Publication

Export 8 results:
Filters: Author is T. Galanti  [Clear All Filters]
2023
Rangamani, A., Lindegaard, M., Galanti, T. & Poggio, T. Feature learning in deep classifiers through Intermediate Neural Collapse. (2023).PDF icon Feature_Learning_memo.pdf (2.16 MB)
Xu, M. et al. The Janus effects of SGD vs GD: high noise and low rank. (2023).PDF icon The_Janus_effects_of_SGD_vs_GD__high_noise_and_low_rank-4.pdf (2.83 MB)PDF icon Updated with appendix showing empirically that the main results extend to deep nonlinear networks (2.95 MB)
Galanti, T., Xu, M., Galanti, L. & Poggio, T. Norm-Based Generalization Bounds for Compositionally Sparse Neural Networks. (2023).PDF icon Norm-based bounds for convnets.pdf (1.2 MB)
Galanti, T., Xu, M., Galanti, L. & Poggio, T. Norm-based Generalization Bounds for Sparse Neural Networks. NeurIPS 2023 (2023). at <https://proceedings.neurips.cc/paper_files/paper/2023/file/8493e190ff1bbe3837eca821190b61ff-Paper-Conference.pdf>PDF icon NeurIPS-2023-norm-based-generalization-bounds-for-sparse-neural-networks-Paper-Conference.pdf (577.69 KB)
Galanti, T., Siegel, Z., Gupte, A. & Poggio, T. SGD and Weight Decay Provably Induce a Low-Rank Bias in Deep Neural Networks. (2023).PDF icon Low-rank bias.pdf (2.38 MB)
2022
Galanti, T. & Galanti, L. On the Implicit Bias Towards Minimal Depth of Deep Neural Networks. arXiv (2022). at <https://arxiv.org/abs/2202.09028>PDF icon 2202.09028.pdf (2 MB)
Galanti, T. & Poggio, T. SGD Noise and Implicit Low-Rank Bias in Deep Neural Networks. (2022).PDF icon Implicit Rank Minimization.pdf (1.76 MB)
2021
Xu, M. et al. Dynamics and Neural Collapse in Deep Classifiers trained with the Square Loss. (2021).PDF icon v1.0 (4.61 MB)PDF icon v1.4corrections to generalization section (5.85 MB)PDF icon v1.7Small edits (22.65 MB)