Speakers
Think Globally, Act Locally: A Deep Neural Network Approach to High-Dimensional Time Series Forecasting
Forecasting high-dimensional time series plays a crucial role in many applications such as demand forecasting and financial predictions. Modern datasets can have millions of correlated time-series that evolve together, i.e they are extremely high dimensional (one dimension for each individual time-series). There is a need for exploiting global patterns and coupling them with local calibration for better prediction. However, most recent deep learning approaches in the literature are one-dimensional, i.e, even though they are trained on the whole dataset, during prediction, the future forecast for a single dimension mainly depends on past values from the same dimension. In this paper, we seek to correct this deficiency and propose DeepGLO, a deep forecasting model which thinks globally and acts locally. In particular, DeepGLO is a hybrid model that combines a global matrix factorization model regularized by a temporal convolution network, along with another temporal network that can capture local properties of each time-series and associated covariates. Our model can be trained effectively on high-dimensional but diverse time series, where different time series can have vastly different scales, without a priori normalization or rescaling. Empirical results demonstrate that DeepGLO can outperform state-of-the-art approaches; for example, we see more than 25% improvement in WAPE over other methods on a public dataset that contains more than 100K-dimensional time series. This is joint work with Rajat Sen and Hsiang-Fu Yu.
BioInderjit Dhillon is the Gottesman Family Centennial Professor of Computer Science and Mathematics at UT Austin, where he is also the Director of the ICES Center for Big Data Analytics. Currently he is on leave from UT Austin and heads the Amazon Research Lab in Berkeley, California, where he is developing and deploying state-of-the-art machine learning methods for Amazon Search. His main research interests are in big data, machine learning, network analysis, linear algebra and optimization. He received his B.Tech. degree from IIT Bombay, and Ph.D. from UC Berkeley. Inderjit has received the following awards: the ICES Distinguished Research Award, the SIAM Outstanding Paper Prize, the Moncrief Grand Challenge Award, the SIAM Linear Algebra Prize, the University Research Excellence Award, and the NSF Career Award. He has published over 175 journal and conference papers, and has served on the Editorial Board of the Journal of Machine Learning Research, the IEEE Transactions of Pattern Analysis and Machine Intelligence, Foundations and Trends in Machine Learning and the SIAM Journal for Matrix Analysis and Applications. Inderjit is actively involved with industry. He is currently an Amazon Fellow at A9/Amazon, where he is developing and deploying state-of-the-art machine learning methods in Amazon search. Prior to joining Amazon, Inderjit worked as a quantitative analyst at a hedge fund, Voleon, which does systematic machine learning statistical arbitrage. In the past, Inderjit has been a consultant for Walmart Labs, Sabre Inc, Yahoo!, Syncata, and Neonyoyo. Inderjit is an ACM Fellow, an IEEE Fellow, a SIAM Fellow and an AAAS Fellow.
Reformer: The Efficient Transformer
Transformer models have been used in a variety of fields and yield great results on many NLP tasks. But between the BERT, GPT-2, and many other variants, they can be inefficient and it can be hard to apply them. I will introduce a new efficient variant of Transformer called the Reformer. I'll take you through the code that implements it and I will show how it runs at high efficiency and addresses the main problems or high memory use and low performance on long sequences that limited the use of some Transformers before. I will finish with new applications of Reformer that open up.
Bio
Łukasz Kaiser joined Google in 2013 and is currently a staff Research Scientist in the Google Brain Team in Mountain View, where he works on fundamental aspects of deep learning and natural language processing. He has co-designed state-of-the-art neural models for machine translation, parsing and other algorithmic and generative tasks and co-authored the TensorFlow system, the Tensor2Tensor and Trax libraries and the Transformer model. Before joining Google, Lukasz was a tenured researcher at University Paris Diderot and worked on logic and automata theory. He received his PhD from RWTH Aachen University in 2008 and his MSc from the University of Wroclaw, Poland.
Machine Learning for Healthcare in the COVID-19 Era
Bio
Mihaela van der Schaar is the John Humphrey Plummer Professor of Machine Learning, Artificial Intelligence and Medicine at the University of Cambridge, a Fellow at The Alan Turing Institute in London, and a Chancellor’s Professor at UCLA. Mihaela was elected IEEE Fellow in 2009. She has received numerous awards, including the Oon Prize on Preventative Medicine from the University of Cambridge (2018), a National Science Foundation CAREER Award (2004), 3 IBM Faculty Awards, the IBM Exploratory Stream Analytics Innovation Award, the Philips Make a Difference Award and several best paper awards, including the IEEE Darlington Award. Mihaela’s work has also led to 35 USA patents (many widely cited and adopted in standards) and 45+ contributions to international standards for which she received 3 International ISO (International Organization for Standardization) Awards. In 2019, she was identified by National Endowment for Science, Technology and the Arts as the most-cited female AI researcher in the UK. She was also elected as a 2019 “Star in Computer Networking and Communications” by N²Women. Her research expertise span signal and image processing, communication networks, network science, multimedia, game theory, distributed systems, machine learning and AI. Mihaela’s current research focus is on machine learning, AI and operations research for healthcare and medicine.
Machine Learning for Aviation Safety
The National Airspace, with all of its aircraft, airports, personnel, and related infrastructure, is an incredibly safe system. It is essential that we keep it safe through the changes that it is experiencing, including expected increases in traffic over time, increasing variety of traffic in the form of Unmanned Aerial Vehicles (UAVs), and sharp reductions and increases in traffic due to transient phenomena such as pandemics. The system is currently monitored for safety issues through a set of exceedances, which are rules describing various known safety issues. By definition, these rules cannot identify previously-unknown safety issues. Additionally, they do not identify precursors to these safety issues—states that may not represent safety issues by themselves, but are circumstances under which safety issues are more likely to occur in the near future. In this talk, I describe machine learning-based methods that we have developed for anomaly detection and precursor identification and the aviation safety results that we have obtained. I also describe the active learning algorithm that we have developed to mitigate the false alarm problem that is common to data-driven anomaly detection methods. Our ultimate aim is to allow for aviation safety analysts to discover new safety issues as they arise.
Bio
Nikunj Oza is the leader of the Data Sciences Group at NASA Ames Research Center. He also leads a NASA project team, which applies machine learning to aviation safety and operations problems. Dr. Oza’s 50+ research papers represent his research interests, which include data mining, machine learning, ensemble learning, anomaly detection, and their applications to Aeronautics and Earth Science. He received the Arch T. Colwell Award for co-authoring one of the five most innovative technical papers selected from 3300+ SAE technical papers in 2005. His data mining team received the 2018 and 2019 NASA Honor Awards and the 2010 NASA Aeronautics Research Mission Directorate Associate Administrator¹s Award. In 2019, he was named by Cognilytica as one of 50 key people in the US government working to move the adoption of AI forward across the industry. He is an Associate Editor for the peer-reviewed journal Information Fusion (Elsevier) and has served as organizer, senior program committee member, and program committee member of several data mining and machine learning conferences. He received his B.S. in Mathematics with Computer Science from MIT in 1994, and M.S. (in 1998) and Ph.D. (in 2001) in Computer Science from the University of California at Berkeley.
