{"about":{"site":"https://codewithpapers.app","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.","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"},"url":"/paper/long-term-series-forecasting-with-query","title":"Long-term series forecasting with Query Selector -- efficient model of sparse attention","arxiv_id":"2107.08687","date":"2021-07-19","proceeding":null,"authors":["Jacek Klimek","Jakub Klimek","Witold Kraskiewicz","Mateusz Topolewski"],"abstract":"Various modifications of TRANSFORMER were recently used to solve time-series forecasting problem. We propose Query Selector - an efficient, deterministic algorithm for sparse attention matrix. Experiments show it achieves state-of-the art results on ETT, Helpdesk and BPI'12 datasets.","url_abs":"https://arxiv.org/abs/2107.08687v2","url_pdf":"https://arxiv.org/pdf/2107.08687v2.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":"long-term-series-forecasting-with-query","repo_url":"https://github.com/moraieu/query-selector","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"long-term-series-forecasting-with-query","repo_url":"https://github.com/moraieu/query-selector-classification","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"multivariate-time-series-forecasting","task_name":"Multivariate Time Series Forecasting"},{"task_slug":"time-series-1","task_name":"Time Series"},{"task_slug":"time-series","task_name":"Time Series Analysis"},{"task_slug":"time-series-forecasting","task_name":"Time Series Forecasting"}],"methods":[{"method_slug":"absolute-position-encodings","method_name":"Absolute Position Encodings"},{"method_slug":"adam","method_name":"Adam"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"bpe","method_name":"BPE"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"label-smoothing","method_name":"Label Smoothing"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"position-wise-feed-forward-layer","method_name":"Position-Wise Feed-Forward Layer"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"transformer","method_name":"Transformer"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/multivariate-time-series-forecasting-on-bpi","task":"Multivariate Time Series Forecasting","dataset":"BPI challenge '12","model":"QuerySelector","rank_in_archive_order":1,"of":2,"metrics":{"Accuracy":"0.79"},"uses_additional_data":false},{"leaderboard":"/sota/multivariate-time-series-forecasting-on","task":"Multivariate Time Series Forecasting","dataset":"Helpdesk","model":"QuerySelector","rank_in_archive_order":1,"of":2,"metrics":{"Accuracy":"0.743"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth1-168-1","task":"Time Series Forecasting","dataset":"ETTh1 (168) Multivariate","model":"QuerySelector","rank_in_archive_order":2,"of":4,"metrics":{"MAE":"0.6088","MSE":"0.6835"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth1-168-1","task":"Time Series Forecasting","dataset":"ETTh1 (168) Multivariate","model":"Transformer","rank_in_archive_order":3,"of":4,"metrics":{"MAE":"0.6325","MSE":"0.7146"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth1-168-2","task":"Time Series Forecasting","dataset":"ETTh1 (168) Univariate","model":"QuerySelector","rank_in_archive_order":2,"of":4,"metrics":{"MAE":"0.2371","MSE":"0.0935"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth1-168-2","task":"Time Series Forecasting","dataset":"ETTh1 (168) Univariate","model":"Transformer","rank_in_archive_order":3,"of":4,"metrics":{"MAE":"0.2539","MSE":"0.1049"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth1-24-1","task":"Time Series Forecasting","dataset":"ETTh1 (24) Multivariate","model":"QuerySelector","rank_in_archive_order":3,"of":5,"metrics":{"MAE":"0.4627","MSE":"0.4226"},"uses_additional_data":true},{"leaderboard":"/sota/time-series-forecasting-on-etth1-24-1","task":"Time Series Forecasting","dataset":"ETTh1 (24) Multivariate","model":"Transformer","rank_in_archive_order":4,"of":5,"metrics":{"MAE":"0.4788","MSE":"0.4496"},"uses_additional_data":true},{"leaderboard":"/sota/time-series-forecasting-on-etth1-24-2","task":"Time