FFSE-2018 Talks
The 1st Forum on Frontiers of Science and Engineering
Everything towards AI
Seattle, WA, USA, May 28 - 30, 2018
Featured Talks
Natural Language Processing R&D in Alibaba
Natural Language Processing (NLP) and related technologies are critical for the success of Internet business like e-commerce. Alibaba’s NLP R&D aims at supporting the business demands of Alibaba’s eco-system, creating new opportunities for Alibaba’s partners and advancing the state-of-the-art of NLP technologies. This talk will introduce our efforts to build NLP technique platform and machine translation (MT) platform that power Alibaba’s eco-system. Furthermore, some recent research work will be presented on product title compression with user-log information, sentiment classification with questions & answers, machine reading comprehension in real-world custom service, and cascade ranking for large-scale e-commerce search. The R&D work attracts hundreds of millions of users and generates significant business value every day.
Luo Si is the Chief Scientist of Natural Language Processing with Alibaba DAMO institute Machine Intelligence Technologies. He leads a cross-country team in China, USA and Singapore with the focus on developing cutting edge technologies in natural language processing, machine translation, text mining and information retrieval. The work attracts hundreds of millions of users and generates millions of revenue each day. Luo has published more than 150 journal and conference papers with substantial citations. His research has obtained many industry awards from Yahoo!, Google and Alibaba as well as NSF career award. Prior to joining Alibaba in 2014, he was an Associate Professor with Purdue University. He obtained degrees in computer science from Tsinghua University and Carnegie Mellon University.
Selected Topics in Deep Learning for Medical Image Analysis
To identify potential collaboration opportunities between TAAC members, this talk gives a brief review of Wang group’s current research efforts at UBC in the areas of Deep Learning for Medical Image Analysis. During the last few years, deep learning, in particular convolutional networks, has rapidly become popular for medical image analysis, including classification, object detection, segmentation, registration, and other tasks. This overview talk contains three major parts: (1) Skin image analysis, with focus on developing automated detection, segmentation, pattern recognition and computer-aided diagnosis based on skin vasculatures. Vascular structures seen in dermoscopy are of great clinical importance, playing an important role in skin cancer diagnosis, lesion growth and treatment monitoring. (2) 2D/3D medical image registration, with focus on accurate 2D/3D registration of pre-operative 3-D data and intra-operative 2-D X-ray images (which is a key enabler for image-guided therapy). We exploit the modeling power of Convolutional Neural Networks to significantly improve registration accuracy and efficiency. We further propose a Pairwise Domain Adaptation (PDA) module to adapt the model trained on source domain (i.e., synthetic data) to target domain (i.e., clinical data) by learning domain invariant features with a few paired real and synthetic data. (3) MRI data analytics, with focus on 3D CNN based automatic diagnosis of Attention Deficit Hyperactivity Disorder using functional and structural MRI data.
Z. Jane Wang received the B.Sc. degree from Tsinghua University in 1996 and the M.Sc. and Ph.D. degrees from the University of Connecticut in 2000 and 2002, respectively, all in electrical engineering. She has been Research Associate at the University of Maryland, College Park. Since Aug. 2004 she has been with the ECE dept. at UBC, Canada, and she is currently Professor. She is an IEEE Fellow and a member of the College of New Scholars, Artists and Scientists of the Royal Society of Canada. Her research interests are in the broad areas of statistical signal processing and machine learning. She has published 110+ journal papers and 100+ peer-reviewed conference papers. She has been key Organizing Committee Member for numerous IEEE conferences and workshops (e.g., the Finance chair of ICASSP13, the co-Technical Chair for ChinaSIP2014, GlobalSIP2017 and ICIP2021, and the co-General Chair of MMSP2018). She has been Elected Member to several SPS TCs, including the Bio Imaging and Signal Processing TC, Multimedia SP TC, Machine Learning SP TC, and IFS TC. She was Associate Editor for the IEEE TSP, SPL, TMM, TIFS and TBME, and Area Editor of SPM. She is currently serving as Associate Editor for SPM and IEEE Transactions on Circuits and Systems-II: Express Briefs.
