{"url":"/task/metric-learning","name":"Metric Learning","slug":"metric-learning","description_markdown":"The goal of **Metric Learning** is to learn a representation function that maps objects into an embedded space. The distance in the embedded space should preserve the objects’ similarity — similar objects get close and dissimilar objects get far away. Various loss functions have been developed for Metric Learning. For example, the **contrastive loss** guides the objects from the same class to be mapped to the same point and those from different classes to be mapped to different points whose distances are larger than a margin. **Triplet loss** is also popular, which requires the distance between the anchor sample and the positive sample to be smaller than the distance between the anchor sample and the negative sample.\r\n\r\n\r\n<span class=\"description-source\">Source: [Road Network Metric Learning for Estimated Time of Arrival ](https://arxiv.org/abs/2006.13477)</span>","categories":[{"name":"Computer Vision","url":"/area/computer-vision"},{"name":"Methodology","url":"/area/methodology"}],"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","slug_source":"archive_url"},"counts":{"papers_tagged":1648,"papers_with_code":613,"benchmarks":8,"benchmark_tables_in_archive":8,"benchmark_tables_shown":8,"benchmark_tables_withheld_as_spam":0,"benchmark_definition":"a leaderboard table with at least one row; benchmark_tables_shown also counts the zero-row tables; benchmark_tables_in_archive adds the tables withheld as spam","datasets":33,"subtasks":0,"parent_tasks":0},"benchmarks":[{"leaderboard":"/sota/metric-learning-on-cars196","slug":"metric-learning-on-cars196","dataset":"CARS196","dataset_url":"/dataset/cars196","rows_in_archive":36,"metrics":["R@1"],"first_row_in_archive_order":{"model":"Unicom+ViT-L@336px","paper_title":"Unicom: Universal and Compact Representation Learning for Image Retrieval","paper_url":"/paper/unicom-universal-and-compact-representation","paper_date":"2023-04-12","arxiv_id":"2304.05884","code_links":[{"title":"OML-Team/open-metric-learning","url":"https://github.com/OML-Team/open-metric-learning"},{"title":"deepglint/unicom","url":"https://github.com/deepglint/unicom"},{"title":"RocketFlash/easy_metric_learning","url":"https://github.com/RocketFlash/easy_metric_learning/tree/master/tools"}],"syntology":{"n":6,"n_ran":3,"n_unverified":3,"n_pointer_only":6}}},{"leaderboard":"/sota/metric-learning-on-stanford-online-products-1","slug":"metric-learning-on-stanford-online-products-1","dataset":"Stanford Online Products","dataset_url":"/dataset/stanford-online-products","rows_in_archive":33,"metrics":["R@1"],"first_row_in_archive_order":{"model":"Unicom+ViT-L@336px","paper_title":"Unicom: Universal and Compact Representation Learning for Image Retrieval","paper_url":"/paper/unicom-universal-and-compact-representation","paper_date":"2023-04-12","arxiv_id":"2304.05884","code_links":[{"title":"OML-Team/open-metric-learning","url":"https://github.com/OML-Team/open-metric-learning"},{"title":"deepglint/unicom","url":"https://github.com/deepglint/unicom"},{"title":"RocketFlash/easy_metric_learning","url":"https://github.com/RocketFlash/easy_metric_learning/tree/master/tools"}],"syntology":{"n":6,"n_ran":3,"n_unverified":3,"n_pointer_only":6}}},{"leaderboard":"/sota/metric-learning-on-cub-200-2011","slug":"metric-learning-on-cub-200-2011","dataset":"CUB-200-2011","dataset_url":"/dataset/cub-200-2011","rows_in_archive":30,"metrics":["R@1"],"first_row_in_archive_order":{"model":"Unicom+ViT-L@336px","paper_title":"Unicom: Universal and Compact Representation Learning for Image