Can LLMs Identify Critical Limitations within Scientific Research? A Systematic Evaluation on AI Research Papers

TL;DR

This paper introduces LimitGen, a benchmark for evaluating LLMs' ability to identify scientific research limitations, and demonstrates that augmenting LLMs with literature retrieval improves their performance in providing constructive peer review feedback.

Contribution

The paper presents the first comprehensive benchmark for LLMs to identify research limitations and shows that literature retrieval augmentation enhances their effectiveness.

Findings

LimitGen benchmark includes synthetic and real limitations datasets.

Augmentation with literature retrieval improves LLMs' limitation identification.

LLMs can support early-stage peer review with enhanced feedback capabilities.

Abstract

Peer review is fundamental to scientific research, but the growing volume of publications has intensified the challenges of this expertise-intensive process. While LLMs show promise in various scientific tasks, their potential to assist with peer review, particularly in identifying paper limitations, remains understudied. We first present a comprehensive taxonomy of limitation types in scientific research, with a focus on AI. Guided by this taxonomy, for studying limitations, we present LimitGen, the first comprehensive benchmark for evaluating LLMs' capability to support early-stage feedback and complement human peer review. Our benchmark consists of two subsets: LimitGen-Syn, a synthetic dataset carefully created through controlled perturbations of high-quality papers, and LimitGen-Human, a collection of real human-written limitations. To improve the ability of LLM systems to identify limitations, we augment them with literature retrieval, which is essential for grounding identifying limitations in prior scientific findings. Our approach enhances the capabilities of LLM systems to generate limitations in research papers, enabling them to provide more concrete and constructive feedback.