Iddo Drori. Ph.D.

Iddo Drori, Ph.D.

iddo.drori@yu.edu | 646-592-4763 | 205 Lexington Avenue, 7th FL, NYC

Iddo Drori, an associate professor, runs a superintelligence lab. He was an associate professor of the practice in the department of computer science at Boston University. Previously, he was a lecturer at MIT in EECS and visiting at MIT CSAIL, a visiting associate professor at Cornell University in operations research and information engineering, and a research scientist and adjunct professor at NYU Center for Data Science, Courant Institute, and NYU Tandon. 

For the past decade he has been teaching as an adjunct at Columbia University computer science. He has over 80 publications with over 7,000 citations, and has taught over 50 courses in computer science. He is the author of the textbook “The Science of Deep Learning,” published by Cambridge University Press, and the forthcoming book “Artificial General Intelligence: Mathematical Foundations”. He has a visiting position at Stanford University, and visiting associate professor at Tel Aviv University. 

Prof. Drori holds a Ph.D. in computer science and was a postdoctoral research fellow in statistics at Stanford University. He also holds graduate degrees in computer science and Mathematics from the Hebrew University of Jerusalem and an MBA in organizational behavior and entrepreneurship from Tel Aviv University, and has a decade of industry research and leadership experience. He has won multiple competitions in AI and computer vision conferences and received multiple best paper awards in machine learning conferences.

Selected Publications

• A neural network solves, explains, and generates university math problems by program synthesis and few-shot learning at human level
• Combinatorial optimization by graph pointer networks and hierarchical reinforcement learning
• Sparse solution of underdetermined systems of linear equations by stagewise orthogonal matching pursuit - IEEE
• Bilateral mesh denoising - ACM
• Human-in-the-loop AI reviewing: Feasibility, opportunities, and risks
• Learning to solve combinatorial optimization problems on real-world graphs in linear time
• AlphaD3M: Machine learning pipeline synthesis 
• Human evaluation of text-to-image models on a multi-task benchmark

Research

• ReviewerArena: Evaluate LLM reviewer quality based on preferences by direct and anonymous comparison of reviews.
• OpenReviewer: Learn to improve your academic papers by generating real-time reviews.
• Papers with reviews: Read top ranked arXiv and open access Nature papers.
• Machine Learning for Education: Can a machine solve, explain, and generate university-level mathematics and STEM courses? Our latest research published in PNAS and featured by MIT news demonstrates that a neural network solves, explains, and generates university math problems by program synthesis and few-shot learning at human level.
• Machine Learning for Climate Science: In the spirit of MIT’s leading efforts to take action against climate change for a better world, we’ve published multidisciplinary work in Nature Scientific Reports on computer vision methods for tracking turbulent structures in the plasma of a fusion reactor, and on predicting the Atlantic Multi-decadal Variability and ocean biogeochemistry, awarded best paper: pathway to impact at NeurIPS CCAI.
• Machine Learning for Autonomous Driving: Can we accurately predict trajectories and learn to drive? we’ve won the ICCV learning to drive challenge and continuously improve performance; society will accept autonomous vehicles once they are orders of magnitude safer than humans. 

Teaching Experiences

• Neural Computation
• Artificial General Intelligence
• Deep Learning
• Artificial Intelligence
• Machine Learning
• Meta Learning
• Reinforcement Learning
• Data Science

Recent Awards

• 2024 International Mathematical Olympiad (IMO): AI X Prize - Zack Meeks, Xi Chen, Mao Mao, Akshat Gurbuxani, Iddo Drori - Silver medal
• 2022 NeurIPS open-ended learning competitions: MineRL BASALT and Neural MMO
• 2021 CCAI NeurIPS Best Paper Award Winner
• 2021 FG Competition Winner, Kinship Verification Challenge
• ACML 2021 Best Student Paper Award Winner
• ICCV 2019 Competition Winner, Learning to Drive Challenge