Abstract
Large-scale machine learning requires blending computational thinking with statistical frameworks. Designing fast, efficient and distributed learning algorithms with statistical guarantees is an outstanding grand challenge. In this lecture, the speaker will present perspectives from theory and practice. She will demonstrate how spectral optimization can reach the globally optimal solution for many learning problems despite being non-convex. This includes unsupervised learning of latent variable models, training neural networks and reinforcement learning of partially observable Markov decision processes. In practice, tensor methods yield enormous gains both in running times and learning accuracy over traditional methods such as variational inference.
The speaker will then talk about the recent advances in large-scale deep learning methods. Her team at Amazon Web Services is actively innovating on the MXNet package. It is a highly flexible and developer-friendly open-source deep learning framework designed for both efficiency and flexibility. It is based on the distributed parameter-server framework. She will also demonstrate how to use preconfigured Deep Learning AMIs and Cloud Formation Templates on Amazon Web Services to help speed up deep learning research and development.
The speaker will conclude on outstanding challenges on how the scientists can bridge the gaps between theory and practice, and how she can design and analyze large-scale learning algorithms.
About the speaker
Prof Anima Anandkumar received her BTech in Electrical Engineering from Indian Institute of Technology Madras in 2004 and PhD from Cornell University in 2009. She was a postdoctoral researcher at Massachusetts Institute of Technology from 2009 to 2010, an assistant professor at University of California at Irvine between 2010 and 2016, and a visiting researcher at Microsoft Research New England in 2012 and 2014. She is currently a principal scientist at Amazon Web Services and will be joining the Department of Computing and Mathematical Sciences of California Institute of Technology in summer 2017 as a Bren endowed chair.
Prof Anandkumar’s research interests are in the areas of large-scale machine learning, non-convex optimization and high-dimensional statistics. In particular, she has been spearheading the development and analysis of tensor algorithms.
Prof Anandkumar received numerous of awards including the Alfred P. Sloan Fellowship, Microsoft Faculty Fellowship, Google research award, ARO and AFOSR Young Investigator Awards, NSF Career Award, Best Thesis Award from the ACM Sigmetrics society, IBM Fran Allen PhD fellowship, and several best paper awards. She has been featured in a number of forums such as the Quora Machine Learning Session, Huffington Post, Forbes, and O’Reilly Media.
|
