REVERE: Reflective Evolving Research Engineer for Scientific Workflows

• URL: http://arxiv.org/abs/2603.20667v1
• Date: Sat, 21 Mar 2026 05:58:30 GMT
• Title: REVERE: Reflective Evolving Research Engineer for Scientific Workflows
• Abstract summary: Existing prompt-optimization techniques rely on local signals to update behavior, often broader and recurring patterns across tasks, leading to poor generalization.<n>We introduce Reflective Evolving Research Engineer (REVERE), a framework that continuously learns from Global Training Context.<n>REVERE, via this reflective optimization framework, improves performance over prior state-of-the-art expert-crafted instructions on research coding tasks by 4.50% on SUPER, 3.51% on ResearchCodeBench, and 4.89% on ScienceAgentBench.
• Score: 44.36828076189902
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: Existing prompt-optimization techniques rely on local signals to update behavior, often neglecting broader and recurring patterns across tasks, leading to poor generalization; they further rely on full-prompt rewrites or unstructured merges, resulting in knowledge loss. These limitations are magnified in research-coding workflows, which involve heterogeneous repositories, underspecified environments, and weak feedback, where reproducing results from public codebases is an established evaluation regime. We introduce Reflective Evolving Research Engineer (REVERE), a framework that continuously learns from Global Training Context, recognizes recurring failure modes in cross-repository execution trajectories, distills them into reusable heuristics, and performs targeted edits across three configurable fields: the system prompt, a task-prompt template, and a cumulative cheatsheet. REVERE, via this reflective optimization framework, improves performance over prior state-of-the-art expert-crafted instructions on research coding tasks by 4.50% on SUPER, 3.51% on ResearchCodeBench, and 4.89% on ScienceAgentBench across their respective metrics. These results demonstrate that agents equipped with mechanisms for continual learning and global memory consolidation can meaningfully evolve their capabilities over time.

Related papers

• ARISE-RL: Agentic Rubric-Grounded Iterative Self-Evolution with Reinforcement Learning [74.19730275669227]
  ARISE-RL is a novel full-cycle self-evolution framework for training open-ended agents.<n>We present Reward-Gated Self-Evolution Distillation (RG-SED), which distills a memory-augmented variant of the same policy back into itself.<n>We also present ECR-Bench, an expert-calibrated rubric benchmark suite covering single-tool deep research and multi-tool travel planning.
  arXiv  Detail & Related papers  (2026-09-01T10:54:13Z)
• MetaResearcher: Scaling Deep Research via Self-Reflective Reinforcement Learning in Adversarial Virtual Environments [15.961083103517872]
  We propose a novel framework that scales deep research agent training across four synergistic dimensions.<n>First, we introduce an Evolving Virtual World that injects temporal dynamics and adversarial misinformation into the training environment.<n>Second, we design Discovery-Oriented Tasks that transcend simple fact retrieval and push agents toward genuine research behaviors.<n>Third, we propose a Self-Reflective Meta-Reward mechanism that jointly optimize for answer correctness, search path efficiency, reflection depth, and tool call diversity.
  arXiv  Detail & Related papers  (2026-06-18T07:50:40Z)
• NanoResearch: Co-Evolving Skills, Memory, and Policy for Personalized Research Automation [59.19858972597063]
  LLM-powered multi-agent systems can now automate the full research pipeline from ideation to paper writing, but a fundamental question remains: automation for whom?<n>We propose NanoResearch, a multi-agent framework that addresses these gaps through tri-level co-evolution.<n>A skill bank distills recurring operations into compact procedural rules reusable across projects.<n>A memory module maintains user- and project-specific experience that grounds planning decisions in each user's research history.<n>A label-free policy learning converts free-form feedback into persistent parameter updates of the planner, reshaping subsequent coordination.
  arXiv  Detail & Related papers  (2026-05-11T16:33:47Z)
• Trajectory-Informed Memory Generation for Self-Improving Agent Systems [4.933717407152962]
  LLM-powered agents face a persistent challenge: learning from their execution experiences to improve future performance.<n>We present a novel framework for automatically extracting actionable learnings from agent execution trajectories.<n>Our framework understands execution patterns, extracts structured learnings with provenance, and retrieves guidance tailored to specific task contexts.
  arXiv  Detail & Related papers  (2026-03-11T09:54:09Z)
• Understanding by Reconstruction: Reversing the Software Development Process for LLM Pretraining [66.89012795621349]
  Large Language Models (LLMs) often struggle with the deep, long-horizon reasoning required for complex software engineering.<n>We propose a novel paradigm: understanding via reconstruction.<n>We introduce a framework that synthesizes latent agentic trajectories using a multi-agent simulation.
  arXiv  Detail & Related papers  (2026-03-11T09:23:20Z)
• BeyondSWE: Can Current Code Agent Survive Beyond Single-Repo Bug Fixing? [61.247730037229815]
  We introduce BeyondSWE, a comprehensive benchmark that broadens existing evaluations along two axes - resolution scope and knowledge scope.<n>To investigate the role of external knowledge, we develop SearchSWE, a framework that integrates deep search with coding abilities.<n>This work offers both a realistic, challenging evaluation benchmark and a flexible framework to advance research toward more capable code agents.
  arXiv  Detail & Related papers  (2026-03-03T17:52:01Z)
• Evaluating Novelty in AI-Generated Research Plans Using Multi-Workflow LLM Pipelines [1.3986052226424095]
  This paper investigates whether agentic systems employing iterative reasoning, evolutionary search, and decomposition can generate more novel and feasible research plans.<n>We benchmark five reasoning architectures: Reflection-based iterative refinement, Sakana AI v2 evolutionary algorithms, Google Co-Scientist multi-agent framework, GPT Deep Research, and Gemini3 Pro multimodal long-context pipeline.<n>Results reveal varied performance across research domains, with high-performing maintaining feasibility without sacrificing creativity.
  arXiv  Detail & Related papers  (2025-12-24T12:41:31Z)
• RECODE-H: A Benchmark for Research Code Development with Interactive Human Feedback [87.97664892075811]
  We present RECODE-H, a benchmark of 102 tasks from research papers and repositories.<n>It includes structured instructions,unit tests, and a five-level feedback hierarchy to reflect realistic researcher-agent collaboration.<n>We also present ReCodeAgent, a framework that integrates feedback into iterative code generation.
  arXiv  Detail & Related papers  (2025-10-07T17:45:35Z)
• SaraCoder: Orchestrating Semantic and Structural Cues for Resource-Optimized Repository-Level Code Completion [34.41683042851225]
  We propose a resource-optimized retrieval augmentation method, SaraCoder.<n>It maximizes information diversity and representativeness in a limited context window.<n>Our work proves that systematically refining retrieval results across multiple dimensions provides a new paradigm for building more accurate and resource-optimized repository-level code completion systems.
  arXiv  Detail & Related papers  (2025-08-13T11:56:05Z)

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.