently in different contexts, and it is challenging to find a vocabulary to describe these differences in enough detail to be useful both for recognizing the words in an automatic speech recognition system and understanding the principles that govern these patterns of variation. This project involves creating an inventory of systematic context-dependent surface variation, determining the factors that govern it, and evaluating hypotheses about the principles that underlie this phenomenon, including massive reductions like Why don’t you → (approximately) wynchah, or I’m going to → (approximately) ahmuhnuh. Candidates with overlapping interests in Course 6, Course 9 and/or Course 24 are particularly appropriate. Contact Dr. Stefanie Shattuck-Hufnagel, Speech Communication Group, sshuf@mit.edu
"High-level processing module for a speech analysis system"
Faculty Advisor: Stefanie Shattuck-Hufnagel
Mentor(s): Jeung-Yoon Choi
Contact e-mail: sshuf@mit.edu
Research Area(s): Communications, Signals and Systems
Current automatic speech recognition systems function usefully, but operate very differently from human speech perception. This UROP involves work on a speech signal analysis system that is modeled more closely on what we know about human speech processing. The project involves developing a consolidator module to integrate acoustic, lexical, and prosodic information derived from the signal into a preliminary hierarchical structure for the entire phrase or utterance, even before the speaker’s intended words are fully recognized. Candidates with overlapping interests in Course 6, Course 9 and/or Course 24 are particularly appropriate. Contact Dr. Stefanie Shattuck-Hufnagel, Speech Communication Group, sshuf@mit.edu.
"Portable Centrifuges"
Faculty Advisor: Jongyoon Han
Mentor(s): Dr. Hyunryul Ryu
Contact e-mail: jyhan@mit.edu
Research Area(s): Applied Physics, BioEECS, Materials and Devices
Many people die from treatable or controllable diseases every year in the third world. One critical bottleneck is the sample preparation and sample transport (out of the endemic region) for biodetection, caused by the lack of basic equipments such as centrifuges. We propose a portable microfluidics system that can automatically perform the function of standard centrifugation on the field automatically. This low-cost platform will reliably isolate leukocyte, bacteria, virus, and other serum factors from the patients' bodily fluid. The safely detoxicated and purified samples can be sent to advanced labs with modern diagnostic techniques, to enable rapid treatment and countermeasures. Students will will working with researchers in the Han group to engineer and build the portable system, by work on hardware (pumping, microfluidics, circuits) and control system programming. They also will carry out some testing to increase usability and reliability. Collaboration with researchers working on infectious diseases are already in place to increase the impact of the students' work. Portable Centrifuges
"Portable Seawater Desalination System"
Faculty Advisor: Jongyoon Han
Mentor(s):
Contact e-mail: jyhan@mit.edu
Research Area(s): Applied Physics, Control, Energy, Materials and Devices
A small scale, portable and on-site water desalination system may be a solution to meet the ever-increasing and fluctuating demand for water for many uses, including human consumption and agriculture. In this project, we will develop a small scale water purification desalination system that can meet the economic and technical specification for rural and remote areas' water needs. Electrical desalination techniques such as electrodialysis, ICP desalination, and electrocoagulation will be integrated into a small package, in order to provide a demonstration of small flow rate (10-100mL/min) operation. Various applications such as camping, disaster relief, and military applications can be served by such a device. Students will be participating in the design of a small-scale system (with the help of other lab members) and work on different aspect of the integration, potentially including control, assembly, and testing with actual field samples. We will also have an opportunity to learn how to carry out techno-economic analysis of such a system. Portable Seawater Desalination System
"3D electronics based on III-V semiconductors"
Faculty Advisor: Jesus A. del Alamo
Mentor(s): Alon Vardi
Contact e-mail: alonva@mit.edu
Research Area(s): Applied Physics, Materials and Devices, Nanotechnology
3D transistors based on III-V compound semiconductors are considered promising candidates for future logic applications. Due to the low effective mass in these materials, quantum size effects become prominent already at 10 nm dimensions. With the advanced technology in our group, this dimensional range is within reach and we have fully developed Fin-Field-Effect Transistor and Nanowire Transistor processes with sub-10 nm dimensions. To explore the transport properties in these extraordinary nano devices, we have developed nano transmission line model (TLM) structures and revolutionary nano Hall devices in which a nano-electromagnet is embedded into them. This device, by its own right may lead to breakthroughs in fast electronics. The scope of the project is wide and may include simulations, electrical measurements and process development. If you want to push at the frontiers of electronic devices and go where no one has gone before, sign up today with the Xtreme Transistor Group. 3D electronics based on III-V semiconductors
"Mobile apps for education"
Faculty Advisor: Hal Abelson
Mentor(s):
Contact e-mail: hal@mit.edu
Research Area(s): Computer Systems
MIT App Inventor is a Web-based development environment that lets anyone build mobile apps for Android phones and tablets -- even kids who have never programmed before. Our group in CSAIL and the Media Lab runs App Inventor as a worldwide service, currently with 3 million users. The system is used by hobbyists, teachers, and school kids, and there are several curriculum project that use App Inventor and mobile app development as a basis for teaching computing.

There are several projects available:


* Graphing/data visualization
Student would be in charge of designing and developing data visualization components for App Inventor, which would help students visualize collected data in mobile apps. Students should have familiarity with Java, JavaScript, CSS, and HTML. Knowledge of charting libraries, such as c3.js, is a plus. By the end of the project, the student will create a production-ready component for App Inventor users to visualize data in mobile apps.

