ACL 2026时间序列论文总结：LLM多模态大模型预测与异常检测
ACL2026将于2026年7月2日至7日在美国圣迭戈举行，收录14篇时间序列论文，涵盖LLM、VLM、多模态大模型、预测、异常检测、推理方向，其中Main论文6篇，Findings论文8篇。
ACL 2026计划于2026年7月2日至7日，在美国加利福尼亚州圣迭戈举办。本文将带大家快速过一遍本届会议中与时间序列（Time Series）相关的论文。根据目前的公开信息，相关论文共计14篇，其中Main有6篇，Findings有8篇。覆盖的方向包括：LLM、VLM、多模态大模型、预测、异常检测、推理等，可谓是相当密集。
先上总览，再逐一细看。
Main 1. Augur: Modeling Covariate Causal Associations in Time Series via Large Language Models 2. Inferring Events from Time Series using Language Models 3. ZARA: Training-Free Motion Time-Series Reasoning via Evidence-Grounded LLM Agents 4. Is the Attention Matrix Really the Key to Self‑Attention in Multivariate Long‑Term Time Series Forecasting? 5. Markovian Linguistic-Temporal Bridge: Unlocking the Potential of LLMs for Time Series Forecasting 6. TimeSAF: Towards LLM-Guided Semantic Asynchronous Fusion for Time Series Forecasting Findings 7. Time-RA: Towards Time Series Reasoning for Anomaly Diagnosis with LLM Feedback 8. CTRL: Control-Based Time Series Forecasting with LLM-Guided Residual Learning 9. Learning Dynamic Representations and Policies from Multimodal Clinical Time-Series with Informative Missingness 10. Probabilistic Depression Detection from Textual Time Series 11. LLaTiSA: Towards Difficulty-Stratified Time Series Reasoning from Visual Perception to Semantics 12. CaTS-Bench: Can Language Models Describe Time Series? 13. Can Large Language Models Adequately Perform Symbolic Reasoning Over Time Series? 14. Enhancing Zero-Shot Time Series Forecasting in Off-the-Shelf LLMs via Noise Injection Prompting |
|---|
Main
1 Augur: Modeling Covariate Causal Associations in Time Series via Large Language Models
链接：https://aclanthology.org/2026.acl-long.32/
作者：Zhiqing Cui, Binwu Wang, Qingxiang Liu, Yeqiang Wang, Zhengyang Zhou, Yuxuan Liang, Yang Wang
关键词：预测，LLM，因果
2 Inferring Events from Time Series using Language Models
链接：https://aclanthology.org/2026.acl-long.157/
作者：Mingtian Tan, Mike A Merrill, Zachary Gottesman, Tim Althoff, Da vid Evans, Thomas Hartvigsen
关键词：时序事件预测，LLM
3 ZARA: Training-Free Motion Time-Series Reasoning via Evidence-Grounded LLM Agents
链接：https://aclanthology.org/2026.acl-long.684/
作者：Zechen Li, Baiyu Chen, Hao Xue, Flora D. Salim
关键词：时序推理，Agents
4 Is the Attention Matrix Really the Key to Self‑Attention in Multivariate Long‑Term Time Series Forecasting?
链接：https://aclanthology.org/2026.acl-long.853/
作者：Xinyu Li, Kexi Chen, Jiajie Shen, Ying Zheng, Hong Lu, Jin Zhao, Xin Wang
关键词：长时预测，自注意力机制
5 Markovian Linguistic-Temporal Bridge: Unlocking the Potential of LLMs for Time Series Forecasting
链接：https://aclanthology.org/2026.acl-long.1014/
作者：Siming Sun, Kai Zhang, Xuejun Jiang, Wenchao Meng, Qinmin Yang
关键词：预测，LLM，马尔可夫状态转移图
6 TimeSAF: Towards LLM-Guided Semantic Asynchronous Fusion for Time Series Forecasting
链接：https://aclanthology.org/2026.acl-long.1208/
作者：Fan Zhang, Shiming Fan, Hua Wang
关键词：预测，LLM，语义聚合
Findings
7 Time-RA: Towards Time Series Reasoning for Anomaly Diagnosis with LLM Feedback
链接：https://aclanthology.org/2026.findings-acl.562/
作者：Yiyuan Yang, Zichuan Liu, Lei Song, Kai Ying, Stephen Wang, Joshua Thomas Bamford, Svitlana Vyetrenko, Jiang Bian, Qingsong Wen
关键词：异常检测，时序推理
8 CTRL: Control-Based Time Series Forecasting with LLM-Guided Residual Learning
链接：https://aclanthology.org/2026.findings-acl.1104/
作者：Minkyoung Kim, Daeun Ji, Yohan Lee, Beomsoo Kim, Beakcheol Jang
关键词：预测，LLM，残差学习
9 Learning Dynamic Representations and Policies from Multimodal Clinical Time-Series with Informative Missingness
链接：https://aclanthology.org/2026.findings-acl.1313/
作者：Zihan Liang, Ziwen Pan, Ruoxuan Xiong
关键词：多模态医疗时序
10 Probabilistic Depression Detection from Textual Time Series
链接：https://aclanthology.org/2026.findings-acl.1630/
作者：Fabian Schmidt, Seyedehmoniba Ra van, Vladimir Vlassov
关键词：时序推理，临床访谈话语序列，抑郁检测，概率框架
11 LLaTiSA: Towards Difficulty-Stratified Time Series Reasoning from Visual Perception to Semantics
链接：https://aclanthology.org/2026.findings-acl.1636/
作者：Yueyang Ding, HaoPeng Zhang, Rui Dai, Yi Wang, Tianyu Zong, Kaikui Liu, Xiangxiang Chu
