{"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/recovering-the-unbiased-scene-graphs-from-the","title":"Recovering the Unbiased Scene Graphs from the Biased Ones","arxiv_id":"2107.02112","date":"2021-07-05","proceeding":null,"authors":["Meng-Jiun Chiou","Henghui Ding","Hanshu Yan","Changhu Wang","Roger Zimmermann","Jiashi Feng"],"abstract":"Given input images, scene graph generation (SGG) aims to produce comprehensive, graphical representations describing visual relationships among salient objects. Recently, more efforts have been paid to the long tail problem in SGG; however, the imbalance in the fraction of missing labels of different classes, or reporting bias, exacerbating the long tail is rarely considered and cannot be solved by the existing debiasing methods. In this paper we show that, due to the missing labels, SGG can be viewed as a \"Learning from Positive and Unlabeled data\" (PU learning) problem, where the reporting bias can be removed by recovering the unbiased probabilities from the biased ones by utilizing label frequencies, i.e., the per-class fraction of labeled, positive examples in all the positive examples. To obtain accurate label frequency estimates, we propose Dynamic Label Frequency Estimation (DLFE) to take advantage of training-time data augmentation and average over multiple training iterations to introduce more valid examples. Extensive experiments show that DLFE is more effective in estimating label frequencies than a naive variant of the traditional estimate, and DLFE significantly alleviates the long tail and achieves state-of-the-art debiasing performance on the VG dataset. We also show qualitatively that SGG models with DLFE produce prominently more balanced and unbiased scene graphs.","url_abs":"https://arxiv.org/abs/2107.02112v1","url_pdf":"https://arxiv.org/pdf/2107.02112v1.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":"recovering-the-unbiased-scene-graphs-from-the","repo_url":"https://github.com/coldmanck/recovering-unbiased-scene-graphs","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"missing-labels","task_name":"Missing Labels"},{"task_slug":"scene-graph-classification","task_name":"Scene Graph Classification"},{"task_slug":"scene-graph-detection","task_name":"Scene Graph Detection"},{"task_slug":"scene-graph-generation","task_name":"Scene Graph Generation"},{"task_slug":"unbiased-scene-graph-generation","task_name":"Unbiased Scene Graph Generation"},{"task_slug":"visual-relationship-detection","task_name":"Visual Relationship Detection"},{"task_slug":null,"task_name":"valid"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/scene-graph-generation-on-visual-genome","task":"Scene Graph Generation","dataset":"Visual Genome","model":"DLFE","rank_in_archive_order":10,"of":19,"metrics":{"Recall@50":"25.4"},"uses_additional_data":false},{"leaderboard":"/sota/unbiased-scene-graph-generation-on-visual","task":"Unbiased Scene Graph Generation","dataset":"Visual Genome","model":"DLFE (MOTIFS-ResNeXt-101-FPN backbone; PredCls mode)","rank_in_archive_order":3,"of":31,"metrics":{"F@100":"37.6","mR@20":"22.1","ng-mR@20":"30.0"},"uses_additional_data":false},{"leaderboard":"/sota/unbiased-scene-graph-generation-on-visual","task":"Unbiased Scene Graph Generation","dataset":"Visual Genome","model":"DLFE (VCTree-ResNeXt-101-FPN backbone; PredCls mode)","rank_in_archive_order":4,"of":31,"metrics":{"F@100":"36.0","mR@20":"20.8","ng-mR@20":"29.1"},"uses_additional_data":false},{"leaderboard":"/sota/unbiased-scene-graph-generation-on-visual","task":"Unbiased Scene Graph Generation","dataset":"Visual Genome","model":"DLFE (VCTree-ResNeXt-101-FPN backbone; SGCls mode)","rank_in_archive_order":8,"of":31,"metrics":{"mR@20":"15.8","ng-mR@20":"21.6"},"uses_additional_data":false},{"leaderboard":"/sota/unbiased-scene-graph-generation-on-visual","task":"Unbiased Scene Graph Generation","dataset":"Visual Genome","model":"DLFE (MOTIFS-ResNeXt-101-FPN backbone; SGCls mode)","rank_in_archive_order":12,"of":31,"metrics":{"F@100":"21.5","mR@20":"12.8","ng-mR@20":"17.6"},"uses_additional_data":false},{"leaderboard":"/sota/unbiased-scene-graph-generation-on-visual","task":"Unbiased Scene Graph Generation","dataset":"Visual Genome","model":"DLFE (VCTree-ResNeXt-101-FPN backbone; SGDet mode)","rank_in_archive_order":17,"of":31,"metrics":{"F@100":"18.1","mR@20":"8.6","ng-mR@20":"11.8"},"uses_additional_data":false},{"leaderboard":"/sota/unbiased-scene-graph-generation-on-visual","task":"Unbiased Scene Graph Generation","dataset":"Visual Genome","model":"DLFE (MOTIFS-ResNeXt-101-FPN backbone; SGDet mode)","rank_in_archive_order":18,"of":31,"metrics":{"F@100":"18.8","mR@20":"8.6","ng-mR@20":"11.7"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2107.02112","atlas_url":"https://app.syntology.ai/?focus=2107.02112","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":"2107.02112"},"arguments_in":{"paper_2":"another paper's arXiv id or title"},"call":"get_citation_path(paper_1=\"2107.02112\", paper_2=\"…\")"},{"tool":"get_concepts_for_paper","arguments":{"arxiv_id":"2107.02112"},"call":"get_concepts_for_paper(arxiv_id=\"2107.02112\")"}],"by_repo":[{"repo":"coldmanck/recovering-unbiased-scene-graphs","kind":"official","samples":0,"ran":0,"instrument":0,"unverified":0,"harvested_for_other_papers":1,"harvested_for":["2002.11949"],"harvested_for_calls":[{"arxiv_id":"2002.11949","tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"2002.11949"},"call":"get_harvested_code_for_paper(arxiv_id=\"2002.11949\")"}]}],"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":"6275d3a1b4584cd7b35cabcfd820200041959069e6c529248bc377112fa0c3d1","record_changed_at":"2026-09-28","record_changed_at_basis":"first_hashed"}