speakers
Tel Aviv University
TBD
TBD
Nadav Cohen is an Assoc. Professor of Computer Science at Tel Aviv University. His research focuses on the theoretical and algorithmic foundations of neural networks (a.k.a. deep learning). In particular, he is interested in mathematically analyzing aspects of expressiveness, optimization and generalization, with the goal of deriving theoretically founded procedures and algorithms that improve practical performance. Nadav is also the CTO, President and a Co-Founder at Imubit, a company that applies neural networks for optimal control of industrial manufacturing, thereby reducing emissions while improving yields. Nadav earned a BSc in electrical engineering and a BSc in mathematics (both summa cum laude) at the Technion Excellence Program for Distinguished Undergraduates. He obtained his PhD (direct track) at the School of Computer Science and Engineering in the Hebrew University of Jerusalem, under the supervision of Prof. Amnon Shashua. Subsequently, he was a postdoctoral research scholar at the School of Mathematics in the Institute for Advanced Study of Princeton. For his contributions to the foundations of neural networks, Nadav received various honors and awards, including the ERC Starting Grant, the Google Research Scholar Award, the Google Doctoral Fellowship in Machine Learning, the Rothschild Postdoctoral Fellowship, and the Zuckerman Postdoctoral Fellowship.
UdeM/MILA
A Tensor Perspective on Second-Order RNNs
After providing a brief overview of the research conducted by my group at Mila and Université de Montréal, I will first introduce three families of models that are closely related: second-order RNNs, weighted automata and tensor networks. I will then provide a formal analysis of how tensor decomposition can be used to compress the parameters of second-order RNNs and how this affects their effectiveness. Specifically, our results demonstrate how the rank interacts with hidden size to control the model's capacity. This talk is based on the following publication from my group: Lizaire, Maude, Michael Rizvi-Martel, Marawan Gamal, and Guillaume Rabusseau. "A Tensor Decomposition Perspective on Second-Order RNNs." In Forty-First International Conference on Machine Learning.
Guillaume Rabusseau is an Associate Professor in the Department of Computer Science and Operations Research (DIRO) at Université de Montréal, a core member of Mila, and a Canada CIFAR AI Chair holder. His research focuses on the intersection of machine learning, theoretical computer science, and multilinear algebra. Specifically, he explores the connections between tensors and machine learning, developing efficient learning schemes for structured data by leveraging linear and multilinear algebra. His broader research interests include tensor decomposition techniques, the use of tensor networks for machine learning, quantum machine learning, kernel methods, (weighted) automata theory, and nonlinear computational models for strings, trees, and graphs.
University of Athens
A Tensor Network Approach to Equivariant Polynomial Networks
In this talk, we will investigate a coordinate-free attempt to formalize polynomial networks, stemming from the representation of polynomial maps as tensor networks. We will discuss how this framework encompasses previously proposed polynomial networks and adheres to the compositional architecture paradigm of deep learning. Ultimately, given group representations on the input and output spaces, we will see that this discussion naturally extends to the equivariant setting and elaborate on implementation aspects.
Alexandros Georgiou is a second year Ph.D. student at the National and Kapodistrian University of Athens, under the supervision of Yannis Panagakis, and a fellow of the Archimedes research unit. Previously, he obtained his diploma and M.Sc. from the School of Applied Mathematical and Physical Sciences of the National Technical University of Athens. His main research interests lie in the areas of geometric deep learning and tensor methods.
