{"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/escaping-saddle-points-for-effective","title":"Escaping Saddle Points for Effective Generalization on Class-Imbalanced Data","arxiv_id":"2212.13827","date":"2022-12-28","proceeding":null,"authors":["Harsh Rangwani","Sumukh K Aithal","Mayank Mishra","R. Venkatesh Babu"],"abstract":"Real-world datasets exhibit imbalances of varying types and degrees. Several techniques based on re-weighting and margin adjustment of loss are often used to enhance the performance of neural networks, particularly on minority classes. In this work, we analyze the class-imbalanced learning problem by examining the loss landscape of neural networks trained with re-weighting and margin-based techniques. Specifically, we examine the spectral density of Hessian of class-wise loss, through which we observe that the network weights converge to a saddle point in the loss landscapes of minority classes. Following this observation, we also find that optimization methods designed to escape from saddle points can be effectively used to improve generalization on minority classes. We further theoretically and empirically demonstrate that Sharpness-Aware Minimization (SAM), a recent technique that encourages convergence to a flat minima, can be effectively used to escape saddle points for minority classes. Using SAM results in a 6.2\\% increase in accuracy on the minority classes over the state-of-the-art Vector Scaling Loss, leading to an overall average increase of 4\\% across imbalanced datasets. The code is available at: https://github.com/val-iisc/Saddle-LongTail.","url_abs":"https://arxiv.org/abs/2212.13827v1","url_pdf":"https://arxiv.org/pdf/2212.13827v1.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":"escaping-saddle-points-for-effective","repo_url":"https://github.com/val-iisc/saddle-longtail","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"long-tail-learning","task_name":"Long-tail Learning"}],"methods":[{"method_slug":"sharpness-aware-minimization","method_name":"Sharpness-Aware Minimization"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/long-tail-learning-on-cifar-10-lt-r-10","task":"Long-tail Learning","dataset":"CIFAR-10-LT (ρ=10)","model":"LDAM + DRW + SAM","rank_in_archive_order":28,"of":50,"metrics":{"Error Rate":"10.6"},"uses_additional_data":false},{"leaderboard":"/sota/long-tail-learning-on-cifar-10-lt-r-100","task":"Long-tail Learning","dataset":"CIFAR-10-LT (ρ=100)","model":"GLMC + SAM","rank_in_archive_order":2,"of":28,"metrics":{"Error Rate":"10.82"},"uses_additional_data":false},{"leaderboard":"/sota/long-tail-learning-on-cifar-10-lt-r-100","task":"Long-tail Learning","dataset":"CIFAR-10-LT (ρ=100)","model":"VS + SAM","rank_in_archive_order":17,"of":28,"metrics":{"Error Rate":"17.6"},"uses_additional_data":false},{"leaderboard":"/sota/long-tail-learning-on-cifar-10-lt-r-200","task":"Long-tail Learning","dataset":"CIFAR-10-LT (ρ=200)","model":"LDAM + DRW + SAM","rank_in_archive_order":1,"of":2,"metrics":{"Error Rate":"21.9"},"uses_additional_data":false},{"leaderboard":"/sota/long-tail-learning-on-cifar-10-lt-r-50","task":"Long-tail Learning","dataset":"CIFAR-10-LT (ρ=50)","model":"GLMC + SAM","rank_in_archive_order":1,"of":8,"metrics":{"Error Rate":"8.44"},"uses_additional_data":false},{"leaderboard":"/sota/long-tail-learning-on-cifar-100-lt-r-100","task":"Long-tail Learning","dataset":"CIFAR-100-LT (ρ=100)","model":"GLMC + SAM","rank_in_archive_order":6,"of":66,"metrics":{"Error Rate":"40.99"},"uses_additional_data":false},{"leaderboard":"/sota/long-tail-learning-on-cifar-100-lt-r-100","task":"Long-tail Learning","dataset":"CIFAR-100-LT (ρ=100)","model":"PaCo + SAM","rank_in_archive_order":21,"of":66,"metrics":{"Error Rate":"47.0"},"uses_additional_data":false},{"leaderboard":"/sota/long-tail-learning-on-cifar-100-lt-r-100","task":"Long-tail Learning","dataset":"CIFAR-100-LT (ρ=100)","model":"VS + SAM","rank_in_archive_order":41,"of":66,"metrics":{"Error Rate":"53.4"},"uses_additional_data":false},{"leaderboard":"/sota/long-tail-learning-on-cifar-100-lt-r-200","task":"Long-tail Learning","dataset":"CIFAR-100-LT (ρ=200)","model":"PaCo + SAM","rank_in_archive_order":1,"of":2,"metrics":{"Error Rate":"52.0"},"uses_additional_data":false},{"leaderboard":"/sota/long-tail-learning-on-cifar-100-lt-r-50","task":"Long-tail Learning","dataset":"CIFAR-100-LT (ρ=50)","model":"GLMC + SAM","rank_in_archive_order":5,"of":25,"metrics":{"Error Rate":"34.72"},"uses_additional_data":false},{"leaderboard":"/sota/long-tail-learning-on-imagenet-lt","task":"Long-tail Learning","dataset":"ImageNet-LT","model":"LDAM + DRW + SAM","rank_in_archive_order":45,"of":69,"metrics":{"Top-1 Accuracy":"53.1"},"uses_additional_data":false},{"leaderboard":"/sota/long-tail-learning-on-inaturalist-2018","task":"Long-tail Learning","dataset":"iNaturalist 2018","model":"LDAM + DRW + SAM","rank_in_archive_order":34,"of":43,"metrics":{"Top-1 Accuracy":"70.1"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2212.13827","atlas_url":"https://app.syntology.ai/?focus=2212.13827","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":"2212.13827"},"arguments_in":{"paper_2":"another paper's arXiv id or title"},"call":"get_citation_path(paper_1=\"2212.13827\", paper_2=\"…\")"},{"tool":"get_concepts_for_paper","arguments":{"arxiv_id":"2212.13827"},"call":"get_concepts_for_paper(arxiv_id=\"2212.13827\")"}],"by_repo":[{"repo":"val-iisc/saddle-longtail","kind":"official","samples":0,"ran":0,"instrument":0,"unverified":0,"harvested_for_other_papers":1,"harvested_for":["2305.08661"],"harvested_for_calls":[{"arxiv_id":"2305.08661","tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"2305.08661"},"call":"get_harvested_code_for_paper(arxiv_id=\"2305.08661\")"}]}],"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":"3e5a1a05aac4c08649decf19e9d940883c7799b794c9629c00819205edd2de47","record_changed_at":"2026-09-28","record_changed_at_basis":"first_hashed"}