{"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/svq-sparse-vector-quantization-for","title":"Does Vector Quantization Fail in Spatio-Temporal Forecasting? Exploring a Differentiable Sparse Soft-Vector Quantization Approach","arxiv_id":"2312.03406","date":"2023-12-06","proceeding":null,"authors":["Chao Chen","Tian Zhou","Yanjun Zhao","Hui Liu","Liang Sun","Rong Jin"],"abstract":"Spatio-temporal forecasting is crucial in various fields and requires a careful balance between identifying subtle patterns and filtering out noise. Vector quantization (VQ) appears well-suited for this purpose, as it quantizes input vectors into a set of codebook vectors or patterns. Although VQ has shown promise in various computer vision tasks, it surprisingly falls short in enhancing the accuracy of spatio-temporal forecasting. We attribute this to two main issues: inaccurate optimization due to non-differentiability and limited representation power in hard-VQ. To tackle these challenges, we introduce Differentiable Sparse Soft-Vector Quantization (SVQ), the first VQ method to enhance spatio-temporal forecasting. SVQ balances detail preservation with noise reduction, offering full differentiability and a solid foundation in sparse regression. Our approach employs a two-layer MLP and an extensive codebook to streamline the sparse regression process, significantly cutting computational costs while simplifying training and improving performance. Empirical studies on five spatio-temporal benchmark datasets show SVQ achieves state-of-the-art results, including a 7.9% improvement on the WeatherBench-S temperature dataset and an average mean absolute error reduction of 9.4% in video prediction benchmarks (Human3.6M, KTH, and KittiCaltech), along with a 17.3% enhancement in image quality (LPIPS). Code is publicly available at https://github.com/Pachark/SVQ-Forecasting.","url_abs":"https://arxiv.org/abs/2312.03406v4","url_pdf":"https://arxiv.org/pdf/2312.03406v4.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":"svq-sparse-vector-quantization-for","repo_url":"https://github.com/Pachark/SVQ-Forecasting","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"attribute","task_name":"Attribute"},{"task_slug":"computational-efficiency","task_name":"Computational Efficiency"},{"task_slug":"quantization","task_name":"Quantization"},{"task_slug":"spatio-temporal-forecasting","task_name":"Spatio-Temporal Forecasting"},{"task_slug":"traffic-prediction","task_name":"Traffic Prediction"},{"task_slug":"transfer-learning","task_name":"Transfer Learning"},{"task_slug":"video-prediction","task_name":"Video Prediction"},{"task_slug":"weather-forecasting","task_name":"Weather Forecasting"},{"task_slug":"regression-1","task_name":"regression"}],"methods":[{"method_slug":"base","method_name":"BASE"},{"method_slug":"mae","method_name":"MAE"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/traffic-prediction-on-bjtaxi","task":"Traffic Prediction","dataset":"BJTaxi","model":"SimVP+SVQ (Learnable)","rank_in_archive_order":5,"of":5,"metrics":{"MAE @ in":"14.64"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2312.03406","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}