dc.contributor.author | Galanti, Tomer | |
dc.contributor.author | Xu, Mengjia | |
dc.contributor.author | Galanti, Liane | |
dc.contributor.author | Poggio, Tomaso | |
dc.date.accessioned | 2023-02-27T18:01:37Z | |
dc.date.available | 2023-02-27T18:01:37Z | |
dc.date.issued | 2023-02-14 | |
dc.identifier.uri | https://hdl.handle.net/1721.1/148230 | |
dc.description.abstract | In this paper, we investigate the Rademacher complexity of deep sparse neural networks, where each neuron receives a small number of inputs. We prove generalization bounds for multilayered sparse ReLU neural networks, including convolutional neural networks. These bounds differ from previous ones, as they consider the norms of the convolutional filters instead of the norms of the associated Toeplitz matrices, independently of weight sharing between neurons.
As we show theoretically, these bounds may be orders of magnitude better than standard norm- based generalization bounds and empirically, they are almost non-vacuous in estimating generalization in various simple classification problems. Taken together, these results suggest that compositional sparsity of the underlying target function is critical to the success of deep neural networks. | en_US |
dc.description.sponsorship | This material is based upon work supported by the Center for Brains, Minds and Machines (CBMM), funded by NSF STC award CCF-1231216. | en_US |
dc.publisher | Center for Brains, Minds and Machines (CBMM) | en_US |
dc.relation.ispartofseries | CBMM Memo;139 | |
dc.title | Norm-Based Generalization Bounds for Compositionally Sparse Neural Network | en_US |
dc.type | Article | en_US |
dc.type | Technical Report | en_US |
dc.type | Working Paper | en_US |