Retrieval","paper_url":"/paper/unicom-universal-and-compact-representation","paper_date":"2023-04-12","arxiv_id":"2304.05884","code_links":[{"title":"OML-Team/open-metric-learning","url":"https://github.com/OML-Team/open-metric-learning"},{"title":"deepglint/unicom","url":"https://github.com/deepglint/unicom"},{"title":"RocketFlash/easy_metric_learning","url":"https://github.com/RocketFlash/easy_metric_learning/tree/master/tools"}],"syntology":{"n":6,"n_ran":3,"n_unverified":3,"n_pointer_only":6}}},{"leaderboard":"/sota/metric-learning-on-in-shop-1","slug":"metric-learning-on-in-shop-1","dataset":"In-Shop","dataset_url":"/dataset/in-shop","rows_in_archive":15,"metrics":["R@1"],"first_row_in_archive_order":{"model":"Unicom+ViT-L@336px","paper_title":"Unicom: Universal and Compact Representation Learning for Image Retrieval","paper_url":"/paper/unicom-universal-and-compact-representation","paper_date":"2023-04-12","arxiv_id":"2304.05884","code_links":[{"title":"OML-Team/open-metric-learning","url":"https://github.com/OML-Team/open-metric-learning"},{"title":"deepglint/unicom","url":"https://github.com/deepglint/unicom"},{"title":"RocketFlash/easy_metric_learning","url":"https://github.com/RocketFlash/easy_metric_learning/tree/master/tools"}],"syntology":{"n":6,"n_ran":3,"n_unverified":3,"n_pointer_only":6}}},{"leaderboard":"/sota/metric-learning-on-cub-200-2011-4","slug":"metric-learning-on-cub-200-2011-4","dataset":"CUB-200-2011","dataset_url":"/dataset/cub-200-2011","rows_in_archive":2,"metrics":["R@1"],"first_row_in_archive_order":{"model":"Hyp-DINO","paper_title":"Hyperbolic Vision Transformers: Combining Improvements in Metric Learning","paper_url":"/paper/hyperbolic-vision-transformers-combining","paper_date":"2022-03-21","arxiv_id":"2203.10833","code_links":[{"title":"OML-Team/open-metric-learning","url":"https://github.com/OML-Team/open-metric-learning"},{"title":"htdt/hyp_metric","url":"https://github.com/htdt/hyp_metric"}],"syntology":{"n":6,"n_ran":4,"n_unverified":2,"n_pointer_only":0}}},{"leaderboard":"/sota/metric-learning-on-dyml-animal","slug":"metric-learning-on-dyml-animal","dataset":"DyML-Animal","dataset_url":"/dataset/dyml-animal","rows_in_archive":2,"metrics":["Average-mAP"],"first_row_in_archive_order":{"model":"HAPPIER","paper_title":"Hierarchical Average Precision Training for Pertinent Image Retrieval","paper_url":"/paper/hierarchical-average-precision-training-for","paper_date":"2022-07-05","arxiv_id":"2207.04873","code_links":[{"title":"elias-ramzi/happier","url":"https://github.com/elias-ramzi/happier"},{"title":"elias-ramzi/suprank","url":"https://github.com/elias-ramzi/suprank"}],"syntology":{"n":9,"n_ran":6,"n_unverified":3,"n_pointer_only":0}}},{"leaderboard":"/sota/metric-learning-on-dyml-product","slug":"metric-learning-on-dyml-product","dataset":"DyML-Product","dataset_url":"/dataset/dyml-product","rows_in_archive":2,"metrics":["Average-mAP"],"first_row_in_archive_order":{"model":"HAPPIER","paper_title":"Hierarchical Average Precision Training for Pertinent Image Retrieval","paper_url":"/paper/hierarchical-average-precision-training-for","paper_date":"2022-07-05","arxiv_id":"2207.04873","code_links":[{"title":"elias-ramzi/happier","url":"https://github.com/elias-ramzi/happier"},{"title":"elias-ramzi/suprank","url":"https://github.com/elias-ramzi/suprank"}],"syntology":{"n":9,"n_ran":6,"n_unverified":3,"n_pointer_only":0}}},{"leaderboard":"/sota/metric-learning-on-dyml-vehicle","slug":"metric-learning-on-dyml-vehicle","dataset":"DyML-Vehicle","dataset_url":"/dataset/dyml-vehicle","rows_in_archive":2,"metrics":["Average-mAP"],"first_row_in_archive_order":{"model":"HAPPIER","paper_title":"Hierarchical Average Precision Training for Pertinent Image Retrieval","paper_url":"/paper/hierarchical-average-precision-training-for","paper_date":"2022-07-05","arxiv_id":"2207.04873","code_links":[{"title":"elias-ramzi/happier","url":"https://github.com/elias-ramzi/happier"},{"title":"elias-ramzi/suprank","url":"https://github.com/elias-ramzi/suprank"}],"syntology":{"n":9,"n_ran":6,"n_unverified":3,"n_pointer_only":0}}}],"datasets":[{"url":"/dataset/cub-200-2011","name":"CUB-200-2011","full_name":"Caltech-UCSD Birds-200-2011","num_papers_in_archive":2235},{"url":"/dataset/casia-webface","name":"CASIA-WebFace","full_name":"","num_papers_in_archive":415},{"url":"/dataset/stanford-online-products","name":"Stanford