| On large-batch training for deep learning: Generalization gap and sharp minima NS Keskar, D Mudigere, J Nocedal, M Smelyanskiy, PTP Tang arXiv preprint arXiv:1609.04836, 2016 | 608 | 2016 |
| Regularizing and optimizing LSTM language models S Merity, NS Keskar, R Socher arXiv preprint arXiv:1708.02182, 2017 | 303 | 2017 |
| Improving generalization performance by switching from adam to sgd NS Keskar, R Socher arXiv preprint arXiv:1712.07628, 2017 | 76 | 2017 |
| The natural language decathlon: Multitask learning as question answering B McCann, NS Keskar, C Xiong, R Socher arXiv preprint arXiv:1806.08730, 2018 | 57 | 2018 |
| An analysis of neural language modeling at multiple scales S Merity, NS Keskar, R Socher arXiv preprint arXiv:1803.08240, 2018 | 54 | 2018 |
| Weighted transformer network for machine translation K Ahmed, NS Keskar, R Socher arXiv preprint arXiv:1711.02132, 2017 | 31 | 2017 |
| Balancing communication and computation in distributed optimization A Berahas, R Bollapragada, NS Keskar, E Wei IEEE Transactions on Automatic Control, 2018 | 20 | 2018 |
| adaqn: An adaptive quasi-newton algorithm for training rnns NS Keskar, AS Berahas Joint European Conference on Machine Learning and Knowledge Discovery in …, 2016 | 20 | 2016 |
| A second-order method for convex 1-regularized optimization with active-set prediction N Keskar, J Nocedal, F Öztoprak, A Waechter Optimization Methods and Software 31 (3), 605-621, 2016 | 14 | 2016 |
| Identifying generalization properties in neural networks H Wang, NS Keskar, C Xiong, R Socher arXiv preprint arXiv:1809.07402, 2018 | 13 | 2018 |
| A nonmonotone learning rate strategy for SGD training of deep neural networks NS Keskar, G Saon 2015 IEEE International Conference on Acoustics, Speech and Signal …, 2015 | 9 | 2015 |
| A limited-memory quasi-Newton algorithm for bound-constrained non-smooth optimization N Keskar, A Wächter Optimization Methods and Software 34 (1), 150-171, 2019 | 7 | 2019 |
| A Closer Look at Deep Learning Heuristics: Learning rate restarts, Warmup and Distillation A Gotmare, NS Keskar, C Xiong, R Socher arXiv preprint arXiv:1810.13243, 2018 | 7 | 2018 |
| Coarse-grain fine-grain coattention network for multi-evidence question answering V Zhong, C Xiong, NS Keskar, R Socher arXiv preprint arXiv:1901.00603, 2019 | 6 | 2019 |
| Ctrl: A conditional transformer language model for controllable generation NS Keskar, B McCann, LR Varshney, C Xiong, R Socher arXiv preprint arXiv:1909.05858, 2019 | 3 | 2019 |
| Unifying Question Answering and Text Classification via Span Extraction NS Keskar, B McCann, C Xiong, R Socher arXiv preprint arXiv:1904.09286, 2019 | 3 | 2019 |
| Using mode connectivity for loss landscape analysis A Gotmare, NS Keskar, C Xiong, R Socher arXiv preprint arXiv:1806.06977, 2018 | 3 | 2018 |
| Neural text summarization: A critical evaluation W Kryściński, NS Keskar, B McCann, C Xiong, R Socher arXiv preprint arXiv:1908.08960, 2019 | 2 | 2019 |
| Pretrained AI Models: Performativity, Mobility, and Change LR Varshney, NS Keskar, R Socher arXiv preprint arXiv:1909.03290, 2019 | | 2019 |
| Multitask Learning As Question Answering NS Keskar, B McCann, C Xiong, R Socher US Patent App. 15/974,075, 2019 | | 2019 |
Richard SocherChief Scientist at SalesforceVerified email at stanford.edu