LDC: Learning to Generate Research Idea with Dynamic Control

TL;DR

This paper introduces LDC, a novel two-stage framework combining supervised fine-tuning and controllable reinforcement learning to generate high-quality research ideas that balance novelty, feasibility, and effectiveness.

Contribution

It is the first to integrate multi-dimensional reward models with dynamic control for research idea generation, improving alignment with expert standards.

Findings

Achieved balanced research idea generation with high quality.

Effectively navigated trade-offs among key idea dimensions.

Demonstrated superior performance over existing prompting methods.

Abstract

Recent advancements in large language models (LLMs) have demonstrated their potential in automating the scientific research ideation. Existing approaches primarily focus on prompting techniques, often producing ideas misaligned with expert standards - novelty, feasibility, and effectiveness, which are widely recognized by the research community as the three key subdimensions of high-quality ideas. Also, balancing these dimensions remains challenging due to their inherent trade-offs. To address these limitations, we propose the first framework that employs a two-stage approach combining Supervised Fine-Tuning (SFT) and controllable Reinforcement Learning (RL) for the task. In the SFT stage, the model learns foundational patterns from pairs of research papers and their corresponding follow-up ideas. In the RL stage, multi-dimensional reward models guided by fine-grained feedback evaluate and optimize the model across key dimensions. During inference, dimensional controllers coordinated by a sentence-level decoder enable dynamic context-aware steering of the idea generation process. Our framework provides a balanced approach to research idea generation, achieving high-quality outcomes in the experiment by dynamically navigating the trade-offs among novelty, feasibility, and effectiveness.