{"url":"/sota/collaborative-filtering-on-movielens-20m","task":{"name":"Recommendation Systems","url":"/task/recommendation-systems","note":null},"dataset":{"name":"MovieLens 20M","url":"/dataset/movielens"},"category":"Miscellaneous","categories":["Graphs","Knowledge Base","Miscellaneous"],"category_note":null,"description":"### **Recommendation System in AI Research**  \r\n\r\nA **Recommendation System** is a specialized AI-driven model that analyzes user preferences and behaviors to suggest relevant content, products, or services. It is widely used in domains like e-commerce, streaming platforms, social media, and personalized learning.  \r\n\r\nAI research in recommendation systems focuses on:  \r\n- **Collaborative Filtering**: Predicting user preferences based on similar users' choices.  \r\n- **Content-Based Filtering**: Recommending items based on user history and item characteristics.  \r\n- **Hybrid Models**: Combining multiple techniques for better accuracy.  \r\n- **Deep Learning & Transformers**: Using neural networks and self-attention mechanisms for personalized recommendations.  \r\n- **Graph-Based Approaches**: Leveraging knowledge graphs for relationship-aware recommendations.  \r\n\r\nKey challenges include data sparsity, scalability, and bias mitigation. Cutting-edge research explores reinforcement learning, explainability, and privacy-preserving methods to enhance recommendation systems.","description_from":"task","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","rank":"the archive's row order at snapshot; not re-ranked","rows_end_at":"2025-07-28","rows_withheld_as_spam":0,"metric_values":"the archive's strings, untouched"},"metrics":["nDCG@10","Recall@10","Recall@20","Recall@50","nDCG@100","Recall@100","HR@10","HR@10 (full corpus)","nDCG@10 (full corpus)","Recall@2"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"nDCG@10":"higher","Recall@10":"higher","Recall@20":"higher","Recall@50":"higher","nDCG@100":"higher","Recall@100":"higher","HR@10":null,"HR@10 (full corpus)":null,"nDCG@10 (full corpus)":"higher","Recall@2":"higher"}},"counts":{"rows":18,"rows_with_code":17,"rows_with_paper_page":18,"rows_dated":18,"rows_using_additional_data":2},"rows":[{"rank_in_archive_order":1,"model":"HyperML","metrics":{"HR@10":"0.8736","nDCG@10":"0.6404"},"uses_additional_data":false,"paper_date":"2018-09-05","paper":"/paper/hyperbolic-recommender-systems","paper_url":"https://arxiv.org/abs/1809.01703v3","paper_title":"HyperML: A Boosting Metric Learning Approach in Hyperbolic Space for Recommender Systems","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":2,"model":"LRML","metrics":{"HR@10":"0.8447","nDCG@10":"0.6152"},"uses_additional_data":false,"paper_date":"2017-07-17","paper":"/paper/latent-relational-metric-learning-via-memory","paper_url":"http://arxiv.org/abs/1707.05176v3","paper_title":"Latent Relational Metric Learning via Memory-based Attention for Collaborative Ranking","code":"https://github.com/vanzytay/WWW2018_LRML","n_code_links":1,"syntology":null},{"rank_in_archive_order":3,"model":"CML","metrics":{"HR@10":"0.7764","Recall@100":"0.6022","Recall@50":"0.4665","nDCG@10":"0.5301"},"uses_additional_data":false,"paper_date":"2017-04-01","paper":"/paper/collaborative-metric-learning","paper_url":"https://ylongqi.com/publication/www17b/","paper_title":"Collaborative Metric Learning","code":"https://github.com/changun/CollMetric","n_code_links":2,"syntology":null},{"rank_in_archive_order":4,"model":"HGN","metrics":{"Recall@10":"0.1255","nDCG@10":"0.1195"},"uses_additional_data":false,"paper_date":"2019-06-21","paper":"/paper/hierarchical-gating-networks-for-sequential","paper_url":"https://arxiv.org/abs/1906.09217v1","paper_title":"Hierarchical Gating Networks for Sequential Recommendation","code":"https://github.com/allenjack/HGN","n_code_links":2,"syntology":{"n_ran":1,"n_unverified":2,"n_samples":3,"n_pointer_only_licence":0}},{"rank_in_archive_order":5,"model":"KGNN-LS","metrics":{"Recall@10":"0.155","Recall@100":"0.458","Recall@2":"0.043","Recall@50":"0.321"},"uses_additional_data":false,"paper_date":"2019-05-11","paper":"/paper/knowledge-graph-convolutional-networks-for","paper_url":"https://arxiv.org/abs/1905.04413v3","paper_title":"Knowledge-aware Graph Neural Networks with Label Smoothness Regularization for Recommender