# DeepInnovator: Triggering the Innovative Capabilities of LLMs

arXiv: https://arxiv.org/abs/2602.18920 · HTML: https://mlsanity.com/paper/2602.18920/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Placed in Web, search, and research agents, one of 16 directions in LLM agents and tool use.
- Builds on 32 of its 45 references that are in the map.
- First posted on arXiv 21 Feb 2026.

## Placed in

- LLM agents and tool use / [Web, search, and research agents](https://mlsanity.com/agents/web-search-and-research-agents/)

## Builds on (most cited first, 30 of 32)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Qwen Technical Report](https://mlsanity.com/paper/2309.16609/) (2023)
- [HybridFlow: A Flexible and Efficient RLHF Framework](https://mlsanity.com/paper/2409.19256/) (2024)
- [Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning](https://mlsanity.com/paper/2503.09516/) (2025)
- [Scaling Laws for Reward Model Overoptimization](https://mlsanity.com/paper/2210.10760/) (2022)
- [AlphaEvolve: A coding agent for scientific and algorithmic discovery](https://mlsanity.com/paper/2506.13131/) (2025)
- [RLAIF vs. RLHF: Scaling Reinforcement Learning from Human Feedback with AI Feedback](https://mlsanity.com/paper/2309.00267/) (2023)
- [Eureka: Human-Level Reward Design via Coding Large Language Models](https://mlsanity.com/paper/2310.12931/) (2023)
- [The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Search](https://mlsanity.com/paper/2504.08066/) (2025)
- [Rubrics as Rewards: Reinforcement Learning Beyond Verifiable Domains](https://mlsanity.com/paper/2507.17746/) (2025)
- [Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers](https://mlsanity.com/paper/2409.04109/) (2024)
- [DeepResearcher: Scaling Deep Research via Reinforcement Learning in Real-world Environments](https://mlsanity.com/paper/2504.03160/) (2025)
- [Seed1.5-Thinking: Advancing Superb Reasoning Models with Reinforcement Learning](https://mlsanity.com/paper/2504.13914/) (2025)
- [Natural Emergent Misalignment from Reward Hacking in Production RL](https://mlsanity.com/paper/2511.18397/) (2025)
- [Checklists Are Better Than Reward Models For Aligning Language Models](https://mlsanity.com/paper/2507.18624/) (2025)
- [CollabLLM: From Passive Responders to Active Collaborators](https://mlsanity.com/paper/2502.00640/) (2025)
- [DR Tulu: Reinforcement Learning with Evolving Rubrics for Deep Research](https://mlsanity.com/paper/2511.19399/) (2025)
- [Reinforcement Learning with Rubric Anchors](https://mlsanity.com/paper/2508.12790/) (2025)
- [AI-Researcher: Autonomous Scientific Innovation](https://mlsanity.com/paper/2505.18705/) (2025)
- [Deep Research: A Survey of Autonomous Research Agents](https://mlsanity.com/paper/2508.12752/) (2025)
- [DeepScientist: Advancing Frontier-Pushing Scientific Findings Progressively](https://mlsanity.com/paper/2509.26603/) (2025)
- [One Token to Fool LLM-as-a-Judge](https://mlsanity.com/paper/2507.08794/) (2025)
- [A Comprehensive Survey of Deep Research: Systems, Methodologies, and Applications](https://mlsanity.com/paper/2506.12594/) (2025)
- [Harnessing Uncertainty: Entropy-Modulated Policy Gradients for Long-Horizon LLM Agents](https://mlsanity.com/paper/2509.09265/) (2025)
- [A Critical Evaluation of AI Feedback for Aligning Large Language Models](https://mlsanity.com/paper/2402.12366/) (2024)
- [Training AI Co-Scientists Using Rubric Rewards](https://mlsanity.com/paper/2512.23707/) (2025)
- [Probing Scientific General Intelligence of LLMs with Scientist-Aligned Workflows](https://mlsanity.com/paper/2512.16969/) (2025)
- [ReCode: Reinforcing Code Generation with Reasoning-Process Rewards](https://mlsanity.com/paper/2508.05170/) (2025)
- [Augmenting Scientific Creativity with Retrieval across Knowledge Domains](https://mlsanity.com/paper/2206.01328/) (2022)

## Built on by (most cited first, 5 of 5)

- [AI for Auto-Research: Roadmap & User Guide](https://mlsanity.com/paper/2605.18661/) (2026)
- [Intern-Atlas: A Methodological Evolution Graph as Research Infrastructure for AI Scientists](https://mlsanity.com/paper/2604.28158/) (2026)
- [PaperGym: Rubric-Centered Evolution for Research-Plan Generation](https://mlsanity.com/paper/2608.31119/) (2026)
- [pAI/MSc: ML Theory Research with Humans on the Loop](https://mlsanity.com/paper/2604.20622/) (2026)
- [Learning to Ideate for Scientific Impact](https://mlsanity.com/paper/2609.29802/) (2026)

## Abstract (arXiv)

The application of Large Language Models (LLMs) in accelerating scientific discovery has garnered increasing attention, with a key focus on constructing research agents endowed with innovative capability, i.e., the ability to autonomously generate novel and significant research ideas. Existing approaches predominantly rely on sophisticated prompt engineering and lack a systematic training paradigm. To address this, we propose DeepInnovator, a training framework designed to trigger the innovative capability of LLMs. Our approach comprises two core components. (1) ``Standing on the shoulders of giants''. We construct an automated data extraction pipeline to extract and organize structured research knowledge from a vast corpus of unlabeled scientific literature. (2) ``Conjectures and refutations''. We introduce a ``Next Idea Prediction'' training paradigm, which models the generation of research ideas as an iterative process of continuously predicting, evaluating, and refining plausible and novel next idea. Both automatic and expert evaluations demonstrate that our DeepInnovator-14B significantly outperforms untrained baselines, achieving win rates of 80.53%-93.81%, and attains performance comparable to that of current leading LLMs. This work provides a scalable training pathway toward building research agents with genuine, originative innovative capability, and will open-source the dataset to foster community advancement. Source code and data are available at: https://github.com/HKUDS/DeepInnovator.