Ten from CIS, engineering faculty win Google research awards
By Melanie Lefkowitz
Ten Cornell faculty members in computer science and engineering have received Google Faculty Research Awards, supporting projects that range from using machine learning to fight neurological diseases to building a database of common errors to help automate fact-checking.
The Google Faculty Research Awards Program’s goal is to recognize cutting-edge research in mutual areas of interest and to “identify and strengthen long-term collaborative relations with faculty working on problems that will impact how future generations use technology,” according to its website.
The selected faculty will receive unrestricted funding to support one graduate student for a year. Cornell has the third-highest number of recipients among the 80 institutions worldwide that received Google awards. The Massachusetts Institute of Technology topped the list, with 13 awards.
“I’m not at all surprised that Cornell emerged as one of the top universities receiving so many awards from Google,” said Greg Morrisett, dean of Computing and Information Science. “It’s because we have the best faculty in the world working on a huge range of topics – from how to engineer, secure and manage data centers more effectively to developing cutting-edge algorithms and applications for machine learning.”
The winners and their projects are:
- Jayadev Acharya, assistant professor of electrical and computer engineering: “Private Heavy Hitters – Simplest and Optimal.” While user data helps design better systems, personal and potentially sensitive data need to remain private. One particular task that Acharya plans to track is heavy-hitter detection, such as when Google receives various search requests from users – or spammers – all occurring at the same time.
- Yoav Artzi, assistant professor of computer science, Cornell Tech; and Daniel Lee, professor of electrical and computer engineering, Cornell Tech: “Joint Learning of Continuous Natural Language Control for Quadcopters in Simulation and Reality.” This project proposes learning a direct mapping of natural language instructions and raw observations for the continuous control of a drone.
- Claire Cardie, professor of computer and information science: “iFashionist Knowledge Graph: a Domain-Specific Knowledge Base and Automatic Taxonomy Construction Method for Fashion.” Cardie is working with fashion experts to develop better methods to represent and search online for garments and accessories. Her project will investigate methods in natural language processing to build a knowledge base about fashion by analyzing and extracting information from online catalogs.
- Christina Delimitrou, assistant professor of electrical and computer engineering: “Using Machine Learning to Improve Datacenter Server Management.” Delimitrou’s proposal aims to use practical machine learning techniques to address the increasing complexity in the hardware and software of warehouse-scale computers.
- Thorsten Joachims, professor of computer and information science: “Counterfactual Learning With Estimated Propensities.” Joachims’ proposed research deals with the uncertainty of whether users clicked on a search engine result because it was the most relevant or because they didn’t notice a more relevant result further down. His project aims to estimate how much the true relevance signal is distorted by the result’s position, and then develop new machine learning methods to combat the bias.
- Mahsa Shoaran, assistant professor of electrical and computer engineering: “Toward Minimally-Invasive Brain Implants With Embedded Classification of Real-time Neural Data.”Shoaran’s proposal aims to employ state-of-the-art machine learning principles and efficient hardware architectures to enable real-time neural data processing on devices and prostheses for neurological diseases.
- Vitaly Shmatikov, professor of computer science, Cornell Tech: “Detecting and Mitigating Unwanted Learning.”Shmatikov’s research focuses on the problem of “unwanted learning” – when machine learning models learn more than intended, such as sensitive information about their training data, privacy-violating tasks without their creators’ knowledge, or even malicious functionality.
- Immanuel Trummer, assistant professor of computer science: “Mining an ‘Anti-Knowledge Base’ for Fact Checking from Wikipedia Updates.” Trummer’s goal is to create an “anti-knowledge base,” containing common factual mistakes and controversial claims, that can be used for automated fact checking through language analysis and unsupervised learning.
- Zhiru Zhang, associate professor of electrical and computer engineering: “Automatic Synthesis for Programmable Hardware Specialization.” Zhang proposes to develop a new compilation framework that can automatically synthesize a high-quality programmable hardware accelerator from instruction set specifications.