关键词：时序推理，VLM
12 CaTS-Bench: Can Language Models Describe Time Series?
链接：https://aclanthology.org/2026.findings-acl.1722/
作者：Luca Zhou, Pratham Yashwante, Marshall Fisher, Alessio Sampieri, Zihao Zhou, Fabio Galasso, Rose Yu
关键词：时序描述，benchmark
13 Can Large Language Models Adequately Perform Symbolic Reasoning Over Time Series?
链接：https://aclanthology.org/2026.findings-acl.1756/
作者：Zewen Liu, Juntong Ni, Xianfeng Tang, Max SY Lau, Qi He, Wenpeng Yin, Wei Jin
关键词：符号推理，LLM
14 Enhancing Zero-Shot Time Series Forecasting in Off-the-Shelf LLMs via Noise Injection Prompting
链接：https://aclanthology.org/2026.findings-acl.2054/
作者：Xingyou Yin, Ceyao Zhang, Min Hu, Kai Chen
关键词：零样本预测，LLM，分布偏移，噪声注入提示词
游乐网为非赢利性网站，所展示的游戏/软件/文章内容均来自于互联网或第三方用户上传分享，版权归原作者所有，本站不承担相应法律责任。如您发现有涉嫌抄袭侵权的内容，请联系youleyoucom@outlook.com。
同类文章
CAD零基础入门教程：坐标输入、图层管理与基础绘图命令
本文面向CAD零基础学习者，系统讲解坐标输入、图层管理与基础绘图命令的核心用法。通过分步实操与常见问题排查，帮助新手建立精确绘图习惯，掌握规范出图的基础能力。
CAD从入门到项目交付：绘图、标注、图块与实战工作流
掌握CAD的核心在于建立“画得准、标得清、复用快、交付稳”的工作流。本文提供从环境设置、高频命令组合、标注规范、图块标准化到项目分阶段交付的完整路径，帮助初学者避免常见返工陷阱，独立完成可检查、可复用、可打印的工程图纸。
Claude Code 登录指南：个人、Teams 与企业账号区分与授权步骤
本文详细解析 Claude Code 登录前的账号类型区分方法，涵盖个人订阅、Teams 席位与企业 Enterprise 席位的授权路径差异。提供终端登录命令、环境变量排查及常见异常处理步骤，帮助用户快速完成正确授权并避免登录路径混淆。
Claude Code 文件修改前的权限模式配置与命令审批指南
本文详细介绍Claude Code在修改文件前的权限模式配置方法，包括defaultMode可选值、permissions allow与deny规则设置、多层级配置文件管理以及 status验证技巧，帮助开发者安全高效地使用AI编程助手。
Claude Code接入VS Code后先测扩展和终端命令
在VS Code中接入Claude Code后，建议优先验证扩展面板与集成终端两条入口。本文提供标准检查顺序、关键命令与常见故障排查路径，帮助你快速确认环境就绪，避免后续开发受阻。
- 热门数据榜
1
2
3
4
5
6
7
8
9
10
相关攻略
2026-09-01 16:53
2026-09-01 16:52
2026-09-01 14:27
2026-09-01 14:12
2026-09-01 14:10
2026-09-01 14:07
2026-09-01 13:55
2026-09-01 13:47
热门教程
- 游戏攻略
- 安卓教程
- 苹果教程
- 电脑教程
