Eran Malach

Most-cited papers

Decoupling "when to update" from "how to update"
Eran Malach, S. Shalev-Shwartz
2017Neural Information Processing Systems671 citations
Semantic Scholar
Proving the Lottery Ticket Hypothesis: Pruning is All You Need
Eran Malach, Gilad Yehudai, S. Shalev-Shwartz, Ohad Shamir
2020International Conference on Machine Learning344 citations
Semantic Scholar
SGD Learns Over-parameterized Networks that Provably Generalize on Linearly Separable Data
Alon Brutzkus, A. Globerson, Eran Malach, S. Shalev-Shwartz
2017International Conference on Learning Representations288 citations
Semantic Scholar
Repeat After Me: Transformers are Better than State Space Models at Copying
Samy Jelassi, David Brandfonbrener, S. Kakade, Eran Malach
2024International Conference on Machine Learning251 citations
Semantic Scholar
Hidden Progress in Deep Learning: SGD Learns Parities Near the Computational Limit
B. Barak, Benjamin L. Edelman, Surbhi Goel, S. Kakade +2 more
2022Neural Information Processing Systems223 citations Open access
Semantic Scholar
Echo Chamber: RL Post-training Amplifies Behaviors Learned in Pretraining
Rosie Zhao, Alexandru Meterez, S. Kakade, Cengiz Pehlevan +2 more
2025arXiv.org151 citations
Semantic Scholar
The Evolution of Statistical Induction Heads: In-Context Learning Markov Chains
Benjamin L. Edelman, Ezra Edelman, Surbhi Goel, Eran Malach +1 more
2024Neural Information Processing Systems143 citations
Semantic Scholar
Learning Parities with Neural Networks
Amit Daniely, Eran Malach
2020Neural Information Processing Systems105 citations
Semantic Scholar
Auto-Regressive Next-Token Predictors are Universal Learners
Eran Malach
2023International Conference on Machine Learning79 citations
Semantic Scholar
LoRA Soups: Merging LoRAs for Practical Skill Composition Tasks
Akshara Prabhakar, Yuanzhi Li, K. Narasimhan, S. Kakade +2 more
2024International Conference on Computational Linguistics67 citations
Semantic Scholar
A New Perspective on Shampoo's Preconditioner
Depen Morwani, Itai Shapira, Nikhil Vyas, Eran Malach +2 more
2024International Conference on Learning Representations66 citations
Semantic Scholar
Is Deeper Better only when Shallow is Good?
Eran Malach, S. Shalev-Shwartz
2019Neural Information Processing Systems54 citations
Semantic Scholar
Don't Stop Me Now: Embedding Based Scheduling for LLMs
Rana Shahout, Eran Malach, Chun-Wei Liu, Wei-Fan Jiang +2 more
2024International Conference on Learning Representations51 citations
Semantic Scholar
Quantifying the Benefit of Using Differentiable Learning over Tangent Kernels
Eran Malach, Pritish Kamath, Emmanuel Abbe, N. Srebro
2021International Conference on Machine Learning46 citations
Semantic Scholar
A Provably Correct Algorithm for Deep Learning that Actually Works
Eran Malach, S. Shalev-Shwartz
2018arXiv.org41 citations
Semantic Scholar
Computational Separation Between Convolutional and Fully-Connected Networks
Eran Malach, S. Shalev-Shwartz
2020International Conference on Learning Representations33 citations
Semantic Scholar
On the Power of Differentiable Learning versus PAC and SQ Learning
Emmanuel Abbe, Pritish Kamath, Eran Malach, Colin Sandon +1 more
2021Neural Information Processing Systems32 citations
Semantic Scholar
Decoupling Gating from Linearity
Jonathan Fiat, Eran Malach, S. Shalev-Shwartz
2018arXiv.org32 citations
Semantic Scholar
To Backtrack or Not to Backtrack: When Sequential Search Limits Model Reasoning
Tian Qin, David Alvarez-Melis, Samy Jelassi, Eran Malach
2025arXiv.org30 citations
Semantic Scholar
Transcendence: Generative Models Can Outperform The Experts That Train Them
Edwin Zhang, Vincent Zhu, Naomi Saphra, Anat Kleiman +4 more
2024Neural Information Processing Systems30 citations
Semantic Scholar
Universal Length Generalization with Turing Programs
Kaiying Hou, David Brandfonbrener, S. Kakade, Samy Jelassi +1 more
2024arXiv.org27 citations
Semantic Scholar
The Connection Between Approximation, Depth Separation and Learnability in Neural Networks
Eran Malach, Gilad Yehudai, S. Shalev-Shwartz, Ohad Shamir
2021Annual Conference Computational Learning Theory25 citations
Semantic Scholar
Mixture of Parrots: Experts improve memorization more than reasoning
Samy Jelassi, Clara Mohri, David Brandfonbrener, A. Gu +6 more
2024International Conference on Learning Representations24 citations
Semantic Scholar
ID3 Learns Juntas for Smoothed Product Distributions
Alon Brutzkus, Amit Daniely, Eran Malach
2019Annual Conference Computational Learning Theory23 citations
