Heuresis: Search Strategies for Autonomous AI Research Agents Across Quality, Diversity and Novelty

• URL: http://arxiv.org/abs/2606.25198v2
• Date: Wed, 01 Jul 2026 17:23:10 GMT
• Title: Heuresis: Search Strategies for Autonomous AI Research Agents Across Quality, Diversity and Novelty
• Abstract summary: Heuresis is a framework that abstracts the research pipeline into a set of general and composable primitives.<n>We find that completely novel ideas are rare. No idea across our scored runs is rated as "Original", and only a few achieve only "Minor Similarity" to prior work.
• Score: 39.97857268041792
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: Autonomous AI Research promises to accelerate the scientific progress of machine learning. To realise this goal, current Large Language Model (LLM)-based agents need to go beyond just writing code, to mastering the exploration of simultaneously performant, diverse and novel ideas. To this end, we introduce Heuresis, a framework that abstracts the research pipeline into a set of general and composable primitives, enabling open-ended scientific exploration in machine learning research. We implement six search strategies: a greedy baseline, two archive-based (MAP-Elites, Go-Explore), one evolutionary (Islands), and two divergent (Curiosity, Omni), and evaluate them across three axes (Quality, Diversity, and Novelty) on three domains (LLM Pretraining, On-Policy RL, and Model Unlearning), totalling 3,222 scored runs. We find that completely novel ideas are rare. No idea across our scored runs is rated as "Original", and only a few achieve only "Minor Similarity" to prior work. Moreover, novel ideas never approach the highest-performing known-recipe scores. Across all six strategies and three domains, only one such idea lands in the top-10 by quality. We also observed agents resorting to a variety of reward-hacking techniques during execution (40 confirmed fabrications across 1,628 scored runs), and detecting them was necessary to keep the search faithful to the task. Our results show that while current search and Quality-Diversity strategies enable us to steer where the generated ideas land on the quality, diversity, and novelty axes, they do not expand the quality-novelty frontier. Bridging this gap is the open challenge towards the ultimate goal of perpetual, autonomous scientific progress. Code is available at github.com/a-antoniades/Heuresis.

