AdaptFlow: Adaptive Workflow Optimization via Meta-Learning
This addresses the adaptability and scalability limitations of agentic workflows for LLM users, representing a novel method rather than an incremental improvement.
The paper tackles the problem of static, manually designed workflows for LLM-based agents by proposing AdaptFlow, a meta-learning framework that learns a generalizable workflow initialization for rapid adaptation to diverse tasks, achieving state-of-the-art results across benchmarks like question answering, code generation, and mathematical reasoning.
Recent advances in large language models (LLMs) have sparked growing interest in agentic workflows, which are structured sequences of LLM invocations intended to solve complex tasks. However, existing approaches often rely on static templates or manually designed workflows, which limit adaptability to diverse tasks and hinder scalability. We propose AdaptFlow, a natural language-based meta-learning framework inspired by model-agnostic meta-learning (MAML). AdaptFlow learns a generalizable workflow initialization that enables rapid subtask-level adaptation. It employs a bi-level optimization scheme: the inner loop refines the workflow for a specific subtask using LLM-generated feedback, while the outer loop updates the shared initialization to perform well across tasks. This setup allows AdaptFlow to generalize effectively to unseen tasks by adapting the initialized workflow through language-guided modifications. Evaluated across question answering, code generation, and mathematical reasoning benchmarks, AdaptFlow consistently outperforms both manually crafted and automatically searched baselines, achieving state-of-the-art results with strong generalization across tasks and models. The source code and data are available at https://github.com/microsoft/DKI_LLM/tree/AdaptFlow/AdaptFlow.