Semantic Scholar
Loss-to-Loss Prediction: Scaling Laws for All Datasets
David Brandfonbrener, Nikhil Anand, Nikhil Vyas, Eran Malach +1 more
2024Trans. Mach. Learn. Res.20 citations
Semantic Scholar
Knowledge Distillation: Bad Models Can Be Good Role Models
Gal Kaplun, Eran Malach, Preetum Nakkiran, S. Shalev-Shwartz
2022Neural Information Processing Systems19 citations Open access
Semantic Scholar
Less is More: Selective Layer Finetuning with SubTuning
Gal Kaplun, Andrey Gurevich, Tal Swisa, Mazor David +2 more
202316 citations
Semantic Scholar
Rethinking JEPA: Compute-Efficient Video SSL with Frozen Teachers
Xianhang Li, Chen Huang, Chun-liang Li, Eran Malach +3 more
2025arXiv.org15 citations
Semantic Scholar
The Implications of Local Correlation on Learning Some Deep Functions
Eran Malach, S. Shalev-Shwartz
2020Neural Information Processing Systems15 citations
Semantic Scholar
Pareto Frontiers in Neural Feature Learning: Data, Compute, Width, and Luck
Benjamin L. Edelman, Surbhi Goel, S. Kakade, Eran Malach +1 more
2023arXiv.org14 citations Open access
Semantic Scholar
Let Me Think! A Long Chain-of-Thought Can Be Worth Exponentially Many Short Ones
Parsa Mirtaheri, Ezra Edelman, Samy Jelassi, Eran Malach +1 more
2025Neural Information Processing Systems12 citations
Semantic Scholar
When Hardness of Approximation Meets Hardness of Learning
Eran Malach, S. Shalev-Shwartz
2020Journal of machine learning research11 citations
Semantic Scholar
How Reinforcement Learning After Next-Token Prediction Facilitates Learning
Nikolaos Tsilivis, Eran Malach, Karen Ullrich, Julia Kempe
2025arXiv.org10 citations
Semantic Scholar
Efficient Learning of CNNs using Patch Based Features
Alon Brutzkus, A. Globerson, Eran Malach, Alon Regev Netser +1 more
2022International Conference on Machine Learning10 citations
Semantic Scholar
On the Optimality of Trees Generated by ID3
Alon Brutzkus, Amit Daniely, Eran Malach
2019arXiv.org10 citations
Semantic Scholar
Decomposing Elements of Problem Solving: What "Math" Does RL Teach?
Tian Qin, C. Park, Mujin Kwun, Aaron Walsman +4 more
2025arXiv.org9 citations
Semantic Scholar
The Role of Sparsity for Length Generalization in Transformers
Noah Golowich, Samy Jelassi, David Brandfonbrener, S. Kakade +1 more
2025arXiv.org9 citations
Semantic Scholar
Pareto Frontiers in Deep Feature Learning: Data, Compute, Width, and Luck
Benjamin L. Edelman, Surbhi Goel, S. Kakade, Eran Malach +1 more
2023Neural Information Processing Systems8 citations
Semantic Scholar
Annotations Mitigate Post-Training Mode Collapse
Jacob Mitchell Springer, Madhu Advani, Lukas Aichberger, Arwen Bradley +6 more
2026arXiv.org5 citations
Semantic Scholar
To Infinity and Beyond: Tool-Use Unlocks Length Generalization in State Space Models
Eran Malach, O. Saremi, Sinead Williamson, Arwen Bradley +4 more
2025arXiv.org5 citations
Semantic Scholar
Learning Boolean Circuits with Neural Networks
Eran Malach, S. Shalev-Shwartz
2019arXiv.org5 citations
Semantic Scholar
The Role of Sparsity for Length Generalization in LLMs
Noah Golowich, Samy Jelassi, David Brandfonbrener, S. Kakade +1 more
2025International Conference on Machine Learning4 citations
Semantic Scholar
Is ML-Based Cryptanalysis Inherently Limited? Simulating Cryptographic Adversaries via Gradient-Based Methods
Avital Shafran, Eran Malach, Thomas Ristenpart, Gil Segev +1 more
2024IACR Cryptology ePrint Archive4 citations
Semantic Scholar
SubTuning: Efficient Finetuning for Multi-Task Learning
Gal Kaplun, Andrey Gurevich, Tal Swisa, Mazor David +2 more
2023arXiv.org4 citations Open access
Semantic Scholar
LLM Priors for ERM over Programs
Shivam Singhal, Priyadarsi Mishra, Eran Malach, Tomer Galanti
20252 citations
Semantic Scholar
A Taxonomy of Transcendence
Natalie Abreu, Edwin Zhang, Eran Malach, Naomi Saphra
2025arXiv.org2 citations
Semantic Scholar
On the Power of Decision Trees in Auto-Regressive Language Modeling
Yu-Lu Gan, Tomer Galanti, Tomaso A. Poggio, Eran Malach
2024Neural Information Processing Systems2 citations
Semantic Scholar
On the Importance of Gating: Memorization vs. In-Context Learning in State Space Models
W. Tong, Aryo Lotfi, Emmanuel Abbé, Kostas Vaggelakos +5 more
20261 citations
Semantic Scholar
Fewer Bits, More Tries: Recovering Capability Under Memory Constraints
Tomer Galanti, Josh M. Susskind, Eran Malach
20260 citations
Semantic Scholar
Weighted State Souping for Efficient Long-Context Retrieval
W. Tong, Aryo Lotfi, Emmanuel Abbé, Josh M. Susskind +2 more
20260 citations
Semantic Scholar