mlsanityData to 7 Oct 2026

MoRI: Learning Motivation-Grounded Reasoning for Scientific Ideation in Large Language Models

arXiv 2603.19044 ↗First posted 19 Mar 2026cs.CLcited by 0

On the map: LLM reasoning → Knowledge- and tool-augmented reasoning

Abstract (arXiv)

Scientific ideation aims to propose novel solutions within a given scientific context. Existing LLM-based agentic approaches emulate human research workflows, yet inadequately model scientific reasoning, resulting in surface-level conceptual recombinations that lack technical depth and scientific grounding. To address this issue, we propose MoRI (Motivation-grounded Reasoning for Scientific Ideation), a framework that enables LLMs to explicitly learn the reasoning process from research motivations to methodologies. The base LLM is initialized via supervised fine-tuning to generate a research motivation from a given context, and is subsequently trained under a composite reinforcement learning reward that approximates scientific rigor: (1) entropy-aware information gain encourages the model to uncover and elaborate high-complexity technical details grounded in ground-truth methodologies, and (2) contrastive semantic gain constrains the reasoning trajectory to remain conceptually aligned with scientifically valid solutions. Empirical results show that MoRI consistently outperforms strong commercial LLMs and complex agentic baselines across multiple dimensions, including novelty, technical rigor, and feasibility. The code is available on hrefhttps://github.com/ECNU-Text-Computing/IdeaGenerationGitHub.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Knowledge- and tool-augmented reasoning, one of 17 directions in LLM reasoning. Builds on 29 of its 48 references that are in the map. First posted on arXiv 19 Mar 2026.

Builds on

29 of 48 references.

Closest ancestors · 2

Same direction, newest first

  1. 2025AI-Researcher: Autonomous Scientific Innovation24 May
  2. 2025Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning12 Mar

Other references · 27

  1. 2023GPT-4 Technical Reportcited by 18,132
  2. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 13,958
  3. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 8,851
  4. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,146
  5. 2025DAPO: An Open-Source LLM Reinforcement Learning System at Scalecited by 2,678
  6. 2024OpenAI o1 System Cardcited by 1,220
  7. 2024The AI Scientist: Towards Fully Automated Open-Ended Scientific Discoverycited by 1,002
  8. 2025Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoningcited by 540
  9. 2025ReTool: Reinforcement Learning for Strategic Tool Use in LLMscited by 377
  10. 2025The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Searchcited by 375
  11. 2024Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researcherscited by 348
  12. 2025Rubrics as Rewards: Reinforcement Learning Beyond Verifiable Domainscited by 328
15 more
  1. 2024MinerU: An Open-Source Solution for Precise Document Content Extractioncited by 255
  2. 2024CycleResearcher: Improving Automated Research via Automated Reviewcited by 142
  3. 2025DR Tulu: Reinforcement Learning with Evolving Rubrics for Deep Researchcited by 98
  4. 2025RLPR: Extrapolating RLVR to General Domains without Verifierscited by 83
  5. 2025Jointly Reinforcing Diversity and Quality in Language Model Generationscited by 79
  6. 2025Reinforcing General Reasoning without Verifierscited by 76
  7. 2025One Token to Fool LLM-as-a-Judgecited by 56
  8. 2025Information Gain-based Policy Optimization: A Simple and Effective Approach for Multi-Turn Search Agentscited by 43
  9. 2025Learning to Reason for Long-Form Story Generationcited by 34
  10. 2025Language Models that Think, Chat Bettercited by 23
  11. 2025ThinkDial: An Open Recipe for Controlling Reasoning Effort in Large Language Modelscited by 11
  12. 2025SCI-IDEA: Context-Aware Scientific Ideation Using Token and Sentence Embeddingscited by 7
  13. 2025Measuring Combinatorial Creativity in Large Language Models and Humanscited by 7
  14. 2025Adaptive Reinforcement for Open-ended Medical Reasoning via Semantic-Guided Reward Collapse Mitigationcited by 4
  15. 2025MotivGraph-SoIQ: Integrating Motivational Knowledge Graphs and Socratic Dialogue for Enhanced LLM Ideationcited by 4

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Debate as Reward: A Multi-Agent Reward System for Scientific Ideation via RL Post-Training
  2. 2026IdeaScientist: Orchestrating Agents for Grounded Scientific Ideation
  3. 2026COGTRL: Training LLMs for Scientific Discovery Assistance using Cognitive Traces via Reinforcement Learning
  4. 2026FALSIFYBENCH: Evaluating Inductive Reasoning in LLMs with Rule Discovery Games
  5. 2025SciReasoner: Laying the Scientific Reasoning Ground Across Disciplines
  6. 2026Teaching Language Models to Forecast Research Success Through Comparative Idea Evaluation
  7. 2026AI scientists produce results without reasoning scientifically
  8. 2026Reinforcing Agentic Creativity in Scientific Ideation with Night Science