Srinadh Bhojanapalli
Srinadh Bhojanapalli
Research Scientist, Google Research
Dirección de correo verificada de - Página principal
Citado por
Citado por
Exploring generalization in deep learning
B Neyshabur, S Bhojanapalli, D McAllester, N Srebro
Advances in neural information processing systems 30, 2017
Large batch optimization for deep learning: Training bert in 76 minutes
Y You, J Li, S Reddi, J Hseu, S Kumar, S Bhojanapalli, X Song, J Demmel, ...
arXiv preprint arXiv:1904.00962, 2019
A pac-bayesian approach to spectrally-normalized margin bounds for neural networks
B Neyshabur, S Bhojanapalli, N Srebro
arXiv preprint arXiv:1707.09564, 2017
Towards understanding the role of over-parametrization in generalization of neural networks
B Neyshabur, Z Li, S Bhojanapalli, Y LeCun, N Srebro
arXiv preprint arXiv:1805.12076, 2018
Implicit regularization in matrix factorization
S Gunasekar, BE Woodworth, S Bhojanapalli, B Neyshabur, N Srebro
Advances in neural information processing systems 30, 2017
Global optimality of local search for low rank matrix recovery
S Bhojanapalli, B Neyshabur, N Srebro
Advances in Neural Information Processing Systems, 3873-3881, 2016
Understanding robustness of transformers for image classification
S Bhojanapalli, A Chakrabarti, D Glasner, D Li, T Unterthiner, A Veit
Proceedings of the IEEE/CVF international conference on computer vision …, 2021
Does label smoothing mitigate label noise?
M Lukasik, S Bhojanapalli, A Menon, S Kumar
International Conference on Machine Learning, 6448-6458, 2020
Are transformers universal approximators of sequence-to-sequence functions?
C Yun, S Bhojanapalli, AS Rawat, SJ Reddi, S Kumar
arXiv preprint arXiv:1912.10077, 2019
Dropping convexity for faster semi-definite optimization
S Bhojanapalli, A Kyrillidis, S Sanghavi
Conference on Learning Theory, 530-582, 2016
Coherent matrix completion.
Y Chen, S Bhojanapalli, S Sanghavi, R Ward
arXiv preprint arXiv:1306.2979, 2013
Universal matrix completion
S Bhojanapalli, P Jain
International Conference on Machine Learning, 1881-1889, 2014
Modifying memories in transformer models
C Zhu, AS Rawat, M Zaheer, S Bhojanapalli, D Li, F Yu, S Kumar
arXiv preprint arXiv:2012.00363, 2020
Stabilizing GAN training with multiple random projections
B Neyshabur, S Bhojanapalli, A Chakrabarti
arXiv preprint arXiv:1705.07831, 2017
Completing any low-rank matrix, provably
Y Chen, S Bhojanapalli, S Sanghavi, R Ward
The Journal of Machine Learning Research 16 (1), 2999-3034, 2015
Low-rank bottleneck in multi-head attention models
S Bhojanapalli, C Yun, AS Rawat, S Reddi, S Kumar
International conference on machine learning, 864-873, 2020
Coping with label shift via distributionally robust optimisation
J Zhang, A Menon, A Veit, S Bhojanapalli, S Kumar, S Sra
arXiv preprint arXiv:2010.12230, 2020
O (n) connections are expressive enough: Universal approximability of sparse transformers
C Yun, YW Chang, S Bhojanapalli, AS Rawat, S Reddi, S Kumar
Advances in Neural Information Processing Systems 33, 13783-13794, 2020
A simple and effective positional encoding for transformers
PC Chen, H Tsai, S Bhojanapalli, HW Chung, YW Chang, CS Ferng
arXiv preprint arXiv:2104.08698, 2021
The lazy neuron phenomenon: On emergence of activation sparsity in transformers
Z Li, C You, S Bhojanapalli, D Li, AS Rawat, SJ Reddi, K Ye, F Chern, ...
arXiv preprint arXiv:2210.06313, 2022
El sistema no puede realizar la operación en estos momentos. Inténtalo de nuevo más tarde.
Artículos 1–20