Towards Execution-Grounded Automated AI Research

• URL: http://arxiv.org/abs/2601.14525v1
• Date: Tue, 20 Jan 2026 22:35:44 GMT
• Title: Towards Execution-Grounded Automated AI Research
• Authors: Chenglei Si, Zitong Yang, Yejin Choi, Emmanuel Candès, Diyi Yang, Tatsunori Hashimoto,
• Abstract summary: Execution grounding may help, but it is unclear whether automated execution is feasible and whether LLMs can learn from the execution feedback.<n>We build an automated executor to implement ideas and launch large-scale parallel GPU experiments to verify their effectiveness.<n>We analyze two methods to learn from the execution feedback: evolutionary search and reinforcement learning.
• Score: 106.90422658528819
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Automated AI research holds great potential to accelerate scientific discovery. However, current LLMs often generate plausible-looking but ineffective ideas. Execution grounding may help, but it is unclear whether automated execution is feasible and whether LLMs can learn from the execution feedback. To investigate these, we first build an automated executor to implement ideas and launch large-scale parallel GPU experiments to verify their effectiveness. We then convert two realistic research problems - LLM pre-training and post-training - into execution environments and demonstrate that our automated executor can implement a large fraction of the ideas sampled from frontier LLMs. We analyze two methods to learn from the execution feedback: evolutionary search and reinforcement learning. Execution-guided evolutionary search is sample-efficient: it finds a method that significantly outperforms the GRPO baseline (69.4% vs 48.0%) on post-training, and finds a pre-training recipe that outperforms the nanoGPT baseline (19.7 minutes vs 35.9 minutes) on pre-training, all within just ten search epochs. Frontier LLMs often generate meaningful algorithmic ideas during search, but they tend to saturate early and only occasionally exhibit scaling trends. Reinforcement learning from execution reward, on the other hand, suffers from mode collapse. It successfully improves the average reward of the ideator model but not the upper-bound, due to models converging on simple ideas. We thoroughly analyze the executed ideas and training dynamics to facilitate future efforts towards execution-grounded automated AI research.

Related papers

• MM-HELIX: Boosting Multimodal Long-Chain Reflective Reasoning with Holistic Platform and Adaptive Hybrid Policy Optimization [103.74675519953898]
  Long-chain reflective reasoning is a prerequisite for solving complex real-world problems.<n>We build a benchmark consisting 1,260 samples of 42 challenging synthetic tasks.<n>We generate post-training data and explore learning paradigms for exploiting such data.
  arXiv  Detail & Related papers  (2025-10-09T17:53:58Z)
• The Automated LLM Speedrunning Benchmark: Reproducing NanoGPT Improvements [87.61432174951891]
  A critical capability toward scientific progress is the ability to reproduce existing work.<n>To evaluate the ability of AI agents to reproduce results in an active research area, we introduce the Automated LLM Speedrunning Benchmark.<n>We find that recent reasoning LLMs combined with SoTA scaffolds struggle to reimplement already-known innovations in our benchmark.
  arXiv  Detail & Related papers  (2025-06-27T17:44:32Z)
• Runaway is Ashamed, But Helpful: On the Early-Exit Behavior of Large Language Model-based Agents in Embodied Environments [54.67512489842682]
  Large language models (LLMs) have demonstrated strong planning and decision-making capabilities in complex embodied environments.<n>We take a first step toward exploring the early-exit behavior for LLM-based agents.
  arXiv  Detail & Related papers  (2025-05-23T08:23:36Z)
• Think or Not? Selective Reasoning via Reinforcement Learning for Vision-Language Models [67.87579664988199]
  TON is a two-stage training strategy for vision-language models (VLMs)<n>It introduces a think-or-not format that serves as a cold start for selective reasoning.<n>TON can reduce the completion length by up to 90% compared to vanilla GRPO.
  arXiv  Detail & Related papers  (2025-05-22T16:13:29Z)
• Escaping Collapse: The Strength of Weak Data for Large Language Model Training [15.77316232527746]
  We develop a theoretical framework to investigate how much curation is needed in order to ensure that LLM performance continually improves.<n>We describe a training procedure that converges to an optimal LLM even if almost all of the non-synthetic training data is of poor quality.
  arXiv  Detail & Related papers  (2025-02-13T03:20:37Z)
• O1 Embedder: Let Retrievers Think Before Action [28.583031173137428]
  We propose O1 Embedder, which generates useful thoughts for the input query before making retrieval for the target documents.<n>Our approach is evaluated by comprehensive experiments, where substantial improvements are achieved across 12 popular datasets.<n>These results highlight O1 Embedder's remarkable accuracy and generalizability, paving the way for the development of next-generation IR foundation models.
  arXiv  Detail & Related papers  (2025-02-11T13:48:10Z)
• Efficient Reinforcement Learning via Decoupling Exploration and Utilization [6.305976803910899]
  Reinforcement Learning (RL) has achieved remarkable success across multiple fields and applications, including gaming, robotics, and autonomous vehicles. In this work, our aim is to train agent with efficient learning by decoupling exploration and utilization, so that agent can escaping the conundrum of suboptimal Solutions. The above idea is implemented in the proposed OPARL (Optimistic and Pessimistic Actor Reinforcement Learning) algorithm.
  arXiv  Detail & Related papers  (2023-12-26T09:03:23Z)
• MLAgentBench: Evaluating Language Agents on Machine Learning Experimentation [96.71370747681078]
  We introduce MLAgentBench, a suite of 13 tasks ranging from improving model performance on CIFAR-10 to recent research problems like BabyLM. For each task, an agent can perform actions like reading/writing files, executing code, and inspecting outputs. We benchmark agents based on Claude v1.0, Claude v2.1, Claude v3 Opus, GPT-4, GPT-4-turbo, Gemini-Pro, and Mixtral and find that a Claude v3 Opus agent is the best in terms of success rate.
  arXiv  Detail & Related papers  (2023-10-05T04:06:12Z)
• Rethinking Closed-loop Training for Autonomous Driving [82.61418945804544]
  We present the first empirical study which analyzes the effects of different training benchmark designs on the success of learning agents. We propose trajectory value learning (TRAVL), an RL-based driving agent that performs planning with multistep look-ahead. Our experiments show that TRAVL can learn much faster and produce safer maneuvers compared to all the baselines.
  arXiv  Detail & Related papers  (2023-06-27T17:58:39Z)

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.