The Challenges of Running Neural Networks on Edge and Mobile Devices
Fast growth of the computation cost associated with training and testing of deep neural networks (DNNs) inspired various acceleration techniques for edge and mobile applications. Reducing topological complexity and simplifying data representation of neural networks are two approaches that popularly adopted in deep learning society: many connections in DNNs can be pruned and the precision of synaptic weights can be reduced, respectively, incurring no or minimum impact on inference accuracy. However, the practical impacts of hardware design are often ignored in these algorithm-level techniques, such as the increase of the random accesses to memory hierarchy and the constraints of memory capacity. On the other side, the limited understanding about the computational needs at algorithm level may lead to unrealistic assumptions during the hardware designs. In this talk, we will discuss this mismatch and show how we can solve it through an interactive design practice across both software and hardware levels.
Yiran Chen received B.S and M.S. from Tsinghua University and Ph.D. from Purdue University in 2005. After five years in industry, he joined University of Pittsburgh in 2010 as Assistant Professor and then promoted to Associate Professor with tenure in 2014, held Bicentennial Alumni Faculty Fellow. He now is a tenured Associate Professor of the Department of Electrical and Computer Engineering at Duke University and serving as the co-director of Duke Center for Evolutionary Intelligence (CEI), focusing on the research of new memory and storage systems, machine learning and neuromorphic computing, and mobile computing systems. Dr. Chen has published one book and more than 300 technical publications and has been granted 93 US patents. He serves or served the associate editor of IEEE TNNLS, IEEE TCAD, IEEE D&T, IEEE ESL, ACM JETC, ACM TCPS, ACM SIGDA E-News and served on the technical and organization committees of more than 40 international conferences. He received 6 best paper awards and 12 best paper nominations from international conferences. He is the recipient of NSF CAREER award and ACM SIGDA outstanding new faculty award. He is a Fellow of IEEE.
Statistical Thinking for AI: Keeping Human in the Loop
Statistics plays a central role in the data science approach. In almost all data science solutions, data scientists need to exercise statistical thinking, in designing data collection, deriving insights from visualizing data, obtaining supporting evidence for data-based decisions and constructing models for predicting future trends from data. Recently developed artificial intelligence methods have shown to possess the power to identify patterns, extract knowledge, and generate prediction from complex data. Machine learning algorithms have been used to automate human tasks and support decision making. However, we soon realize that AI algorithms often produce biased and misleading results. In this talk, I will discuss the roots for these biases and present how statistical thinking can be used to introduce human wisdom and values into an AI workflow.
Tian Zheng grew up on the campus of Tsinghua University. She attended Tsinghua from 1994-1998 in the Department of Applied Statistics and obtained her PhD in Statistics from Columbia University in 2002. She is currently Professor of Statistics and Associate Director for Education of the Data Science Institute at Columbia University. She develops novel methods for exploring and understanding patterns in complex data from different application domains such as biology, psychology, climatology, and etc. Her current projects are in the fields of statistical machine learning, spatiotemporal modeling and social network analysis. Professor Zheng’s research has been recognized by the 2008 Outstanding Statistical Application Award from the American Statistical Association (ASA), the Mitchell Prize from ISBA and a Google faculty research award. She became a Fellow of American Statistical Association in 2014. Professor Zheng is the receipt of 2017 Columbia’s Presidential Award for Outstanding Teaching. In 2018, she was elected as the chair-elect for ASA’s section on Statistical Learning and Data Science. Professor Zheng was an associate editor for Journal of American Statistical Association - Applications and Case Studies from 2007 to 2013 and a current AE for Statistical analysis and data mining (SAM) and Statistics in Biosciences (SIBS), also a Faculty member of F1000 Prime. She is on the advisory board for STATS at Sense About Science America that targets to develop a statistical literate citizenry.