| Cleanrl: High-quality single-file implementations of deep reinforcement learning algorithms S Huang, RFJ Dossa, C Ye, J Braga, D Chakraborty, K Mehta, ... Journal of Machine Learning Research 23 (274), 1-18, 2022 | 344 | 2022 |
| Mitigating harm in language models with conditional-likelihood filtration H Ngo, C Raterink, JGM Araújo, I Zhang, C Chen, A Morisot, N Frosst arXiv preprint arXiv:2108.07790, 2021 | 34 | 2021 |
| Transformers need glasses! information over-squashing in language tasks F Barbero, A Banino, S Kapturowski, D Kumaran, J Madeira Araújo, ... Advances in Neural Information Processing Systems 37, 98111-98142, 2024 | 16 | 2024 |
| Categorical deep learning: An algebraic theory of architectures B Gavranović, P Lessard, A Dudzik, T von Glehn, JGM Araújo, ... arXiv preprint arXiv:2402.15332, 2024 | 13 | 2024 |
| Open rl benchmark: Comprehensive tracked experiments for reinforcement learning S Huang, Q Gallouédec, F Felten, A Raffin, RFJ Dossa, Y Zhao, ... arXiv preprint arXiv:2402.03046, 2024 | 9 | 2024 |
| Position: Categorical deep learning is an algebraic theory of all architectures B Gavranović, P Lessard, A Dudzik, T Von Glehn, JGM Araújo, ... arXiv preprint arXiv:2402.15332, 2024 | 7 | 2024 |
| No news is good news: A critique of the one billion word benchmark H Ngo, JGM Araújo, J Hui, N Frosst arXiv preprint arXiv:2110.12609, 2021 | 6 | 2021 |
| On the consistency of hyper-parameter selection in value-based deep reinforcement learning J Obando-Ceron, JGM Araújo, A Courville, PS Castro arXiv preprint arXiv:2406.17523, 2024 | 5 | 2024 |
| Scalable training of language models using JAX pjit and TPUv4 J Yoo, K Perlin, SR Kamalakara, JGM Araújo arXiv preprint arXiv:2204.06514, 2022 | 4 | 2022 |
| Lifting the veil on hyper-parameters for value-based deep reinforcement learning JGM Araújo, JSO Ceron, PS Castro NeurIPS 2021 Workshop LatinX in AI, 2021 | 4 | 2021 |
| What makes a good feedforward computational graph? A Vitvitskyi, JGM Araújo, M Lackenby, P Veličković arXiv preprint arXiv:2502.06751, 2025 | | 2025 |
Shengyi HuangAllen Institute for Artificial IntelligenceПотвърден имейл адрес: allenai.org
