T-SMART: Mechanism-Level Attribution for Tool-Augmented Time-Series Question Answering

2026-09-12

AI Summary

为解决时间序列问答中语言模型计算困难问题，提出T-SMART框架，通过分离语言理解、数值计算和感知功能，显著提升准确性。

Abstract

Large language models (LLMs) can struggle with time-series question answering (TS-QA), especially when numerical signals are serialized as text and require explicit computation. Tool-augmented approaches improve performance, but existing systems often intertwine language reasoning, computation, and perception, making it difficult to determine which components drive the gains. We present T-SMART, a neurosymbolic framework that separates these roles: a frozen LLM interprets questions and selects operations, deterministic tools perform numerical computation, and structured perception is invoked only when needed. Controlled paired ablations show that deterministic computation provides the dominant benefit, improving accuracy by 31.7 percentage points over direct LLM reasoning on serialized time series, while language understanding and perception offer smaller complementary gains. These results indicate that tool-augmented TS-QA benefits primarily from reliable numerical execution rather than additional language-model reasoning and provide a controlled framework for analyzing component contributions in neurosymbolic time-series systems.

Problem

• time-series question answering
• large language models
• tool-augmented approaches

Innovation

• neurosymbolic framework
• deterministic computation
• tool-augmented TS-QA