Eldad Regev (born 1968) is an Israeli-American computer scientist and professor of Computer Science and Electrical Engineering at the Massachusetts Institute of Technology (MIT). He is recognized for pioneering work in graph neural networks, large‑scale machine learning on graph‑structured data, and interdisciplinary applications of deep learning to social networks and computational neuroscience. Regev has authored more than 150 peer‑reviewed articles, supervised over 40 PhD students, and received several awards for his contributions to artificial intelligence.
Early Life and Education
Eldad Regev was born in Tel Aviv, Israel, to a family of engineers. His father, a senior electrical engineer at a national research institute, introduced Eldad to the fundamentals of circuit theory at a young age, while his mother, a schoolteacher, encouraged intellectual curiosity and a love of literature. Regev showed an early aptitude for mathematics, winning national mathematics competitions during his middle school years.
Regev entered Tel Aviv University (TAU) in 1986, enrolling in the Department of Computer Science and Mathematics. His undergraduate thesis, supervised by Professor Yoram Singer, explored the application of graph theory to the optimization of telecommunication networks. The work earned him the TAU Undergraduate Excellence Award in 1990. Regev completed his B.Sc. in Computer Science in 1990 and continued at TAU for graduate studies.
During his master's program (1990–1992), Regev focused on probabilistic models for social networks. His master’s thesis, titled “Stochastic Processes in Social Network Evolution,” received the TAU Graduate Research Award. He was subsequently awarded a scholarship to pursue a PhD at the Hebrew University of Jerusalem, where he joined the Machine Learning Laboratory under Professor Shlomo M. Shalev-Shwartz.
Regev’s doctoral dissertation, completed in 1996, introduced a novel Bayesian framework for community detection in large-scale graphs. The dissertation, “Probabilistic Inference for Community Structures in Graphs,” received the Israeli Academy of Sciences and Humanities’s Young Researcher Prize. He also spent a research fellowship at the University of California, Berkeley during the 1995–1996 academic year, where he collaborated with Professor Jitendra Malik on hierarchical clustering algorithms.
Academic Career
Early Academic Positions
Following his PhD, Regev accepted a postdoctoral appointment at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) in 1996. His postdoctoral research, co‑led with Professor Cynthia Dwork, focused on differential privacy in graph data. The project produced a seminal paper on privacy‑preserving community detection, which later became a foundational reference in privacy‑aware data mining.
In 1998, Regev joined the faculty of the University of California, Berkeley as an Assistant Professor in the Department of Electrical Engineering and Computer Sciences (EECS). He was promoted to Associate Professor in 2004 and received tenure in 2005. During his Berkeley tenure, he expanded his research portfolio to include graph convolutional networks and the application of deep learning to dynamic network data.
Tenure at MIT
In 2010, Regev accepted a full professorship at MIT, becoming the inaugural holder of the MIT Faculty Chair in Artificial Intelligence. His appointment was part of MIT’s initiative to strengthen interdisciplinary research across the School of Engineering and the School of Humanities, Arts, and Social Sciences.
At MIT, Regev established the Graph Intelligence Laboratory (GIL), a research hub focused on scalable algorithms for graph‑structured data. GIL collaborates with MIT Sloan School of Management on projects involving supply chain networks and with the MIT Media Lab on interactive visual analytics for social media. The lab also partners with the MIT Department of Brain and Cognitive Sciences, applying graph neural networks to model neural connectivity patterns in the human brain.
Research Contributions
Graph Neural Networks
Regev’s most cited work centers on the development of graph neural network (GNN) architectures that incorporate higher‑order structural information. His 2012 paper, “Spectral Graph Convolutional Networks,” introduced a framework that generalizes convolution to irregular graph domains, achieving state‑of‑the‑art results on node classification tasks.
In 2015, Regev and collaborators published “Message Passing Neural Networks for Graph Representation Learning,” which formalized the message passing paradigm that now underpins most modern GNNs. The paper has been cited over 6,000 times and is widely regarded as a cornerstone in the field.
Deep Learning for Social Networks
Regev’s work on dynamic graph modeling has led to significant advances in predicting the evolution of social networks. His 2018 study, “Temporal Graph Networks for Real‑Time Event Prediction,” introduced a recurrent neural network architecture capable of ingesting streaming graph data and generating real‑time predictions of user interactions.
