{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/hausdorff-dimension-stochastic-differential","title":"Hausdorff Dimension, Heavy Tails, and Generalization in Neural Networks","arxiv_id":"2006.09313","date":"2020-06-16","proceeding":"NeurIPS 2020 12","authors":["Umut Şimşekli","Ozan Sener","George Deligiannidis","Murat A. Erdogdu"],"abstract":"Despite its success in a wide range of applications, characterizing the generalization properties of stochastic gradient descent (SGD) in non-convex deep learning problems is still an important challenge. While modeling the trajectories of SGD via stochastic differential equations (SDE) under heavy-tailed gradient noise has recently shed light over several peculiar characteristics of SGD, a rigorous treatment of the generalization properties of such SDEs in a learning theoretical framework is still missing. Aiming to bridge this gap, in this paper, we prove generalization bounds for SGD under the assumption that its trajectories can be well-approximated by a \\emph{Feller process}, which defines a rich class of Markov processes that include several recent SDE representations (both Brownian or heavy-tailed) as its special case. We show that the generalization error can be controlled by the \\emph{Hausdorff dimension} of the trajectories, which is intimately linked to the tail behavior of the driving process. Our results imply that heavier-tailed processes should achieve better generalization; hence, the tail-index of the process can be used as a notion of \"capacity metric\". We support our theory with experiments on deep neural networks illustrating that the proposed capacity metric accurately estimates the generalization error, and it does not necessarily grow with the number of parameters unlike the existing capacity metrics in the literature.","url_abs":"https://arxiv.org/abs/2006.09313v3","url_pdf":"https://arxiv.org/pdf/2006.09313v3.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"hausdorff-dimension-stochastic-differential","repo_url":"https://github.com/umutsimsekli/Hausdorff-Dimension-and-Generalization","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"generalization-bounds","task_name":"Generalization Bounds"}],"methods":[{"method_slug":"sgd","method_name":"SGD"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2006.09313","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2006.09313"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/umutsimsekli/Hausdorff-Dimension-and-Generalization","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"unverified":12},"by_repo_kind":{"official":{"samples":12,"ran":0,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"8516b1e3797f3626","entry":"compute_loss_accuracy","repo":"umutsimsekli/Hausdorff-Dimension-and-Generalization","repo_kind":"official","path":"src/eval_lenet.py","file_url":"https://github.com/umutsimsekli/Hausdorff-Dimension-and-Generalization/blob/HEAD/src/eval_lenet.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"8516b1e3797f3626"}},{"code_sha256_prefix":"7fba23b864d78e0e","entry":"compute_loss_accuracy","repo":"umutsimsekli/Hausdorff-Dimension-and-Generalization","repo_kind":"official","path":"src/eval_vgg.py","file_url":"https://github.com/umutsimsekli/Hausdorff-Dimension-and-Generalization/blob/HEAD/src/eval_vgg.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"7fba23b864d78e0e"}},{"code_sha256_prefix":"df1d0d9922d01dfd","entry":"get_5_good_fiv","repo":"umutsimsekli/Hausdorff-Dimension-and-Generalization","repo_kind":"official","path":"src/alpha_estimator_vgg.py","file_url":"https://github.com/umutsimsekli/Hausdorff-Dimension-and-Generalization/blob/HEAD/src/alpha_estimator_vgg.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"df1d0d9922d01dfd"}},{"code_sha256_prefix":"81aabb5979983838","entry":"get_grads","repo":"umutsimsekli/Hausdorff-Dimension-and-Generalization","repo_kind":"official","path":"src/utils.py","file_url":"https://github.com/umutsimsekli/Hausdorff-Dimension-and-Generalization/blob/HEAD/src/utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"81aabb5979983838"}},{"code_sha256_prefix":"0b07fe0f42cc6c92","entry":"get_layerWise_norms","repo":"umutsimsekli/Hausdorff-Dimension-and-Generalization","repo_kind":"official","path":"src/utils.py","file_url":"https://github.com/umutsimsekli/Hausdorff-Dimension-and-Generalization/blob/HEAD/src/utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"0b07fe0f42cc6c92"}},{"code_sha256_prefix":"7715f76656a8f363","entry":"get_ms","repo":"umutsimsekli/Hausdorff-Dimension-and-Generalization","repo_kind":"official","path":"src/alpha.py","file_url":"https://github.com/umutsimsekli/Hausdorff-Dimension-and-Generalization/blob/HEAD/src/alpha.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"7715f76656a8f363"}},{"code_sha256_prefix":"f9856a1f95a31bd1","entry":"get_ms","repo":"umutsimsekli/Hausdorff-Dimension-and-Generalization","repo_kind":"official","path":"src/alpha_estimator_vgg.py","file_url":"https://github.com/umutsimsekli/Hausdorff-Dimension-and-Generalization/blob/HEAD/src/alpha_estimator_vgg.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"f9856a1f95a31bd1"}},{"code_sha256_prefix":"edb70818f0dae2b8","entry":"linear_hinge_loss","repo":"umutsimsekli/Hausdorff-Dimension-and-Generalization","repo_kind":"official","path":"src/utils.py","file_url":"https://github.com/umutsimsekli/Hausdorff-Dimension-and-Generalization/blob/HEAD/src/utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"edb70818f0dae2b8"}},{"code_sha256_prefix":"13d5dc2899f822db","entry":"peek_model_size","repo":"umutsimsekli/Hausdorff-Dimension-and-Generalization","repo_kind":"official","path":"src/alpha.py","file_url":"https://github.com/umutsimsekli/Hausdorff-Dimension-and-Generalization/blob/HEAD/src/alpha.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"13d5dc2899f822db"}},{"code_sha256_prefix":"d62121af35bc57d8","entry":"peek_model_size","repo":"umutsimsekli/Hausdorff-Dimension-and-Generalization","repo_kind":"official","path":"src/alpha_estimator_vgg.py","file_url":"https://github.com/umutsimsekli/Hausdorff-Dimension-and-Generalization/blob/HEAD/src/alpha_estimator_vgg.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"d62121af35bc57d8"}},{"code_sha256_prefix":"df5d0650abc8d23f","entry":"test_on_cifar","repo":"umutsimsekli/Hausdorff-Dimension-and-Generalization","repo_kind":"official","path":"src/eval_lenet.py","file_url":"https://github.com/umutsimsekli/Hausdorff-Dimension-and-Generalization/blob/HEAD/src/eval_lenet.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"df5d0650abc8d23f"}},{"code_sha256_prefix":"8b6ff57568a221d1","entry":"test_on_cifar","repo":"umutsimsekli/Hausdorff-Dimension-and-Generalization","repo_kind":"official","path":"src/eval_vgg.py","file_url":"https://github.com/umutsimsekli/Hausdorff-Dimension-and-Generalization/blob/HEAD/src/eval_vgg.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"8b6ff57568a221d1"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}