Series Forecasting","dataset":"ETTh1 (24) Univariate","model":"QuerySelector","rank_in_archive_order":2,"of":5,"metrics":{"MAE":"0.1616","MSE":"0.0436"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth1-24-2","task":"Time Series Forecasting","dataset":"ETTh1 (24) Univariate","model":"Transformer","rank_in_archive_order":4,"of":5,"metrics":{"MAE":"0.183","MSE":"0.0548"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth1-336-1","task":"Time Series Forecasting","dataset":"ETTh1 (336) Multivariate","model":"Transformer","rank_in_archive_order":67,"of":72,"metrics":{"MAE":"0.7041","MSE":"0.8321"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth1-336-1","task":"Time Series Forecasting","dataset":"ETTh1 (336) Multivariate","model":"QuerySelector","rank_in_archive_order":68,"of":72,"metrics":{"MAE":"0.7039","MSE":"0.8503"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth1-336-2","task":"Time Series Forecasting","dataset":"ETTh1 (336) Univariate","model":"QuerySelector","rank_in_archive_order":8,"of":10,"metrics":{"MAE":"0.2844","MSE":"0.1267"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth1-336-2","task":"Time Series Forecasting","dataset":"ETTh1 (336) Univariate","model":"Transformer","rank_in_archive_order":9,"of":10,"metrics":{"MAE":"0.3201","MSE":"0.1541"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth1-48-1","task":"Time Series Forecasting","dataset":"ETTh1 (48) Multivariate","model":"QuerySelector","rank_in_archive_order":2,"of":4,"metrics":{"MAE":"0.4878","MSE":"0.4581"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth1-48-1","task":"Time Series Forecasting","dataset":"ETTh1 (48) Multivariate","model":"Transformer","rank_in_archive_order":3,"of":4,"metrics":{"MAE":"0.4968","MSE":"0.4668"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth1-48-2","task":"Time Series Forecasting","dataset":"ETTh1 (48) Univariate","model":"QuerySelector","rank_in_archive_order":2,"of":4,"metrics":{"MAE":"0.2118","MSE":"0.0721"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth1-48-2","task":"Time Series Forecasting","dataset":"ETTh1 (48) Univariate","model":"Transformer","rank_in_archive_order":3,"of":4,"metrics":{"MAE":"0.2144","MSE":"0.074"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth1-720-1","task":"Time Series Forecasting","dataset":"ETTh1 (720) Multivariate","model":"Transformer","rank_in_archive_order":21,"of":22,"metrics":{"MAE":"0.8399","MSE":"1.108"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth1-720-1","task":"Time Series Forecasting","dataset":"ETTh1 (720) Multivariate","model":"QuerySelector","rank_in_archive_order":22,"of":22,"metrics":{"MAE":"0.8428","MSE":"1.115"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth1-720-2","task":"Time Series Forecasting","dataset":"ETTh1 (720) Univariate","model":"QuerySelector","rank_in_archive_order":11,"of":12,"metrics":{"MAE":"0.373","MSE":"0.2136"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth1-720-2","task":"Time Series Forecasting","dataset":"ETTh1 (720) Univariate","model":"Transformer","rank_in_archive_order":12,"of":12,"metrics":{"MAE":"0.4213","MSE":"0.2501"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth2-168-1","task":"Time Series Forecasting","dataset":"ETTh2 (168) Multivariate","model":"Transformer","rank_in_archive_order":3,"of":4,"metrics":{"MAE":"0.9726","MSE":"1.6225"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth2-168-1","task":"Time Series Forecasting","dataset":"ETTh2 (168) Multivariate","model":"QuerySelector","rank_in_archive_order":4,"of":4,"metrics":{"MAE":"1.0125","MSE":"1.7385"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth2-168-2","task":"Time Series Forecasting","dataset":"ETTh2 (168) Univariate","model":"QuerySelector","rank_in_archive_order":3,"of":4,"metrics":{"MAE":"0.3322","MSE":"0.1753"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth2-168-2","task":"Time