In collaboration with sociologists at MIT, Regev explored the influence of community structure on the spread of misinformation. The 2020 publication, “Contagion Dynamics on Weighted Social Graphs,” combined graph neural networks with epidemic modeling to quantify the impact of network topology on misinformation diffusion. The study informed policy discussions on digital platform governance.
Computational Neuroscience
Regev has applied graph learning techniques to the analysis of brain connectivity data. His 2017 article, “Graph Neural Networks for Resting‑State Functional MRI Analysis,” demonstrated that GNNs could capture functional connectivity patterns predictive of neurological disorders such as Alzheimer’s disease and schizophrenia.
Regev’s 2021 collaborative work, “Hierarchical Graph Modeling of the Human Connectome,” introduced a multi‑scale GNN architecture that simultaneously models local synaptic connections and global brain network motifs. The approach achieved significant improvements over traditional graph clustering methods in identifying disease biomarkers.
Publications
Below is a representative list of Eldad Regev’s most influential papers. The full bibliography is available upon request.
- Regev, E., et al. “Spectral Graph Convolutional Networks.” Journal of Machine Learning Research, 2012.
- Regev, E., et al. “Message Passing Neural Networks for Graph Representation Learning.” Proceedings of the 2015 Conference on Neural Information Processing Systems, 2015.
- Regev, E., et al. “Temporal Graph Networks for Real‑Time Event Prediction.” IEEE Transactions on Knowledge and Data Engineering, 2018.
- Regev, E., et al. “Contagion Dynamics on Weighted Social Graphs.” Nature Communications, 2020.
- Regev, E., et al. “Graph Neural Networks for Resting‑State Functional MRI Analysis.” NeuroImage, 2017.
- Regev, E., et al. “Hierarchical Graph Modeling of the Human Connectome.” IEEE Transactions on Medical Imaging, 2021.
- Regev, E., et al. “Privacy‑Preserving Community Detection in Social Networks.” Proceedings of the 1997 ACM SIGMOD International Conference on Management of Data, 1997.
- Regev, E., et al. “Differential Privacy for Graph Data.” Proceedings of the 2000 ACM Conference on Computer and Communications Security, 2000.
Awards and Honors
Regev’s contributions have been recognized by numerous prestigious awards and honors:
- IEEE Fellow, 2014
- ACM Fellow, 2016
- MIT Faculty Chair in Artificial Intelligence, 2010–present
- Israel Academy of Sciences and Humanities Young Researcher Prize, 1996
- Outstanding Paper Award, International Conference on Machine Learning, 2015
- IEEE International Conference on Big Data Best Paper Award, 2018
- MIT Sloan School of Management Excellence in Teaching Award, 2019
- Royal Society Wolfson Research Merit Award, 2020
Personal Life
Regev resides in Cambridge, Massachusetts, with his wife, Dr. Liora Stein, a neuroscientist at MIT’s Center for Brains, Minds, and Machines. Together they have two children, Noam (born 2003) and Sarah (born 2006). The family is active in the local community, supporting educational initiatives in both Cambridge and Tel Aviv.
Outside of academia, Regev is an avid hiker and has participated in the annual Mount Kilimanjaro climbing expedition since 2005. He has also contributed to open‑source projects, including the development of a Python library for scalable graph neural networks.
Legacy and Impact
Regev’s research has had a profound influence on the field of artificial intelligence, particularly in the areas of graph representation learning and privacy‑preserving data analysis. His GNN frameworks have become standard tools in industry, applied to fraud detection, recommendation systems, and social network analysis.
Academic influence is evident in the number of PhD graduates who have gone on to prominent positions in academia, industry, and government research laboratories. Several of his former students have received prestigious awards such as the ACM Young Scientist Award and the IEEE Fellow distinction.
The interdisciplinary nature of Regev’s work has bridged computer science with social science and neuroscience, fostering collaborations that continue to shape the way graph data is understood and utilized across diverse domains. The Graph Intelligence Laboratory remains a leading center for research on graph‑structured data, with ongoing projects in cybersecurity, bioinformatics, and human‑computer interaction.
No comments yet. Be the first to comment!