²⁰¹²Sarah Cannon

∫ Female Awardee ∫

Senior at Tufts University

Sarah Cannon is a Senior at Tufts University majoring in Mathematics with a Computer Science minor.
Sarah has contributed to an impressive variety of research projects, working with a broad range of researchers. Most recently, she has been looking into spontaneous self-assembly of nanoscale devices. The process can be viewed as a tiling problem with constraints, but several different abstract models have been proposed. Sarah has proved new results comparing these models and their ability to simulate one another.

In separate work, she has devised algorithms and complexity results for the problem of orienting undirected graphs, and made contributions to minimal surface theory.

Sarah works as an academic tutor and taught a robotics workshop for kindergarten and elementary school students. She is also a dedicated field hockey player, and has been recognized by the NCAA as an outstanding scholar-athlete.

²⁰¹²Justin Cheng

∫ Male Awardee ∫

Senior at Cornell University

Justin Cheng is a Senior at Cornell University majoring in Computer Science.

Social networks support both one-way communication (e.g., broadcasts by a celebrity) and two-way communication (e.g., discussions). Justin has analyzed communication patterns to automatically classify relationships as reciprocal or non-reciprocal, paving the way for improved recommendation systems.

In other work, Justin has investigated ways that computer interfaces can motivate behavior, leading to papers on fun interfaces that encourage users to provide tags for images, and contemplative interfaces that promote slowness and self-reflection.

As a teaching assistant, Justin has mentored students from a large range of backgrounds.

²⁰¹²Karthik Narayan

∫ Male Awardee ∫

Senior at Georgia Institute of Technology

Karthik Narayan is a Senior at the Georgia Institute of Technology majoring in Computer Science and Discrete Mathematics.

Games and other modern applications require computers to create natural language text dynamically. One method is to combine pre-constructed templates, but when the collection of possible templates becomes large, managing their interactions can be difficult. Karthik has applied machine-learning techniques to templates, using controlled user studies to show that his work reduces the effort required to construct and maintain large collections.

Between this and other research projects, including algorithms for multi-agent reinforcement learning, Karthik has co-authored four publications as an undergraduate.

Beyond his own work, Karthik has helped freshman and sophomore students get involved in computer science research at Georgia Tech.

²⁰¹²Sarah Chasins

∫ Female Runner-Up ∫

Senior at Swarthmore College

Sarah Chasins is a Senior at Swarthmore College majoring in Computer Science and Behavioral Economics.

Although we expect programs to run quickly, novel language features can make efficient implementations difficult. Sarah designed and produced a new way of compiling the Plaid language, whose support for typestate abstractions permits objects to change their fields and methods on the fly, in well-structured ways. Compared to a previous Java-based implementation, Sarah’s JavaScript-based implementation produces code that is 47 times faster.

In an earlier project, Sarah worked on using embedded sensors in smartphones to draw conclusions about the user and the environment.

Sarah works as a teaching assistant and, as a head TA, trains new teaching assistants in the CS Department.

²⁰¹²Lavanya Jose

∫ Female Runner-Up ∫

Senior at Princeton University

Lavanya Jose is a Senior at Princeton University majoring in Computer Science.

Lavanya’s research has focused on efficient online algorithms for measuring Internet traffic, studying the tradeoffs between accuracy and overhead for problems such as finding Hierarchical Heavy Hitters. Based on this work she developed OpenSketch, a configurable three-stage pipeline designed to run on network switches; it permits many different building blocks that can be combined to do interesting traffic-measurement tasks.

As a separate project, Lavanya has worked to improve WordNet, a lexical database of English widely used in computational linguistics and natural language processing.

²⁰¹²Peter Chapman

∫ Male Runner-Up ∫

Senior at University of Virginia

Peter Chapman is a Senior at the University of Virginia majoring in Computer Science and Cognitive Science.

Computer security and privacy is a critical concern, especially when medical issues are involved. Peter developed a method for automatically searching web applications to find side-channel vulnerabilities in web applications. He applied new statistical tools to better describe these vulnerabilities. In the end, he determined that 88% of queries to Google Health could be recovered by an eavesdropping adversary.

Peter has also worked on secure computation, where parties collaborate on computing a function of two inputs without exposing the inputs to each other. He has proposed novel applications of secure computation in smartphones, and is working on an improved approach to mobile secure computation, relying on the network carrier to provide suitable streams of randomness.

²⁰¹²Svilen Kanev

∫ Male Runner-Up ∫

Senior at Harvard University

Svilen Kanev is a Senior at Harvard University majoring in Computer Science.

Energy-efficient hardware can produce unwanted power fluctuations due to subsystems powering down and powering up. Svilen showed that this voltage noise can be mitigated in multicore systems by intelligently scheduling threads and by controlling compiler optimization levels. A very strong experimentalist, he has also built detailed performance simulations of the Atom CPU microarchitecture, including power modeling, and validated them against real hardware.

Additionally, Svilen is the lead contributor to the AI system used in competition by the Harvard-MIT Robot Soccer Club, working on everything from the high-level strategy language down to low-level embedded controls.