SEA-TS: Self-Evolving Agent for Autonomous Code Generation of Time Series Forecasting Algorithms
This work is significant for machine learning engineers and domain experts who struggle with data scarcity, distribution shift, and manual iteration in developing accurate time series forecasting models, by offering an autonomous code generation solution.
This paper introduces SEA-TS, a framework for autonomously generating, validating, and optimizing time series forecasting code through an iterative self-evolution loop. On the Solar-Energy benchmark, SEA-TS achieved a 40% MAE reduction compared to TimeMixer, and on proprietary datasets, it reduced WAPE by 8.6% for solar PV and 7.7% for residential load forecasting against human-engineered baselines.
Accurate time series forecasting underpins decision-making across domains, yet conventional ML development suffers from data scarcity in new deployments, poor adaptability under distribution shift, and diminishing returns from manual iteration. We propose Self-Evolving Agent for Time Series Algorithms (SEA-TS), a framework that autonomously generates, validates, and optimizes forecasting code via an iterative self-evolution loop. Our framework introduces three key innovations: (1) Metric-Advantage Monte Carlo Tree Search (MA-MCTS), which replaces fixed rewards with a normalized advantage score for discriminative search guidance; (2) Code Review with running prompt refinement, where each executed solution undergoes automated review followed by prompt updates that encode corrective patterns, preventing recurrence of similar errors; and (3) Global Steerable Reasoning, which compares each node against global best and worst solutions, enabling cross-trajectory knowledge transfer. We adopt a MAP-Elites archive for architectural diversity. On the public Solar-Energy benchmark, SEA-TS generated code achieves a 40% MAE reduction relative to TimeMixer, surpassing state-of-the-art methods. On proprietary datasets, SEA-TS generated code reduces WAPE by 8.6% on solar PV forecasting and 7.7% on residential load forecasting compared to human-engineered baselines, and achieves 26.17% MAPE on load forecasting versus 29.34% by TimeMixer. Notably, the evolved models discover novel architectural patterns--including physics-informed monotonic decay heads encoding solar irradiance constraints, per-station learned diurnal cycle profiles, and learnable hourly bias correction--demonstrating that autonomous ML engineering can generate genuinely novel algorithmic ideas beyond manual design.