Can Large Language Models Adequately Perform Symbolic Reasoning Over Time Series?
By: Zewen Liu , Juntong Ni , Xianfeng Tang and more
Potential Business Impact:
Finds hidden rules in data, like planets moving.
Uncovering hidden symbolic laws from time series data, as an aspiration dating back to Kepler's discovery of planetary motion, remains a core challenge in scientific discovery and artificial intelligence. While Large Language Models show promise in structured reasoning tasks, their ability to infer interpretable, context-aligned symbolic structures from time series data is still underexplored. To systematically evaluate this capability, we introduce SymbolBench, a comprehensive benchmark designed to assess symbolic reasoning over real-world time series across three tasks: multivariate symbolic regression, Boolean network inference, and causal discovery. Unlike prior efforts limited to simple algebraic equations, SymbolBench spans a diverse set of symbolic forms with varying complexity. We further propose a unified framework that integrates LLMs with genetic programming to form a closed-loop symbolic reasoning system, where LLMs act both as predictors and evaluators. Our empirical results reveal key strengths and limitations of current models, highlighting the importance of combining domain knowledge, context alignment, and reasoning structure to improve LLMs in automated scientific discovery.
Similar Papers
Toward Reasoning-Centric Time-Series Analysis
Artificial Intelligence
Helps computers understand why things change.
Neuro-Symbolic Artificial Intelligence: Towards Improving the Reasoning Abilities of Large Language Models
Artificial Intelligence
Teaches AI to think better and solve harder problems.
When LLM Meets Time Series: Can LLMs Perform Multi-Step Time Series Reasoning and Inference
Machine Learning (CS)
Tests computers on predicting future events.