Series Forecasting","dataset":"ETTh2 (168) Univariate","model":"Transformer","rank_in_archive_order":4,"of":4,"metrics":{"MAE":"0.3547","MSE":"0.1974"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth2-24-1","task":"Time Series Forecasting","dataset":"ETTh2 (24) Multivariate","model":"QuerySelector","rank_in_archive_order":2,"of":4,"metrics":{"MAE":"0.4864","MSE":"0.4124"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth2-24-1","task":"Time Series Forecasting","dataset":"ETTh2 (24) Multivariate","model":"Transformer","rank_in_archive_order":3,"of":4,"metrics":{"MAE":"0.5013","MSE":"0.4237"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth2-24-2","task":"Time Series Forecasting","dataset":"ETTh2 (24) Univariate","model":"QuerySelector","rank_in_archive_order":3,"of":5,"metrics":{"MAE":"0.2239","MSE":"0.0843"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth2-24-2","task":"Time Series Forecasting","dataset":"ETTh2 (24) Univariate","model":"Transformer","rank_in_archive_order":4,"of":5,"metrics":{"MAE":"0.2479","MSE":"0.0999"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth2-336-1","task":"Time Series Forecasting","dataset":"ETTh2 (336) Multivariate","model":"QuerySelector","rank_in_archive_order":19,"of":20,"metrics":{"MAE":"1.1859","MSE":"2.3168"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth2-336-1","task":"Time Series Forecasting","dataset":"ETTh2 (336) Multivariate","model":"Transformer","rank_in_archive_order":20,"of":20,"metrics":{"MAE":"1.2189","MSE":"2.6617"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth2-336-2","task":"Time Series Forecasting","dataset":"ETTh2 (336) Univariate","model":"QuerySelector","rank_in_archive_order":8,"of":10,"metrics":{"MAE":"0.371","MSE":"0.2088"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth2-336-2","task":"Time Series Forecasting","dataset":"ETTh2 (336) Univariate","model":"Transformer","rank_in_archive_order":10,"of":10,"metrics":{"MAE":"0.3805","MSE":"0.2191"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth2-48-1","task":"Time Series Forecasting","dataset":"ETTh2 (48) Multivariate","model":"QuerySelector","rank_in_archive_order":3,"of":4,"metrics":{"MAE":"0.9317","MSE":"1.4074"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth2-48-1","task":"Time Series Forecasting","dataset":"ETTh2 (48) Multivariate","model":"Transformer","rank_in_archive_order":4,"of":4,"metrics":{"MAE":"0.9488","MSE":"1.522"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth2-48-2","task":"Time Series Forecasting","dataset":"ETTh2 (48) Univariate","model":"QuerySelector","rank_in_archive_order":3,"of":4,"metrics":{"MAE":"0.2622","MSE":"0.1117"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth2-48-2","task":"Time Series Forecasting","dataset":"ETTh2 (48) Univariate","model":"Transformer","rank_in_archive_order":4,"of":4,"metrics":{"MAE":"0.2763","MSE":"0.1218"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth2-720-1","task":"Time Series Forecasting","dataset":"ETTh2 (720) Multivariate","model":"QuerySelector","rank_in_archive_order":19,"of":20,"metrics":{"MAE":"1.3084","MSE":"3.0664"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth2-720-1","task":"Time Series Forecasting","dataset":"ETTh2 (720) Multivariate","model":"Transformer","rank_in_archive_order":20,"of":20,"metrics":{"MAE":"1.3668","MSE":"3.1805"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth2-720-2","task":"Time Series Forecasting","dataset":"ETTh2 (720) Univariate","model":"QuerySelector","rank_in_archive_order":8,"of":11,"metrics":{"MAE":"0.413","MSE":"0.2585"},"uses_additional_data":false},{"leaderboard":"/sota/time-series-forecasting-on-etth2-720-2","task":"Time Series Forecasting","dataset":"ETTh2 (720) Univariate","model":"Transformer","rank_in_archive_order":10,"of":11,"metrics":{"MAE":"0.434","MSE":"0.2853"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2107.08687","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}