| Decoupling" when to update" from" how to update" E Malach, S Shalev-Shwartz Advances in neural information processing systems 30, 2017 | 688 | 2017 |
| Proving the lottery ticket hypothesis: Pruning is all you need E Malach, G Yehudai, S Shalev-Schwartz, O Shamir International Conference on Machine Learning, 6682-6691, 2020 | 316 | 2020 |
| SGD learns over-parameterized networks that provably generalize on linearly separable data A Brutzkus, A Globerson, E Malach, S Shalev-Shwartz arXiv preprint arXiv:1710.10174, 2017 | 303 | 2017 |
| Hidden progress in deep learning: Sgd learns parities near the computational limit B Barak, B Edelman, S Goel, S Kakade, E Malach, C Zhang Advances in Neural Information Processing Systems 35, 21750-21764, 2022 | 135 | 2022 |
| Learning parities with neural networks A Daniely, E Malach Advances in Neural Information Processing Systems 33, 20356-20365, 2020 | 97 | 2020 |
| Is deeper better only when shallow is good? E Malach, S Shalev-Shwartz Advances in Neural Information Processing Systems 32, 2019 | 54 | 2019 |
| Repeat after me: Transformers are better than state space models at copying S Jelassi, D Brandfonbrener, SM Kakade, E Malach arXiv preprint arXiv:2402.01032, 2024 | 52 | 2024 |
| Quantifying the benefit of using differentiable learning over tangent kernels E Malach, P Kamath, E Abbe, N Srebro International Conference on Machine Learning, 7379-7389, 2021 | 51 | 2021 |
| A provably correct algorithm for deep learning that actually works E Malach, S Shalev-Shwartz arXiv preprint arXiv:1803.09522, 2018 | 33 | 2018 |
| The evolution of statistical induction heads: In-context learning markov chains BL Edelman, E Edelman, S Goel, E Malach, N Tsilivis arXiv preprint arXiv:2402.11004, 2024 | 30 | 2024 |
| Decoupling gating from linearity J Fiat, E Malach, S Shalev-Shwartz arXiv preprint arXiv:1906.05032, 2019 | 30 | 2019 |
| On the power of differentiable learning versus PAC and SQ learning E Abbe, P Kamath, E Malach, C Sandon, N Srebro Advances in Neural Information Processing Systems 34, 24340-24351, 2021 | 23 | 2021 |
| Auto-regressive next-token predictors are universal learners E Malach arXiv preprint arXiv:2309.06979, 2023 | 22 | 2023 |
| The connection between approximation, depth separation and learnability in neural networks E Malach, G Yehudai, S Shalev-Schwartz, O Shamir Conference on Learning Theory, 3265-3295, 2021 | 22 | 2021 |
| ID3 learns juntas for smoothed product distributions A Brutzkus, A Daniely, E Malach Conference on Learning Theory, 902-915, 2020 | 21 | 2020 |
| Knowledge distillation: Bad models can be good role models G Kaplun, E Malach, P Nakkiran, S Shalev-Shwartz Advances in Neural Information Processing Systems 35, 28683-28694, 2022 | 20 | 2022 |
| Computational separation between convolutional and fully-connected networks E Malach, S Shalev-Shwartz arXiv preprint arXiv:2010.01369, 2020 | 19 | 2020 |
| On the optimality of trees generated by id3 A Brutzkus, A Daniely, E Malach arXiv preprint arXiv:1907.05444, 2019 | 13 | 2019 |
| When hardness of approximation meets hardness of learning E Malach, S Shalev-Shwartz Journal of Machine Learning Research 23 (91), 1-24, 2022 | 12 | 2022 |
| The implications of local correlation on learning some deep functions E Malach, S Shalev-Shwartz Advances in Neural Information Processing Systems 33, 1322-1332, 2020 | 12 | 2020 |
Shai Shalev-ShwartzThe Hebrew UniversityVerified email at cs.huji.ac.il