Online Products","full_name":"Stanford Online Products","num_papers_in_archive":231},{"url":"/dataset/in-shop","name":"In-Shop","full_name":"In-shop Clothes Retrieval Benchmark","num_papers_in_archive":154},{"url":"/dataset/viper","name":"VIPeR","full_name":"Viewpoint Invariant Pedestrian Recognition","num_papers_in_archive":134},{"url":"/dataset/structured3d","name":"Structured3D","full_name":"","num_papers_in_archive":85},{"url":"/dataset/abo","name":"ABO","full_name":"Amazon Berkeley Objects","num_papers_in_archive":82},{"url":"/dataset/wilddeepfake","name":"WildDeepfake","full_name":"","num_papers_in_archive":49},{"url":"/dataset/cars196","name":"CARS196","full_name":"","num_papers_in_archive":43},{"url":"/dataset/veri-wild","name":"VeRi-Wild","full_name":"VeRi-Wild","num_papers_in_archive":40},{"url":"/dataset/google-landmarks-dataset-v2","name":"Google Landmarks Dataset v2","full_name":null,"num_papers_in_archive":35},{"url":"/dataset/cometa","name":"COMETA","full_name":"","num_papers_in_archive":21},{"url":"/dataset/moviegraphs","name":"MovieGraphs","full_name":"","num_papers_in_archive":13},{"url":"/dataset/airport","name":"Airport","full_name":"Airport","num_papers_in_archive":8},{"url":"/dataset/hotels-50k","name":"Hotels-50K","full_name":"","num_papers_in_archive":7},{"url":"/dataset/toplogo-10","name":"TopLogo-10","full_name":"","num_papers_in_archive":6},{"url":"/dataset/n-digit-mnist","name":"N-Digit MNIST","full_name":"","num_papers_in_archive":5},{"url":"/dataset/birdvox-full-night","name":"BirdVox-full-night","full_name":"BirdVox-full-night","num_papers_in_archive":3},{"url":"/dataset/goldfinch","name":"Goldfinch","full_name":"GOogLe image-search Dataset","num_papers_in_archive":3},{"url":"/dataset/large-age-gap","name":"Large Age-Gap","full_name":"Large Age-Gap","num_papers_in_archive":3},{"url":"/dataset/opereid","name":"OpeReid","full_name":"OpeReid","num_papers_in_archive":3},{"url":"/dataset/dyml-animal","name":"DyML-Animal","full_name":"Dynamic Metric Learning Animal","num_papers_in_archive":2},{"url":"/dataset/dyml-product","name":"DyML-Product","full_name":"Dynamic Metric Learning Product","num_papers_in_archive":2},{"url":"/dataset/dyml-vehicle","name":"DyML-Vehicle","full_name":"Dynamic Metric Learning Vehicle","num_papers_in_archive":2},{"url":"/dataset/freiburg-spatial-relations","name":"Freiburg Spatial Relations","full_name":"Freiburg Spatial Relations","num_papers_in_archive":2},{"url":"/dataset/imemnet","name":"IMEMNET","full_name":"Image-MusicEmotion-Matching-Net","num_papers_in_archive":2},{"url":"/dataset/live-comment-dataset","name":"Live Comment Dataset","full_name":"","num_papers_in_archive":2},{"url":"/dataset/videoforensicshq","name":"VideoForensicsHQ","full_name":"","num_papers_in_archive":2},{"url":"/dataset/a-view-from-somewhere-avfs","name":"A View From Somewhere (AVFS)","full_name":"","num_papers_in_archive":1},{"url":"/dataset/ham","name":"HAM","full_name":"Human-annotated Mappings","num_papers_in_archive":1},{"url":"/dataset/luna-1","name":"Luna-1","full_name":"","num_papers_in_archive":1},{"url":"/dataset/s3o4d","name":"S3O4D","full_name":"Stanford 3D Objects for Disentangling","num_papers_in_archive":1},{"url":"/dataset/tsinghua-dogs","name":"Tsinghua Dogs","full_name":"","num_papers_in_archive":1}],"subtasks":[],"parent_tasks":[],"papers":{"order":"repositories listed in the archive (desc), then date (desc); the archive holds no stars","population":"papers tagged with this task that list at least one repository in the archive","shown":30,"of":613,"tagged_in_all":1648,"items":[{"url":"/paper/in-defense-of-the-triplet-loss-for-person-re","title":"In Defense of the Triplet Loss for Person Re-Identification","date":"2017-03-22","arxiv_id":"1703.07737","repositories_listed":31,"syntology":{"n":46,"n_ran":6,"n_unverified":40,"n_pointer_only":0}},{"url":"/paper/matching-networks-for-one-shot-learning","title":"Matching Networks for One Shot