Related papers

• Personalized Auto-Research: Towards a True AI Co-Scientist [90.01790420664169]
  Given a research goal, state-of-the-art systems remain researcher-agnostic.<n>What counts as novel, valuable, or feasible depends on the researcher.<n>We propose a framework that threads a graph-grounded researcher context through retrieval, hypothesis search, experimentation, writing, and review.
  arXiv  Detail & Related papers  (2026-08-14T20:43:27Z)
• IDEAgent: Agentic Quality-Diversity Search for Research Idea Generation [43.08002957114291]
  We argue that research ideation should be treated as a conjunction of both objectives and framed as a Quality-Diversity (QD) search.<n>We jointly drive Quality using multi-objective feedback for dedicated repair and refinement, while Diversity is achieved through lightweight sequential memory.<n>We show that IDEAgent outperforms the best baseline by 3.89x on Yield, while achieving non-zero Yield on 8x more topics.
  arXiv  Detail & Related papers  (2026-07-24T15:03:14Z)
• AI Research Agents Narrow Scientific Exploration [12.949322198287417]
  We generate 37,802 scientific ideas from shared seed literature across citation-defined research areas in AI and machine learning.<n>We compare the resulting AI ideas against human-authored papers from the same research areas, follow-on human research emerging from the same seed literature, and the seed literature itself.<n>Overall, current AI research agents appear better suited to local elaboration than to broadening scientific exploration.
  arXiv  Detail & Related papers  (2026-05-27T03:26:43Z)
• GEAR: Genetic AutoResearch for Agentic Code Evolution [11.641689406945202]
  GEAR, or Genetic AutoResearch, replaces a single-path search with a population-based search over multiple research states.<n>It keeps a set of strong candidate solutions, selects parents based on productivity, novelty, and coverage, and explores new ideas through mutation and crossover.<n>The paper studies three versions of GEAR: one controlled through prompting, one using a fixed programmatic search controller, and one where the controller itself can evolve during the run.
  arXiv  Detail & Related papers  (2026-05-08T00:25:09Z)
• AutoResearchBench: Benchmarking AI Agents on Complex Scientific Literature Discovery [55.70879973230979]
  AutoResearchBench is a benchmark for autonomous scientific literature discovery.<n>Compared to previous benchmarks on agentic web browsing, AutoResearchBench is research-oriented.<n>Even the most powerful LLMs, despite having largely conquered general agentic web-browsing benchmarks such as BrowseComp, achieve only 9.39% accuracy on Deep Research and 9.31% IoU on Wide Research.
  arXiv  Detail & Related papers  (2026-04-28T06:05:17Z)
• Enhancing Research Idea Generation through Combinatorial Innovation and Multi-Agent Iterative Search Strategies [9.063254770910948]
  This study proposes a multi-agent iterative planning search strategy inspired by com-binatorial innovation theory.<n>The framework combines iterative knowledge search with an LLM-based multi-agent system to generate, evaluate, and re-fine research ideas.<n>Experiments in the natural language processing domain show that the proposed method outperforms state-of-the-art base-lines in both diversity and novelty.
  arXiv  Detail & Related papers  (2026-04-22T13:31:12Z)
• CORAL: Towards Autonomous Multi-Agent Evolution for Open-Ended Discovery [67.47220507982522]
  Large language model (LLM)-based evolution is a promising approach for openended discovery.<n>We present CORAL, the first framework for autonomous multi-agent evolution on open-ended problems.<n>CORAL replaces rigid control with long-running agents that explore, reflect, and collaborate.
  arXiv  Detail & Related papers  (2026-04-02T05:59:06Z)
• AI Co-Scientist for Ranking: Discovering Novel Search Ranking Models alongside LLM-based AI Agents with Cloud Computing Access [60.37646376282544]
  We present an AI Co-Scientist framework that automates the full search ranking research pipeline.<n>This is the first study in the ranking community to utilize an AI Co-Scientist framework for algorithmic research.
  arXiv  Detail & Related papers  (2026-03-23T09:29:08Z)
• Training AI Co-Scientists Using Rubric Rewards [36.857905469178796]
  A crucial feature of AI co-scientists is the ability to generate a research plan given a set of aims and constraints.<n>In this work, we study how to leverage the vast corpus of existing research papers to train language models that generate better research plans.<n>We build a scalable, diverse training corpus by automatically extracting research goals and goal-specific grading rubrics from papers across several domains.
  arXiv  Detail & Related papers  (2025-12-29T18:59:33Z)
• Reinforcement Learning for Generative AI: State of the Art, Opportunities and Open Research Challenges [2.9008806248012333]
  Reinforcement Learning (RL) has emerged as a very successful paradigm for a variety of machine learning tasks. We will discuss three types of applications, namely, RL as an alternative way for generation without specified objectives. We conclude the survey with an in-depth discussion of the opportunities and challenges in this fascinating emerging area.
  arXiv  Detail & Related papers  (2023-07-31T18:00:02Z)
• Discovering and Achieving Goals via World Models [61.95437238374288]
  We introduce Latent Explorer Achiever (LEXA), a unified solution to this problem. LEXA learns a world model from image inputs and uses it to train an explorer and an achiever policy from imagined rollouts. After the unsupervised phase, LEXA solves tasks specified as goal images zero-shot without any additional learning.
  arXiv  Detail & Related papers  (2021-10-18T17:59:58Z)
• Cooperative Exploration for Multi-Agent Deep Reinforcement Learning [127.4746863307944]
  We propose cooperative multi-agent exploration (CMAE) for deep reinforcement learning. The goal is selected from multiple projected state spaces via a normalized entropy-based technique. We demonstrate that CMAE consistently outperforms baselines on various tasks.
  arXiv  Detail & Related papers  (2021-07-23T20:06:32Z)

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.