Melanie is an Assistant Professor of Applied Mathematics and of Computer Science at Harvard University. Her research focuses on utilizing geometric structure in data for the design of efficient Machine Learning and Optimization methods. In 2021-2022, she was a Hooke Research Fellow at the Mathematical Institute in Oxford. Previously, she received her PhD from Princeton University (2021), held visiting positions at MIT and the Simons Institute in Berkeley, and interned in the research labs of Facebook, Google, and Microsoft. In addition to her academic work, she is the Chief Scientist of the Legal Artificial Intelligence startup Claudius.

Abstract: 

The problem of identifying geometric structure in heterogeneous, high-dimensional data is a cornerstone of Representation Learning. In this talk, we study the problem of data geometry from the perspective of Discrete Geometry. We start by reviewing discrete notions of curvature with a focus on discrete Ricci curvature. Then we discuss how curvature is linked to mesoscale structure in graphs, which gives rise to applications of discrete curvature in node clustering and community detection. For downstream machine learning and data science applications, it is often beneficial to represent graph-structured data in a continuous space, which may be Euclidean or Non-Euclidean. We show that discrete curvature allows for characterizing the geometry of a suitable embedding space both locally and in the sense of global curvature bounds, which have implications for graph-based learning.

Michael Bronstein joined the Department of Computing as Professor in 2018. He has served as a professor at USI Lugano, Switzerland since 2010 and held visiting positions at Stanford, Harvard, MIT, TUM, and Tel Aviv University. Michael received his PhD with distinction from the Technion (Israel Institute of Technology) in 2007. His main expertise is in theoretical and computational geometric methods for machine learning and data science, and his research encompasses a broad spectrum of applications ranging from computer vision and pattern recognition to geometry processing, computer graphics, and biomedicine. Michael has authored over 200 papers, a book, and holds over 35 granted patents. He was awarded five ERC grants, two Google Faculty Research awards, two Amazon ML Research awards, Facebook Computational Social Science award, Dalle Molle prize, Royal Society Wolfson Merit award, and Royal Academy of Engineering Silver Medal. He is a PI and ML Lead in Project CETI, a TED Audacious Prize winning collaboration aimed at understanding the communication of sperm whales. During 2017-2018 he was a fellow at the Radcliffe Institute for Advanced Study at Harvard University and since 2017, he is a Rudolf Diesel fellow at TU Munich. He was invited as a Young Scientist to the World Economic Forum, an honour bestowed on forty world’s leading scientists under the age of forty. Michael is a Member of Academia Europaea, Fellow of IEEE, IAPR, ELLIS, and BCS, alumnus of the Technion Excellence Program and the Academy of Achievement, and ACM Distinguished Speaker. In addition to academic work, Michael's industrial experience includes technological leadership in multiple startup companies, including Novafora, Videocites, Invision (acquired by Intel in 2012), and Fabula AI (acquired by Twitter in 2019). Following the acquisition of Fabula, he joined Twitter as Head of Graph Learning Research. He previously served as Principal Engineer at Intel Perceptual Computing (2012-2019) and was one of the key developers of the Intel RealSense 3D camera technology. He is also an angel investor and supporter of multiple early-stage startups. 

Title: Graph Rewiring in GNNs

Yusu Wang is currently Professor in the Halicioglu Data Science Institute at University of California, San Diego, where she also serves as the Associate Director for Research for the NSF National AI Institute TILOS. She obtained her PhD degree from Duke University in 2004, and from 2004-2005, she was a post-doctoral fellow at Stanford University. Yusu Wang primarily works in the fields of Computational geometry, Computational and applied topology, and appliations to data analysis. She received DOE Early Career Principal Investigator Award in 2006, and NSF Career Award in 2008. She is currently a member of the Computational Geometry Steering Committee, and also serves on the editorial boards for SIAM Journal on Computing (SICOMP) and Journal of Computational Geometry (JoCG).

Abstract: 

Machine learning, especially the use of neural netowrks have shown great success in a broad range of applications. Recently, neural approaches have also shown promise in tackling (combinatorial) optimization problems in a data-driven manner. On the other hand, for many problems, especially geometric optimization problems, many beautiful geometric ideas and algorithmic insights have been developed in fields such as theoretical computer science and computational geometry. Our goal is to infuse geometric and algorithmic ideas to the design of neural frameworks so that they can be more effective and generalize better. In this talk, I will give two examples in this direction. The first one is what we call a mixed Neural-algorithmic framework for the Steiner Tree problem in the Euclidean space, leveraging the celebrated PTAS algorithm by Arora. Interestingly, here the model complexity can be made independent of the input point set size. The second one is an neural architecture for approximating the Wasserstein distance between point sets, whose design /analysis uses a geometric coreset idea. 

Tess Smidt is an Assistant Professor of Electrical Engineering and Computer Science at MIT. Tess earned her SB in Physics from MIT in 2012 and her PhD in Physics from the University of California, Berkeley in 2018. Her research focuses on machine learning that incorporates physical and geometric constraints, with applications to materials design. Prior to joining the MIT EECS faculty, she was the 2018 Alvarez Postdoctoral Fellow in Computing Sciences at Lawrence Berkeley National Laboratory and a Software Engineering Intern on the Google Accelerated Sciences team where she developed Euclidean symmetry equivariant neural networks which naturally handle 3D geometry and geometric tensor data.

Abstract:

3D Euclidean symmetry-equivariant neural networks (E(3)NNs) are emerging as an effective machine learning paradigm in molecular modeling, protein design, computer graphics, and beyond. In this talk, I'll discuss the fundamental building blocks of E(3)NNs and how these pieces are combined to create the growing zoo of E(3)NNs available today.