Learning","date":"2016-06-13","arxiv_id":"1606.04080","repositories_listed":26,"syntology":{"n":16,"n_ran":6,"n_unverified":10,"n_pointer_only":6}},{"url":"/paper/circle-loss-a-unified-perspective-of-pair","title":"Circle Loss: A Unified Perspective of Pair Similarity Optimization","date":"2020-02-25","arxiv_id":"2002.10857","repositories_listed":16,"syntology":{"n":2,"n_ran":2,"n_unverified":0,"n_pointer_only":1}},{"url":"/paper/additive-margin-softmax-for-face-verification","title":"Additive Margin Softmax for Face Verification","date":"2018-01-17","arxiv_id":"1801.05599","repositories_listed":10,"syntology":{"n":1,"n_ran":0,"n_unverified":1,"n_pointer_only":1}},{"url":"/paper/revisiting-training-strategies-and","title":"Revisiting Training Strategies and Generalization Performance in Deep Metric Learning","date":"2020-02-19","arxiv_id":"2002.08473","repositories_listed":8,"syntology":{"n":22,"n_ran":3,"n_unverified":19,"n_pointer_only":1}},{"url":"/paper/semantic-instance-segmentation-with-a","title":"Semantic Instance Segmentation with a Discriminative Loss Function","date":"2017-08-08","arxiv_id":"1708.02551","repositories_listed":8,"syntology":{"n":2,"n_ran":0,"n_unverified":2,"n_pointer_only":2}},{"url":"/paper/deep-learning-for-person-re-identification-a","title":"Deep Learning for Person Re-identification: A Survey and Outlook","date":"2020-01-13","arxiv_id":"2001.04193","repositories_listed":7,"syntology":{"n":22,"n_ran":8,"n_unverified":14,"n_pointer_only":3}},{"url":"/paper/time-contrastive-networks-self-supervised","title":"Time-Contrastive Networks: Self-Supervised Learning from Video","date":"2017-04-23","arxiv_id":"1704.06888","repositories_listed":7,"syntology":null},{"url":"/paper/metric-learn-metric-learning-algorithms-in","title":"metric-learn: Metric Learning Algorithms in Python","date":"2019-08-13","arxiv_id":"1908.04710","repositories_listed":6,"syntology":null},{"url":"/paper/sampling-matters-in-deep-embedding-learning","title":"Sampling Matters in Deep Embedding Learning","date":"2017-06-23","arxiv_id":"1706.07567","repositories_listed":6,"syntology":{"n":2,"n_ran":2,"n_unverified":0,"n_pointer_only":0}},{"url":"/paper/exploring-cross-image-pixel-contrast-for","title":"Exploring Cross-Image Pixel Contrast for Semantic Segmentation","date":"2021-01-28","arxiv_id":"2101.11939","repositories_listed":5,"syntology":{"n":1,"n_ran":0,"n_unverified":1,"n_pointer_only":0}},{"url":"/paper/cross-batch-memory-for-embedding-learning","title":"Cross-Batch Memory for Embedding Learning","date":"2019-12-14","arxiv_id":"1912.06798","repositories_listed":5,"syntology":null},{"url":"/paper/softtriple-loss-deep-metric-learning-without","title":"SoftTriple Loss: Deep Metric Learning Without Triplet Sampling","date":"2019-09-11","arxiv_id":"1909.05235","repositories_listed":5,"syntology":{"n":4,"n_ran":2,"n_unverified":2,"n_pointer_only":2}},{"url":"/paper/panet-few-shot-image-semantic-segmentation","title":"PANet: Few-Shot Image Semantic Segmentation with Prototype Alignment","date":"2019-08-18","arxiv_id":"1908.06391","repositories_listed":5,"syntology":null},{"url":"/paper/deep-cosine-metric-learning-for-person-re","title":"Deep Cosine Metric Learning for Person Re-Identification","date":"2018-12-02","arxiv_id":"1812.00442","repositories_listed":5,"syntology":{"n":10,"n_ran":0,"n_unverified":10,"n_pointer_only":0}},{"url":"/paper/batch-feature-erasing-for-person-re","title":"Batch DropBlock Network for Person Re-identification and Beyond","date":"2018-11-17","arxiv_id":"1811.07130","repositories_listed":5,"syntology":{"n":12,"n_ran":3,"n_unverified":9,"n_pointer_only":0}},{"url":"/paper/a-closer-look-at-the-training-dynamics-of","title":"Understanding the Role of the Projector in Knowledge Distillation","date":"2023-03-20","arxiv_id":"2303.11098","repositories_listed":4,"syntology":null},{"url":"/paper/nbc-softmax-darkweb-author-fingerprinting-and","title":"NBC-Softmax : Darkweb Author fingerprinting and migration