Search papers, labs, and topics across Lattice.
FlowEvo introduces a novel framework where workflows and executable skills co-evolve during inference, allowing agents to adaptively construct and refine their problem-solving strategies. This approach enables the persistent storage of successful workflows as callable skills, which can be utilized for future tasks, significantly enhancing the agent's performance. The results show that FlowEvo outperforms eight baseline models across multiple datasets, achieving a remarkable 85.6% accuracy on ALFWorld, far surpassing the nearest competitor while using fewer tokens.
FlowEvo enables agents to continuously learn and optimize their skills in real-time, achieving unprecedented accuracy in complex task environments.
Large language model agents can adapt to complex tasks by constructing workflows at inference time, but procedures discovered in one episode are usually discarded after execution. Existing skill libraries provide reusable executable routines, but are typically assembled offline and do not grow from the agent's own workflows. We introduce FlowEvo, a training-free framework in which workflows and skills co-evolve at inference time. FlowEvo compiles successful workflows into callable skills, stores them in a persistent bank, and uses retrieved skills either through direct execution or as context for constructing new workflows. It also tracks each skill's downstream utility and suppresses skills that cause negative transfer. Using a shared GPT-4o-mini backbone, FlowEvo achieves the highest accuracy among 8 baselines on the full standard splits of ALFWorld, HumanEval, MBPP, GSM8K, and MATH-500. On ALFWorld, it reaches 85.6%, 26.4 points above the strongest baseline, while using roughly one third as many tokens. Across 10 base models spanning 7B to 671B parameters, FlowEvo outperforms ExpeL in 49 of 50 model-dataset comparisons. Code is available at https://github.com/DEFENSE-SEU/FlowEvo.
