Problem
Research questions and friction points this paper is trying to address.
Time-Series Question Answering
Zero-Shot Generalization
Cross-Dataset Transfer
Overfitting
Zero-Shot Time-Series Question Answering via Decoupled Perception and Reasoning
🤖 AI Summary
This study addresses the susceptibility of existing time series question-answering methods to overfitting and their limited generalizability under joint variations in input, context, and format. To overcome these limitations, this work proposes TSHarness, an agent-based framework that introduces a novel decoupling mechanism separating perception from reasoning. By leveraging structured states to isolate numerical extraction from semantic inference, the framework incorporates reusable analytical memory, an adaptive tool selector, and iterative feedback loops to ensure evidential sufficiency. Notably, this approach achieves efficient zero-shot question answering across diverse datasets without requiring target-domain training or answer-level feedback. Consequently, it establishes a highly generalizable and cost-effective foundational solution for time series analysis.
📝 Abstract
Time-series question answering (TSQA) requires grounding linguistic queries and diverse answer formats in complex numerical observations. However, existing methods heavily overfit to specific datasets and struggle to generalize when input series, question contexts, and answer requirements shift simultaneously. To address this challenge, we propose TSHarness, an agentic framework that establishes a decoupled workflow for cross-dataset zero-shot TSQA. At its core, TSHarness divides and conquers numerical perception and contextual reasoning via a structured Time-Series Perception State (TPS). Guided by a reusable memory of analytical knowledge, a learned Tool Selector adaptively invokes numerical tools to extract salient statistical and temporal features into the TPS. The answering agent then performs semantic reasoning over the TPS to generate target outputs, triggering iterative re-perception through the feedback loop when evidence is deemed insufficient. By separating numerical feature extraction from question-specific reasoning, TSHarness eliminates the need for target-side training or answer feedback, providing a generalizable, cost-efficient foundation for zero-shot TSQA.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Zero-Shot Time-Series Question Answering
Decoupled Perception and Reasoning
Agentic Framework
Time-Series Perception State
Iterative Re-perception
🔎 Similar Papers
No similar papers found.