tracking","date":"2022-12-15","arxiv_id":"2212.08184","repositories_listed":4,"syntology":null},{"url":"/paper/a-metric-learning-reality-check","title":"A Metric Learning Reality Check","date":"2020-03-18","arxiv_id":"2003.08505","repositories_listed":4,"syntology":{"n":2,"n_ran":2,"n_unverified":0,"n_pointer_only":2}},{"url":"/paper/unicom-universal-and-compact-representation","title":"Unicom: Universal and Compact Representation Learning for Image Retrieval","date":"2023-04-12","arxiv_id":"2304.05884","repositories_listed":3,"syntology":{"n":6,"n_ran":3,"n_unverified":3,"n_pointer_only":6}},{"url":"/paper/equiangular-basis-vectors","title":"Equiangular Basis Vectors","date":"2023-03-21","arxiv_id":"2303.11637","repositories_listed":3,"syntology":null},{"url":"/paper/a-deep-metric-learning-approach-to-account","title":"A Deep Metric Learning Approach to Account Linking","date":"2021-05-15","arxiv_id":"2105.07263","repositories_listed":3,"syntology":{"n":6,"n_ran":0,"n_unverified":6,"n_pointer_only":0}},{"url":"/paper/distribution-free-risk-controlling-prediction","title":"Distribution-Free, Risk-Controlling Prediction Sets","date":"2021-01-07","arxiv_id":"2101.02703","repositories_listed":3,"syntology":{"n":3,"n_ran":3,"n_unverified":0,"n_pointer_only":0}},{"url":"/paper/channel-augmented-joint-learning-for-visible","title":"Channel Augmented Joint Learning for Visible-Infrared Recognition","date":"2021-01-01","arxiv_id":null,"repositories_listed":3,"syntology":null},{"url":"/paper/quasi-dense-instance-similarity-learning","title":"Quasi-Dense Similarity Learning for Multiple Object Tracking","date":"2020-06-11","arxiv_id":"2006.06664","repositories_listed":3,"syntology":null},{"url":"/paper/proxy-anchor-loss-for-deep-metric-learning","title":"Proxy Anchor Loss for Deep Metric Learning","date":"2020-03-31","arxiv_id":"2003.13911","repositories_listed":3,"syntology":null},{"url":"/paper/event-sequence-metric-learning","title":"CoLES: Contrastive Learning for Event Sequences with Self-Supervision","date":"2020-02-19","arxiv_id":"2002.08232","repositories_listed":3,"syntology":null},{"url":"/paper/cross-view-kernel-similarity-metric-learning","title":"Cross-View Kernel Similarity Metric Learning Using Pairwise Constraints for Person Re-identification","date":"2019-09-25","arxiv_id":"1909.11316","repositories_listed":3,"syntology":null},{"url":"/paper/relational-knowledge-distillation","title":"Relational Knowledge Distillation","date":"2019-04-10","arxiv_id":"1904.05068","repositories_listed":3,"syntology":{"n":3,"n_ran":1,"n_unverified":2,"n_pointer_only":1}},{"url":"/paper/holistic-and-comprehensive-annotation-of","title":"Holistic and Comprehensive Annotation of Clinically Significant Findings on Diverse CT Images: Learning from Radiology Reports and Label Ontology","date":"2019-04-09","arxiv_id":"1904.04661","repositories_listed":3,"syntology":null}],"syntology_records":17,"syntology_note":"a paper without a record is not a recorded non-run: it may lack an arXiv id or simply be absent from the graph layer"},"description_links":{"kept":0,"unwrapped_to_text":0,"bare_urls_linked":0,"relative_images_dropped":0,"rule":"internal links are kept only when the target slug exists in the catalog"},"syntology":{"read_at":"2026-09-24T18:15:14+00:00","claim":"Per-sample execution status on synthesized fixtures ('ran N of M samples'); not a correctness claim and not a ranking signal.","status_vocabulary":{"ran_honours":"ran, honoured the contract we drafted","ran_violates":"ran, violated the contract we drafted","ran_draft_wrong":"ran; our contract draft was wrong, not the code","ran_fixture":"ran; our fixture could not drive it","ran":"ran on a synthesized input","unverified":"unverified (harvested, no recorded run)"}},"not_shown":{"libraries":"the archive has no per-task library table","trend_sparklines":"the Trend column of the benchmarks table was a rendered image; it is not in the archive","social_and_latest_sorts":"stars and social signals are not in the archive"}}