FML-bench: A Controlled Study of AI Research Agent Strategies from the Perspective of Search Dynamics

• URL: http://arxiv.org/abs/2605.17373v1
• Date: Sun, 17 May 2026 10:30:38 GMT
• Title: FML-bench: A Controlled Study of AI Research Agent Strategies from the Perspective of Search Dynamics
• Abstract summary: FML-Bench is a benchmark of 18 fundamental ML research tasks across 10 domains.<n>It separates agent strategy from execution infrastructure and defines 12 process-level behavioral metrics.<n>We find greedy search tends to be more effective when opportunities are dense, while tree-search and evolutionary strategies tend to be more effective when opportunities are sparse.
• Score: 24.125726163497742
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: AI research agents accelerate ML research by automating hypothesis generation, experimentation, and empirical refinement. Existing agent strategies range from greedy hill-climbing to tree search and evolutionary optimization, yet which strategy choices drive performance remains unclear. Answering this question requires a benchmark that separates agent strategy (e.g., search topology) from execution infrastructure (e.g., code editor), so that performance differences are attributable to strategy rather than infrastructure, and that provides process-level metrics beyond final scores to analyze exploration behaviors. Existing benchmarks offer limited support. We propose FML-Bench, a benchmark of 18 fundamental ML research tasks across 10 domains that separates agent strategy from execution infrastructure and defines 12 process-level behavioral metrics. Evaluating six representative agents, we find that: (1) strategy complexity alone does not guarantee strong performance: a simple greedy hill-climber nearly matches the best-performing tree-search agent, both well above the remaining agents; (2) our analysis suggests this pattern relates to improvement opportunity structure: greedy search tends to be more effective when opportunities are dense, while tree-search and evolutionary strategies tend to be more effective when opportunities are sparse; an adaptive agent built on this insight switches to broader exploration upon detecting improvement stagnation and outperforms the other six agents, lending initial support to this observation; and (3) process-level analysis reveals that early convergence and directionally focused exploration are significantly associated with final performance, while solution diversity and compute cost are not. Our benchmark is available at: https://github.com/qrzou/FML-bench.

Related papers

• The Optimizer Is the Agent: Reasoning-Driven Search across Prompts, Programs, and ML Workflows [72.9051182568928]
  We present ReASearch, a unified framework for reasoning-driven optimization.<n>With a shared agent loop and domain-specific tools, ReASearch instantiates the exact same scaffold to optimize prompts, programs, and ML.<n>We observe that complex search behaviors emerge naturally from the agent's reasoning process.
  arXiv  Detail & Related papers  (2026-08-07T02:15:39Z)
• CoSearch: Joint Training of Reasoning and Document Ranking via Reinforcement Learning for Agentic Search [51.911048955965136]
  CoSearch is a framework that jointly trains a multi-step reasoning agent and a generative document ranking model.<n>Our results show that joint training of the reasoning agent and retrieval system is both feasible and strongly performant.
  arXiv  Detail & Related papers  (2026-04-19T17:48:17Z)
• Beyond Stochastic Exploration: What Makes Training Data Valuable for Agentic Search [5.481048072631037]
  We propose a novel framework, Hierarchical Experience (HiExp), to enhance the performance and training of search agents.<n>We extract empirical knowledge through contrastive analysis and a multi-level clustering mechanism, transforming raw reasoning trajectories into hierarchical experience knowledge.<n>Our approach achieves substantial performance gains and also exhibits strong cross-task and cross-algorithm generalization.
  arXiv  Detail & Related papers  (2026-04-09T11:44:44Z)
• Search More, Think Less: Rethinking Long-Horizon Agentic Search for Efficiency and Generalization [64.61432234404276]
  emphSearch More, Think Less (SMTL) is a framework for long-horizon agentic search that targets both efficiency and generalization.<n>We train an end-to-end agent using supervised fine-tuning and reinforcement learning, achieving strong and often state of the art performance across benchmarks.
  arXiv  Detail & Related papers  (2026-02-26T06:46:41Z)
• TreePS-RAG: Tree-based Process Supervision for Reinforcement Learning in Agentic RAG [71.06073770344732]
  Agentic retrieval-augmented generation (RAG) formulates question answering as a multi-step interaction between reasoning and information retrieval.<n>We present TreePS-RAG, an online, tree-based RL framework for agentic RAG that enables step-wise credit assignment while retaining outcome-only rewards.
  arXiv  Detail & Related papers  (2026-01-11T14:07:30Z)
• Beyond Fast and Slow: Cognitive-Inspired Elastic Reasoning for Large Language Models [39.03483371038282]
  CogER is a framework inspired by human hierarchical reasoning.<n>For queries requiring external tools, we introduce Cognitive Tool-Assisted Reasoning.<n>CogER outperforms state-of-the-art Test-Time scaling methods.
  arXiv  Detail & Related papers  (2025-12-17T05:11:58Z)
• AgentSwift: Efficient LLM Agent Design via Value-guided Hierarchical Search [58.98450205734779]
  Large language model (LLM) agents have demonstrated strong capabilities across diverse domains.<n>Existing agent search methods suffer from three major limitations.<n>We introduce a comprehensive framework to address these challenges.
  arXiv  Detail & Related papers  (2025-06-06T12:07:23Z)
• Learning to reset in target search problems [0.6554326244334868]
  We propose a reinforcement learning framework to train agents capable of optimizing their search efficiency in environments by learning how to reset.<n>First, we validate the approach in a well-established benchmark: the Brownian search with resetting.<n>We then extend the framework by allowing agents to control not only when to reset, but also their spatial dynamics through turning actions.
  arXiv  Detail & Related papers  (2025-03-14T11:57:51Z)
• Alchemy: A structured task distribution for meta-reinforcement learning [52.75769317355963]
  We introduce a new benchmark for meta-RL research, which combines structural richness with structural transparency. Alchemy is a 3D video game, which involves a latent causal structure that is resampled procedurally from episode to episode. We evaluate a pair of powerful RL agents on Alchemy and present an in-depth analysis of one of these agents.
  arXiv  Detail & Related papers  (2021-02-04T23:40:44Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.