SOTAVerified

Long-term series forecasting with Query Selector -- efficient model of sparse attention

2021-07-19Code Available1· sign in to hype

Jacek Klimek, Jakub Klimek, Witold Kraskiewicz, Mateusz Topolewski

Code Available — Be the first to reproduce this paper.

Reproduce

Code

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.

Tasks

Benchmark Results

DatasetModelMetricClaimedVerifiedStatus
BPI challenge '12QuerySelectorAccuracy0.79Unverified
HelpdeskQuerySelectorAccuracy0.74Unverified

Reproductions