Systems","code":"https://github.com/hwwang55/KGNN-LS","n_code_links":6,"syntology":{"n_ran":1,"n_unverified":3,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":6,"model":"Multi-Gradient Descent","metrics":{"Recall@20":"0.418"},"uses_additional_data":false,"paper_date":"2019-12-09","paper":"/paper/multi-gradient-descent-for-multi-objective","paper_url":"https://arxiv.org/abs/2001.00846v3","paper_title":"Multi-Gradient Descent for Multi-Objective Recommender Systems","code":"https://github.com/swisscom/ai-research-mamo-framework","n_code_links":1,"syntology":null},{"rank_in_archive_order":7,"model":"RecVAE","metrics":{"Recall@20":"0.414","Recall@50":"0.553","nDCG@100":"0.442"},"uses_additional_data":false,"paper_date":"2019-12-24","paper":"/paper/recvae-a-new-variational-autoencoder-for-top","paper_url":"https://arxiv.org/abs/1912.11160v1","paper_title":"RecVAE: a New Variational Autoencoder for Top-N Recommendations with Implicit Feedback","code":"https://github.com/ilya-shenbin/RecVAE","n_code_links":3,"syntology":{"n_ran":4,"n_unverified":0,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":8,"model":"VASP","metrics":{"Recall@20":"0.414","Recall@50":"0.552","nDCG@100":"0.448"},"uses_additional_data":true,"paper_date":"2021-02-10","paper":"/paper/deep-variational-autoencoder-with-shallow","paper_url":"https://arxiv.org/abs/2102.05774v1","paper_title":"Deep Variational Autoencoder with Shallow Parallel Path for Top-N Recommendation (VASP)","code":"https://github.com/zombak79/vasp","n_code_links":1,"syntology":null},{"rank_in_archive_order":9,"model":"H+Vamp Gated","metrics":{"Recall@20":"0.41308","Recall@50":"0.55109","nDCG@100":"0.44522"},"uses_additional_data":true,"paper_date":"2019-11-03","paper":"/paper/enhancing-vaes-for-collaborative-filtering","paper_url":"https://arxiv.org/abs/1911.00936v1","paper_title":"Enhancing VAEs for Collaborative Filtering: Flexible Priors & Gating Mechanisms","code":"https://github.com/psywaves/EVCF","n_code_links":1,"syntology":null},{"rank_in_archive_order":10,"model":"RaCT","metrics":{"Recall@20":"0.403","Recall@50":"0.543","nDCG@100":"0.434"},"uses_additional_data":false,"paper_date":"2019-06-10","paper":"/paper/towards-amortized-ranking-critical-training","paper_url":"https://arxiv.org/abs/1906.04281v2","paper_title":"Towards Amortized Ranking-Critical Training for Collaborative Filtering","code":"https://github.com/samlobel/RaCT_CF","n_code_links":1,"syntology":null},{"rank_in_archive_order":11,"model":"Mult-VAE PR","metrics":{"Recall@20":"0.395","Recall@50":"0.537","nDCG@100":"0.426"},"uses_additional_data":false,"paper_date":"2018-02-16","paper":"/paper/variational-autoencoders-for-collaborative","paper_url":"http://arxiv.org/abs/1802.05814v1","paper_title":"Variational Autoencoders for Collaborative Filtering","code":"https://github.com/microsoft/recommenders","n_code_links":18,"syntology":{"n_ran":2,"n_unverified":19,"n_samples":21,"n_pointer_only_licence":1}},{"rank_in_archive_order":12,"model":"EASE","metrics":{"Recall@20":"0.391","Recall@50":"0.521","nDCG@100":"0.420"},"uses_additional_data":false,"paper_date":"2019-05-08","paper":"/paper/190503375","paper_url":"https://arxiv.org/abs/1905.03375v1","paper_title":"Embarrassingly Shallow Autoencoders for Sparse Data","code":"https://github.com/PreferredAI/cornac","n_code_links":9,"syntology":{"n_ran":5,"n_unverified":1,"n_samples":6,"n_pointer_only_licence":0}},{"rank_in_archive_order":13,"model":"Mult-DAE","metrics":{"Recall@20":"0.387","Recall@50":"0.524","nDCG@100":"0.419"},"uses_additional_data":false,"paper_date":"2018-02-16","paper":"/paper/variational-autoencoders-for-collaborative","paper_url":"http://arxiv.org/abs/1802.05814v1","paper_title":"Variational Autoencoders for Collaborative Filtering","code":"https://github.com/microsoft/recommenders","n_code_links":18,"syntology":{"n_ran":2,"n_unverified":19,"n_samples":21,"n_pointer_only_licence":1}},{"rank_in_archive_order":14,"model":"LED","metrics":{"Recall@20":"0.375","Recall@50":"0.516"},"uses_additional_data":false,"paper_date":"2021-01-04","paper":"/paper/scalable-representation-learning-and","paper_url":"https://arxiv.org/abs/2101.00870v2","paper_title":"Lightweight representation learning for efficient and scalable recommendation","code":"https://github.com/criteo/deepr","n_code_links":1,"syntology":null},{"rank_in_archive_order":15,"model":"HSTU","metrics":{"HR@10 (full corpus)":"0.3556","nDCG@10 (full corpus)":"0.2098"},"uses_additional_data":false,"paper_date":"2024-02-27","paper":"/paper/actions-speak-louder-than-words-trillion","paper_url":"https://arxiv.org/abs/2402.17152v3","paper_title":"Actions Speak Louder than Words: Trillion-Parameter Sequential Transducers for Generative Recommendations","code":"https://github.com/facebookresearch/generative-recommenders","n_code_links":12,"syntology":{"n_ran":15,"n_unverified":27,"n_samples":42,"n_pointer_only_licence":3}},{"rank_in_archive_order":16,"model":"SASRec","metrics":{"HR@10 (full corpus)":"0.2889","nDCG@10 (full corpus)":"0.1621"},"uses_additional_data":false,"paper_date":"2018-08-20","paper":"/paper/180809781","paper_url":"http://arxiv.org/abs/1808.09781v1","paper_title":"Self-Attentive Sequential Recommendation","code":"https://github.com/microsoft/recommenders","n_code_links":8,"syntology":{"n_ran":3,"n_unverified":17,"n_samples":20,"n_pointer_only_licence":0}},{"rank_in_archive_order":17,"model":"BERT4Rec","metrics":{"HR@10 (full corpus)":"0.2816","nDCG@10 (full corpus)":"0.1703"},"uses_additional_data":false,"paper_date":"2019-04-14","paper":"/paper/bert4rec-sequential-recommendation-with","paper_url":"https://arxiv.org/abs/1904.06690v2","paper_title":"BERT4Rec: Sequential Recommendation with Bidirectional Encoder Representations from Transformer","code":"https://github.com/PaddlePaddle/PaddleRec/tree/master/models/rank/bert4rec","n_code_links":8,"syntology":{"n_ran":2,"n_unverified":4,"n_samples":6,"n_pointer_only_licence":1}},{"rank_in_archive_order":18,"model":"GRU4Rec","metrics":{"HR@10 (full corpus)":"0.2813","nDCG@10 (full corpus)":"0.1730"},"uses_additional_data":false,"paper_date":"2015-11-21","paper":"/paper/session-based-recommendations-with-recurrent","paper_url":"http://arxiv.org/abs/1511.06939v4","paper_title":"Session-based Recommendations with Recurrent Neural Networks","code":"https://github.com/microsoft/recommenders","n_code_links":25,"syntology":{"n_ran":2,"n_unverified":13,"n_samples":15,"n_pointer_only_licence":4}}],"since_archive":{"claim":"Results that newer papers report for their own method, placed here by Syntology. A model pointed at the cell in the paper's own table; the number was read from that cell and checked against this leaderboard's metric, dataset, split and scale; an independent check that saw this leaderboard's other rows and every other leaderboard on the same dataset accepted it. Not reviewed by the paper's authors or by the archive's editors, and not ranked against the archive rows.","extraction_file_present":true,"measurement":{"test_papers":883,"papers_with_output":881,"judged_true":108,"judged":110,"wilson95_lower":0.9361,"measured_on":"2026-09-24","frozen_commit":"0e3de0df94"},"measurement_note":"blind adjudication of accepted entries on a held-out split of archive papers, rules frozen before the test","coverage":{"sentence":"Syntology has checked 6,885 of the 9,623 papers on this site that are newer than the archive; results from the others appear after they are checked.","complete":false,"papers_newer_than_archive":9623,"papers_checked":6885,"papers_extracted_not_yet_verified":0,"boards_without_verdict":2,"papers_not_yet_extracted":2737},"order":"newest first by month (arXiv date, else the arXiv-id month), then arXiv id descending","columns":[],"entries":[]},"syntology":{"read_at":"2026-09-25T09:33:49+00:00","claim":"Per row: N of M harvested code samples from that row's paper executed on a synthesized fixture; the other M-N are unverified. Not a reproduction of the row's number; not a correctness claim. n_pointer_only_licence counts samples the site points at rather than redistributes (a licence axis, independent of ran/unverified).","rows_with_graph_line":10,"rows_with_any_sample_ran":10,"distinct_papers_with_graph_line":9,"distinct_papers_with_any_sample_ran":9,"samples_over_distinct_papers":{"n_ran":35,"n_unverified":86,"n_samples":121,"n_pointer_only_licence":9,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":37,"n_unverified":105,"n_samples":142,"n_pointer_only_licence":10,"note":"row-weighted: a paper behind several rows is counted once per row; inflated relative to samples_over_distinct_papers by design, kept for readers summing the per-row syntology blocks"}}}