{"about":{"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.","site":"https://codewithpapers.app","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","syntology":{"site":"https://syntology.ai","developers":"https://syntology.ai/developers","mcp":{"server":"https://syntology.ai/mcp","transport":"streamable-http","server_card":"https://syntology.ai/.well-known/mcp/server-card.json","auth":{"type":"trial token, no account","trial_token":"https://syntology.ai/api/oauth/trial/token","method":"POST","docs":"https://syntology.ai/developers"}},"have":"https://syntology.ai/api/graph/have?x=<method, arXiv id or title> (free, answers coverage only)","paper_base":"https://syntology.ai/paper/","atlas_base":"https://app.syntology.ai/?focus="},"machine_readable":[{"url":"https://codewithpapers.app/llms.txt","what":"the machine catalog: every machine-readable file, counted"},{"url":"https://codewithpapers.app/index/manifest.json","what":"paper-to-code index by arXiv id, with Syntology's counts"},{"url":"https://codewithpapers.app/search/manifest.json","what":"site search index (titles, authors) and its files"},{"url":"https://codewithpapers.app/download","what":"bulk files: Syntology's layer, described there"},{"url":"https://codewithpapers.app/build_manifest.json","what":"the build record: inputs, counts, exclusions, probes"}]},"url":"/paper/understanding-deep-learning-requires","title":"Understanding deep learning requires rethinking generalization","arxiv_id":"1611.03530","date":"2016-11-10","proceeding":null,"authors":["Chiyuan Zhang","Samy Bengio","Moritz Hardt","Benjamin Recht","Oriol Vinyals"],"abstract":"Despite their massive size, successful deep artificial neural networks can\nexhibit a remarkably small difference between training and test performance.\nConventional wisdom attributes small generalization error either to properties\nof the model family, or to the regularization techniques used during training.\n  Through extensive systematic experiments, we show how these traditional\napproaches fail to explain why large neural networks generalize well in\npractice. Specifically, our experiments establish that state-of-the-art\nconvolutional networks for image classification trained with stochastic\ngradient methods easily fit a random labeling of the training data. This\nphenomenon is qualitatively unaffected by explicit regularization, and occurs\neven if we replace the true images by completely unstructured random noise. We\ncorroborate these experimental findings with a theoretical construction showing\nthat simple depth two neural networks already have perfect finite sample\nexpressivity as soon as the number of parameters exceeds the number of data\npoints as it usually does in practice.\n  We interpret our experimental findings by comparison with traditional models.","url_abs":"http://arxiv.org/abs/1611.03530v2","url_pdf":"http://arxiv.org/pdf/1611.03530v2.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":"understanding-deep-learning-requires","repo_url":"https://github.com/2xic/notebooks","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"understanding-deep-learning-requires","repo_url":"https://github.com/KellyHwong/rethinking_generalization","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"understanding-deep-learning-requires","repo_url":"https://github.com/aaronpeikert/methods-meetup","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"Unlicense"}},{"paper_slug":"understanding-deep-learning-requires","repo_url":"https://github.com/glouppe/info8010-deep-learning","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"understanding-deep-learning-requires","repo_url":"https://github.com/jessemzhang/dl_spectral_normalization","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"understanding-deep-learning-requires","repo_url":"https://github.com/mdv3101/Rethinking_Generalization","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"understanding-deep-learning-requires","repo_url":"https://github.com/pluskid/fitting-random-labels","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"deep-learning","task_name":"Deep Learning"},{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"image-classification","task_name":"image-classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1611.03530","atlas_url":"https://app.syntology.ai/?focus=1611.03530","mcp":null,"developers":"https://syntology.ai/developers","held":true,"reason":"arxiv_id_in_atlas","checked_against":{"atlas_ids":549449,"atlas_ids_sha256":"afd72399dafd9906e9062927f70655952385a3d75488558e1f0484e950ab09b7","read_at":"2026-10-05T01:46:21Z"},"agent_calls":[{"tool":"get_citation_path","arguments":{"paper_1":"1611.03530"},"arguments_in":{"paper_2":"another paper's arXiv id or title"},"call":"get_citation_path(paper_1=\"1611.03530\", paper_2=\"…\")"},{"tool":"get_concepts_for_paper","arguments":{"arxiv_id":"1611.03530"},"call":"get_concepts_for_paper(arxiv_id=\"1611.03530\")"}],"by_repo":[{"repo":"2xic/notebooks","kind":"listed","samples":0,"ran":0,"instrument":0,"unverified":0},{"repo":"KellyHwong/rethinking_generalization","kind":"listed","samples":0,"ran":0,"instrument":0,"unverified":0,"harvested_for_other_papers":1,"harvested_for":["1512.03385"],"harvested_for_calls":[{"arxiv_id":"1512.03385","tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"1512.03385"},"call":"get_harvested_code_for_paper(arxiv_id=\"1512.03385\")"}]},{"repo":"aaronpeikert/methods-meetup","kind":"listed","samples":0,"ran":0,"instrument":0,"unverified":0},{"repo":"glouppe/info8010-deep-learning","kind":"listed","samples":0,"ran":0,"instrument":0,"unverified":0,"harvested_for_other_papers":1,"harvested_for":["1706.03762"],"harvested_for_calls":[{"arxiv_id":"1706.03762","tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"1706.03762"},"call":"get_harvested_code_for_paper(arxiv_id=\"1706.03762\")"}]},{"repo":"jessemzhang/dl_spectral_normalization","kind":"listed","samples":0,"ran":0,"instrument":0,"unverified":0},{"repo":"mdv3101/Rethinking_Generalization","kind":"listed","samples":0,"ran":0,"instrument":0,"unverified":0},{"repo":"pluskid/fitting-random-labels","kind":"listed","samples":0,"ran":0,"instrument":0,"unverified":0,"harvested_for_other_papers":1,"harvested_for":["2203.08390"],"harvested_for_calls":[{"arxiv_id":"2203.08390","tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"2203.08390"},"call":"get_harvested_code_for_paper(arxiv_id=\"2203.08390\")"}]}],"by_repo_harvested_for_is":"harvested_for: the arXiv ids (as Syntology's graph holds them) of every paper whose samples Syntology's graph links to this repository, in arXiv-id order, complete; the page names the first 3 and says how many more. A link only: it does not say the papers are the same work, that one copied the other, or that the code implements this paper's method; harvested_for_calls gives the get_harvested_code_for_paper call for each arXiv-shaped id among them","by_repo_is":"one row per repository in the archive's code links, Syntology's graph links or the samples; samples 0 means no sample is linked to this paper, a statement about Syntology's coverage, not about the repository: with harvested_for_other_papers present, Syntology harvested the repository for that many other papers, and without it, Syntology harvested nothing from it; placed_by_identical_code counts samples whose link records no repository, placed here because identical code was harvested from this repository for another paper; ran counts executions on a synthesized input and instrument counts failures of Syntology's instrument, not of the code"},"code_links_note":null,"arxiv_metadata":null,"syntology_extracted_results":null,"record_sha256":"9e4aa9fe70a47056bad37652ab3fa77ed9ad5cd094c44b368e6e399c8bdb5d4a","record_changed_at":"2026-09-28","record_changed_at_basis":"first_hashed"}