*Text to block
App Inventor uses a blocks based programming language to make it easier for users to create Android applications. However, there are some tasks that are easier to do using text, including searching and writing mathematical expressions. For this project, students will begin to improve the user experience of creating apps with our text based language by adding features like autocomplete, error checking, and internationalization. Experience with Ace editor, acorn.js, or programming language grammars is a plus.

*Tabs in blocks editor for decluttering and organizing
In App Inventor, users snap together blocks to control their Android app. As their apps become larger and more complicated, they need ways to be able to organize their code. While in text based languages, programmers put their code in separate files, methods of structuring code in blocks languages has not been standardized. Your project will be to create ways for users separate their blocks into multiple pages, and provide guidance for how they can create cleaner and clearer code in a blocks language. Experience with Javascript and svg is a plus.

*Dictionaries and complex data structures
Student would be in charge of implementing support for object-oriented class objects in App Inventor's blocks editors, which would help users of App Inventor learn about data structures. Students should have familiarity with Java, JavaScript, and Scheme. Previous experience working with Google's Blockly is a plus. By the end of the project, the student will produce and merge production-ready extensions to the App Inventor blocks editor and companion app to enable App Inventor users to design and leverage custom data structures.

*Virtual Reality in App Inventor
App Inventor wants to be able to support Google Cardboard applications, however this requires changing the way that users build projects to support 3d environments. We’ve begun developing an editor for creating VR experiences, but there is a lot more to do. You would help build on our current progress, test to see what difficulties students have when creating VR experiences, and modify the editor based on their feedback. Experience with three.js or other 3d graphics library is a plus.

*CloudDB data + log history
CloudDB is an experimental component that allows app developers to share data in their apps as tag/value pairs. This is currently implemented using Redis through App Inventor. The SuperUROP student would be in charge of creating a webpage that app developers can use to view their CloudDB data buckets. For example, the webpage could show the list of tags and values for each of the developer’s buckets as well as a last changed timestamp and a brief history. The SuperUROP student will work closely with a PhD student and will have some autonomy in UI/UX design. Students should have a strong background in Java, creating websites, and working with databases. Experience with Redis is preferred but can be easily learned. By the end of the project, we expect the student’s contribution to go live on the App Inventor tool.


For more information see appinventor.mit.edu or send mail to appinventor-urop@mit.edu.
Mobile apps for education
"Information Policy"
Faculty Advisor: Hal Abelson
Mentor(s):
Contact e-mail: hal@mit.edu
Research Area(s):
The Internet is now the central nervous system of our global economy and the essential infrastructure for communication, commerce, and civic discourse. Yet at this transformative moment, many important public debates concerning information policy occur without adequate technical understanding and scholarship. The new MIT Information Policy Project seeks to fill this gap with technically-informed research and politically-engaged dialogue, aimed at guiding Internet policymakers around the world.

Based within MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL), the Information Policy Project builds on MIT’s proven approach to engineering research and education. We view policy development, like technology development, as a design discipline that should be driven by methodical study, and we aims to train a new generation of technology policy leaders in government, civil society, academia and industry.

In its first three years, the Project will tackle research challenges such as:

Privacy and surveillance, including accountable systems and their potential; mobile apps privacy; civil liberties in the age of big data; EU-US privacy agreements; and surveillance and national security.

Network architecture, including the evolution of wireless and wireline access; the economics and flow of payments across network layers; and alternate service provision models.

Internet governance, including assessments of existing global mechanisms and the growth of policy expertise within the global internet community.

Other research priorities will include cybersecurity, online copyright protection, Internet content regulation, and intermediary liability worldwide.

There are several SuperUROP projects for students who want to be part of this new effort. It would be good have some familiarity with the material covered in 6.805/STS085 and also 6.033, but these are not necessarily prerequisites.

For additional information, contact Taylor Reynolds <treyn@mit.edu> or Danny Weitzner <djweitzner@mit.edu>
Information Policy
"Assistive Devices for Healthcare"
Faculty Advisor: Dorothy Curtis
Mentor(s): Prof Pino & Prof Aqueveque, University of Concepcion
Contact e-mail: dcurtis@csail.mit.edu
Research Area(s): Computer Systems, Graphics and Human-Computer Interfaces, Signals and Systems
This project involves analyzing data collected by a non-invasive monitoring and alerting system for people with advanced Multiple Sclerosis (MS). Several sensors were deployed on an electric wheelchair acquiring vital signs, activity level and ambient information.

At this time, we are looking for assistance from students in the analysis of the 260 hours of ballistocardiogram (for heart rate) and accelerometer data (for activity level) that has been collected by the monitoring platform. from MS volunteers at The Boston Home (TBH).
Assistive Devices for Healthcare
"Re-Imagining the Collection and Archiving of Alumnae Oral Histories"
Faculty Advisor: Dorothy Curtis
Mentor(s):
Contact e-mail: dcurtis@csail.mit.edu
Research Area(s): Computer Systems, Graphics and Human-Computer Interfaces
As recently as 1966, when about 10 women graduated from MIT each year, the AMITA Margaret MacVicar Oral History project was established, so that students could learn from these pioneering women. Now, many more women graduate from MIT each year (between 400 and 800) and the process of collection of their histories needs to be revisited.

The goal of this project is to review existing oral history collection apps and understand the process that AMITA MacVicar Oral Histories Project uses to collect MIT alumnae oral histories. It is expected that an app to collect oral histories as well as an archival system will need to be developed.

This project is appropriate for CS-HASS funding.
Re-Imagining the Collection and Archiving of Alumnae Oral Histories

Total: 91

Technical questions related to this site? Please contact Search-Master@superurop.eecs.mit.edu