{"id":24866423,"url":"https://github.com/jihyeonseong/som-tp","last_synced_at":"2025-03-26T19:21:27.075Z","repository":{"id":222425375,"uuid":"757237046","full_name":"jihyeonseong/SoM-TP","owner":"jihyeonseong","description":"Towards Diverse Perspective Learning with Selection over Multiple Temporal Poolings (AAAI24)","archived":false,"fork":false,"pushed_at":"2024-05-02T06:17:22.000Z","size":1358,"stargazers_count":1,"open_issues_count":0,"forks_count":1,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-02-01T01:11:23.801Z","etag":null,"topics":["classification","ensemble-learning","pooling","time-series"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/jihyeonseong.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2024-02-14T04:26:39.000Z","updated_at":"2025-01-02T12:08:22.000Z","dependencies_parsed_at":null,"dependency_job_id":"839e97b9-bd31-4587-8854-f6ef3c87849a","html_url":"https://github.com/jihyeonseong/SoM-TP","commit_stats":null,"previous_names":["jihyeonseong/som-tp"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jihyeonseong%2FSoM-TP","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jihyeonseong%2FSoM-TP/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jihyeonseong%2FSoM-TP/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jihyeonseong%2FSoM-TP/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/jihyeonseong","download_url":"https://codeload.github.com/jihyeonseong/SoM-TP/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":245718765,"owners_count":20661161,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["classification","ensemble-learning","pooling","time-series"],"created_at":"2025-02-01T01:11:24.700Z","updated_at":"2025-03-26T19:21:27.036Z","avatar_url":"https://github.com/jihyeonseong.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Towards Diverse Perspective Learning with Selection over Multiple Temporal Poolings (AAAI24)\n* This is the author code implements \"Towards Diverse Perspective Learning with Selection over Multiple Temporal Poolings,\" a paper accepted at AAAI 2024.\n* It builds upon the official code of [DTP github](https://github.com/donalee/DTW-Pool) and [softDTW github](https://github.com/Maghoumi/pytorch-softdtw-cuda) based on PyTorch.\n* For further details, please refer to the original [DTP](https://arxiv.org/abs/2104.02577) and [softDTW](https://arxiv.org/abs/1703.01541) papers.\n## Overview\n![image](https://github.com/jihyeonseong/SoM-TP/assets/159874470/c15390f3-3e6c-477b-b019-3ae1b08bda3f)\nIn Time Series Classification (TSC), temporal pooling methods that consider sequential information have been proposed. However, we found that each temporal pooling has a distinct mechanism, and can perform better or worse depending on time series data. We term this fixed pooling mechanism a single perspective of temporal poolings. In this paper, we propose a novel temporal pooling method with diverse perspective learning: Selection over Multiple Temporal Poolings (SoM-TP). \n* We investigate data dependency arising from distinct perspectives of existing temporal poolings.\n* We propose SoM-TP, a new temporal pooling method that fully utilizes the diverse temporal pooling mechanisms through an MCL-inspired selection ensemble.\n* We employ an attention mechanism to enable a non-iterative ensemble in a single classifier.\n* We define DPLN and perspective loss as a regularizer to promote diverse pooling selection.\n## Running the codes\n### STEP 1. Download the benchmark datsets for time series classification\n* The datasets can be downloaded form the [UCR/UEA repository](https://www.timeseriesclassification.com/).\n* Create a directory named \"data\" and store downloaded datasets within it.\n### STEP 2. Train the CNN classifier with various temporal poolings including SoM-TP\nFor traditional temporal poolings,\n```\npython main.py --model=ConvPool --pool=DTP\n```\nand for SoM-TP\n```\npython main.py --model=SoMTP\n```\n### STEP 3. Run LRP (Layer-wise Relevance Propagation: XAI input attribution method)\nFor traditional temporal poolings,\n```\npython LRP.py --model=ConvPool --pool=DTP\n```\nand for SoM-TP\n```\npython LRP.py --model=SoMTP\n```\n### SoM-TP performance\n1. Comparison with traditional temporal poolings\n![image](https://github.com/jihyeonseong/SoM-TP/assets/159874470/c9b862b8-7b2f-45eb-ad12-ca12af0ac7e0)\n2. Comparison with advanced TSC methods\n![image](https://github.com/jihyeonseong/SoM-TP/assets/159874470/dd0eb53a-f287-4943-bd62-85be70ac65c6)\n3. SoM-TP dynamic selection \n![image](https://github.com/jihyeonseong/SoM-TP/assets/159874470/65842d09-6f27-46c7-ae71-e817982e1465)\n4. LRP comparison\n![image](https://github.com/jihyeonseong/SoM-TP/assets/159874470/4e46bfd8-7a4b-4a89-8308-54399a53d275)\n\n## Citation\n```\n@article{Seong_Kim_Choi_2024,\ntitle={Towards Diverse Perspective Learning with Selection over Multiple Temporal Poolings}, volume={38},\nurl={https://ojs.aaai.org/index.php/AAAI/article/view/28743}, DOI={10.1609/aaai.v38i8.28743},\nnumber={8},\njournal={Proceedings of the AAAI Conference on Artificial Intelligence},\nauthor={Seong, Jihyeon and Kim, Jungmin and Choi, Jaesik}, year={2024}, month={Mar.}, pages={8948-8956} }\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjihyeonseong%2Fsom-tp","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fjihyeonseong%2Fsom-tp","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjihyeonseong%2Fsom-tp/lists"}