Three CEAS Faculty Receive Google Research Awards
From teaching machines to think autonomously and customizing touch screen technology, to mapping out floor plans for indoor location services, the recipients of this year's Google Research Awards from Stony Brook University's College of Engineering and Applied Sciences are driving the cutting-edge research that will impact the world for years to come.
Each year, Google announces an open call for PhD students and faculty at accredited Universities to submit proposals on computer science related topics. Those chosen gain the opportunity to work hand-in-hand with Google researchers and engineers to develop their research.
This year, three faculty members in the College of Engineering and Applied Sciences have been awarded, highlighting the College's rising competitive presence among other top engineering universities on Google’s 2017 winner list, which include MIT, Stanford University, and University of California - Berkeley.
Hailing from the Departments of Computer Science and Electrical and Computer Engineering, the 2017 Google Faculty Research Award recipients include Professors Xiaojun Bi, Francesco Orabona, and Fan Ye.
“At Stony Brook, we know our faculty drives innovation not only in the classroom but in the real world as well. Receiving three Google Research Awards is a demonstration of the truly transformative work being done by our outstanding faculty,” said Stony Brook University President Samuel L. Stanley. “ I congratulate Professors Bi, Orabona and Ye on this award, and thank them for their contributions to the Stony Brook community.”
"I extend my congratulations to Xiaojun Bi, Francesco Orabona, and Fan Ye for receiving 2017 Google Faculty Research Awards. We are immensely proud of the elite company that Stony Brook now joins as one of the few universities to have multiple recipients of this award,” said Michael A. Bernstein, Provost and Senior Vice President for Academic Affairs at Stony Brook University. “ We are thrilled that the innovative research of these fine colleagues has been recognized with this esteemed honor."
Xiaojun Bi , Computer Science : to develop a model-based approach that will address user interface issues caused by the imprecision of current touch screen technology.
Finger touch screens are the dominant input modality for most common Post-PC computing devices such as smartphones, tablets, or smart watches. Because of the difficulty in creating a user interface (UI) device that is compatible with all finger shapes and sizes, selecting target applications with finger touch is still extremely error-prone compared to a mouse pointer.
This problem has left Assistant Professor Xiaojun Bi wondering, “Can we personalize UIs for each user?” Google thought this question was worth answering, so with a grant of $38,500, they are helping him find out.
Xiaojun Bi’s team will use a model-based approach to address the UI design issues caused by the imprecision of touch. They will derive models that predict the accuracy of touch by using the basic human motor control theories, and then apply them to guide UI design, to optimize and personalize UIs.
“ We are stepping into an era where touch is one of the most important input modalities,” Bi explained. “Our research will benefit millions of users who use smartphones, tablets and other touchscreen devices.”
Francesco Orabona , Computer Science: to design parameter-free, automatic machine learning algorithms.
Machine Learning (ML) algorithms are tools to learn rules and prediction strategies automatically from data. The problem is that no existing ML algorithms work automatically, because in order to realize positive results, they require a human to set parameters.
Assistant Professor Francesco Orabona wants to know, “Can we do better?” and Google wants to help him find out. He believes there is no reason why ML algorithms can not be truly automatic, and in previous work he proved that they can be when applied to simple ML algorithms. In his proposal to Google, he suggested going beyond these simple applications to find a way to Deep Learning algorithms (the latest tools in ML) parameter-free too.
“There is nothing without ML these days,” he explained, citing examples such as intelligent thermostats and cameras that detect smiles in photos. “There has been a lot of press about artificial intelligence lately, but what they actually mean is just ML. And it’s ML that will be the protagonist of the next industrial revolution.”
Fan Ye , Electrical and Computer Engineering , to investigate a low cost, scalable solution for mapping indoor floor plans, to serve as the foundation for indoor location based services.
Indoor floor plans are the foundational data that enable location based services (LBS) such as navigation, however obtaining this data traditionally consists of large scale business negotiations and often expensive gear for people to carry and walk inside buildings with.
Assistant Professor Fan Ye believes that instead, by leveraging smartphone speakers and microphones, he can measure the distance of major surfaces in complex buildings, such as walls, and then eventually construct complete, accurate floor maps. His proposed solution to leverage already existing commodity smartphone technology, eliminates the logistic and financial costs of business negotiation and specialized hardware. Google agreed, his idea sounds promising.
“Indoor maps are the foundation for indoor LBS which will have huge business impacts,” Fan Ye explained. “Navigation to get to a specific room for instance, location based ads, essentially all things that you use with outdoor maps, should have indoor maps for counterparts.”