Selected Talks

AI generated classes
What questions does a student need to answer to fulfill the graduation requirements of MIT in Mathematics and EECS in any major? Can a language model fulfill the graduation requirements? How can we ensure that a language model generates correct and complete answers? Which learning approaches allow language models to pass the MIT curriculum and why? What are the most challenging classes, concepts, tasks, and types of questions? What questions, topics, and classes help answer other questions, understand other topics, and are prerequisites for other classes based on data? How can we evaluate student performance while allowing the usage of large language models? How can we automatically generate questions at scale with correct answers? How can we rapidly write complete books while ensuring quality and originality? How can we drive a photo-realistic speaking avatar to act as a TA or lecturer? Our work answers these questions.
 Massachusetts Institute of Technology, Faculty Seminar
February, 2023
 Columbia University, Senate
Host: Julia Hirschberg
January, 2023

Learning to learn courses slides
 MIT-CalTech-UPenn-Stanford, NeuroSym Seminar
Host: Armando Solar-Lezama
July, 2022
 Boston University, Department of Computer Science
Host: Kate Saenko
April, 2022
 McMaster University, Distinguished Lecture Seminar Series, Department of Computing and Software
Host: Mark Lawford
March, 2022
 Massachusetts Institute of Technology, 18.065
Host: Gilbert Strang
March, 2022
 Harvard University, Center of Mathematical Sciences and Applications
Host: Michael Douglas
March, 2022
 Cornell University, Department of Computer Science
Host: Thorsten Joachims
February, 2022
 Worcester Polytechnic Institute, Computer Science Department
Host: Daniel Reichman
January, 2022
 The City University of New York
January, 2022
 Illinois Institute of Technology, Department of Computer Science
Host: Shlomo Argamon
January, 2022
 Bar-Ilan University, Israel
August, 2021
 Massachusetts Institute of Technology
τβπ, Da Vinci Lecture
February, 2021

SARS-CoV-2 proteins structure prediction
 Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science
 Columbia University, Zuckerman Institute, COVID-19 Virtual Symposium
 Tel-Aviv University, School of Computer Science, Structural Bioinformatics seminar
April 2020

The science of deep learning
 Carnegie Mellon University, Machine Learning Department
March 2019

Learning deep learning
 Columbia University, Department of Computer Science
February 2019

Automatic machine learning for medical imaging
 Columbia University, Data Science Institute
Workshop, February 2019

Automatic machine learning
 Tel Aviv University, Information Systems
January 2019

AlphaX: generalizing AlphaZero
 Ben-Gurion University, Department of Computer Science
Computer Science seminar
January 2019

Automatic machine learning
 International Conference on Machine Learning, AutoML, panel member
July 2018

Adaptive dual process theory, neural network architectures, and applications
 NYU, Center for Data Science
Host: Kyunghyun Cho
Data Science seminar, April 2018

Visual task recommendation
 Tel Aviv University
Institute for Internet Studies, May 2017

Sparse solution of underdetermined systems of equations
 Ben-Gurion University, Department of Computer Science
Computer Science seminar, February 2007
 Hebrew University of Jerusalem, Dept of Statistics
Statistics seminar, November 2006

Fast L1 minimization
 Stanford University, Department of Statistics
Host: Bradley Efron
Statistics seminar, July 2006

Iterative thresholding for rapid sparse solution of underdetermined linear systems
 Stanford University, ICME
Host: Gene Golub
Linear Algebra and Optimization seminar, September 2005

Multi-scale representations for manifold valued data
 Technion, Faculty of Industrial Engineering
Statistics seminar, March 2007
 Weizmann Institute
Vision and Robotics seminar, August 2005

Video operations in the gradient domain
 MIT, CSAIL
Host: Fredo Durand
Graphics meeting, August 2004

Spectral sound gap filling
 Weizmann Institute
Vision and Robotics seminar, June 2004

Gradient video compositing, matting and completion
 Hebrew University of Jerusalem
Computer Vision seminar, May 2004
 Technion - Tel Aviv University
3rd Workshop on Geometric Computing, May 2004

Gradient video compositing
 Tel-Aviv University
Israel SIGGRAPH Chapter meeting, November 2003

Fragment-based image completion
 Technion - Tel Aviv University
2nd Workshop on Geometric Computing, May 2003
 Weizmann Institute
Vision and Robotics seminar, June 2003

Example-based image completion
 The Interdisciplinary Center
Israel SIGGRAPH Chapter meeting, November 2002

Example-based style synthesis
 Weizmann Institute
Vision and Robotics seminar, April 2002

Example-based rendering
 Weizmann Institute
Computer Vision annual seminar, December 2001
 Tel Aviv University
Center of Geometric Computing, October 2001

Fast multi-resolution image operations in the wavelet domain
 Tel Aviv University
Israel SIGGRAPH Chapter meeting, February 2000