Search papers, labs, and topics across Lattice.
EvoTS-Agent is a self-evolving LLM agent designed for autonomous change-point detection in financial time series, addressing the challenges posed by non-stationary and heterogeneous data. By employing a validation-guided approach, it adapts its detection pipeline through curated exploratory data analysis and evolves experiment trajectories using three operators: Revision, Alternative Strategy, and Recombination. Experiments show that EvoTS-Agent outperforms existing LLM-based agents while achieving a 100% execution success rate across various datasets, highlighting its scalability and adaptability in financial applications.
EvoTS-Agent not only outperforms existing models in financial change-point detection but does so with a flawless execution success rate across multiple datasets.
Financial time series exhibit non-stationary and heterogeneous statistical properties, making change-point detection challenging because no single unsupervised algorithm performs consistently across assets and market regimes. Conventional workflows consequently depend heavily on expert-driven model selection, feature design, and hyperparameter tuning, limiting their scalability and adaptability. We propose EvoTS-Agent, a validation-guided self-evolving LLM agent for autonomous financial time-series change-point detection. EvoTS-Agent first performs curated exploratory data analysis to characterize dataset properties and initialize candidate detection models. It then evolves executable experiment trajectories through three complementary operators: \textit{Revision} exploits the current best solution, \textit{Alternative Strategy} explores fundamentally different modeling directions when progress stagnates, and \textit{Recombination} synthesizes complementary evidence from high-performing trajectories. Validation feedback guides trajectory evolution throughout the search, enabling the agent to adapt its detection pipeline to the statistical characteristics of each dataset while preserving reliable optimization. Experiments across four benchmark datasets demonstrate that EvoTS-Agent consistently outperforms existing LLM-based agents while maintaining a 100\% execution success rate across all evaluated backbone LLMs.
