analiza aiqbench
This research introduces MarsTSC, the first framework leveraging Visual Language Model (VLM) agentic reasoning to solve few-shot multimodal time series classification. By employing a self-evolving knowledge bank and a three-role collaborative system, it significantly outperforms classical baselines while providing interpretable rationales grounded in temporal features. The work advances the intersection of agentic workflows and specialized ML tasks, offering a robust approach for handling distribution shifts without relying on massive labeled datasets.
Empowering VLMs for Few-Shot Multimodal Time Series Classification via Tailored Agentic Reasoning
arXiv:2605.09395v3 Announce Type: replace-cross Abstract: In this paper, we propose the first VL\underline{\textbf{M}} \underline{\textbf{a}}gentic \underline{\textbf{r}}easoning framework for few-\underline{\textbf{s}}hot multimodal \underline{\textbf{T}}ime \underline{\textbf{S}}eries \underline{\textbf{C}}lassification (\textsc{MarsTSC}), which introduces a self-evolving knowledge bank as a dynamic context iteratively refined via reflective agentic reasoning. The framework comprises three collaborative roles: i) Generator conducts reliable classification via reasoning; ii) Reflector diagnoses the root causes of reasoning errors to yield discriminative insights targeting the temporal features overlooked by Generator; iii) Modifier applies verified updates to the knowledge bank to prevent context collapse. We further introduce a test-time update strategy to enable cautious, continuous knowledge bank refinement to mitigate few-shot bias and distribution shift. Extensive experiments across 12 mainstream time series benchmark datasets demonstrate that \method{} delivers substantial and consistent performance gains across 5 VLM backbones, outperforming both classical and foundation model-based time series baselines under few-shot conditions, while producing interpretable rationales that ground each classification decision in human-readable feature evidence. Code is available at https://github.com/HuangJW0821/MarsTSC.
Abstract:In this paper, we propose the first VL\underline{\textbf{M}} \underline{\textbf{a}}gentic \underline{\textbf{r}}easoning framework for few-\underline{\textbf{s}}hot multimodal \underline{\textbf{T}}ime \underline{\textbf{S}}eries \underline{\textbf{C}}lassification (\textsc{MarsTSC}), which introduces a self-evolving knowledge bank as a dynamic context iteratively refined via reflective agentic reasoning. The framework comprises three collaborative roles: i) Generator conducts reliable classification via reasoning; ii) Reflector diagnoses the root causes of reasoning errors to yield discriminative insights targeting the temporal features overlooked by Generator; iii) Modifier applies verified updates to the knowledge bank to prevent context collapse. We further introduce a test-time update strategy to enable cautious, continuous knowledge bank refinement to mitigate few-shot bias and distribution shift. Extensive experiments across 12 mainstream time series benchmark datasets demonstrate that \method{} delivers substantial and consistent performance gains across 5 VLM backbones, outperforming both classical and foundation model-based time series baselines under few-shot conditions, while producing interpretable rationales that ground each classification decision in human-readable feature evidence. Code is available at this https URL.
17 pages, 12 figures, 8 tables. Accepted by ACM MM 2026
Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Multiagent Systems (cs.MA); Multimedia (cs.MM)
I.2.0; I.2.4; I.5.4
arXiv:2605.09395 [cs.AI]
(or arXiv:2605.09395v3 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2605.09395 Focus to learn more arXiv-issued DOI via DataCite
https://doi.org/10.1145/3767308.3836170 Focus to learn more DOI(s) linking to related resources
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 279 cuvinte
Subiecte conexe
MedTVL: Harnessing Vision and Language for Medical Time Series Classification
arXiv · 1 sept. 2026
Amazon is investing in the Lean Focused Research Organization
Amazon Science · 26 iul. 2026
Vertical Advantage: Transforming Industries with Lakebase and Agentic AI
Databricks · 27 aug. 2026
34 Amazon Research Awards Build on Trainium recipients announced
Amazon Science · 5 aug. 2026