What Makes AI Research Replicable? Executable Knowledge Graphs as Scientific Knowledge Representations

TL;DR

This paper introduces Executable Knowledge Graphs (xKG), a structured, paper-centric knowledge base that enhances AI research replication by integrating code snippets and technical insights, significantly improving performance across multiple frameworks.

Contribution

The paper presents xKG, a novel, extensible knowledge representation that captures technical details and code snippets from scientific literature to improve AI research replication.

Findings

xKG improves replication performance by 10.9% on PaperBench.

Integration of xKG into LLM agents enhances their ability to reproduce AI research.

xKG effectively captures technical details and code snippets from scientific papers.

Abstract

Replicating AI research is a crucial yet challenging task for large language model (LLM) agents. Existing approaches often struggle to generate executable code, primarily due to insufficient background knowledge and the limitations of retrieval-augmented generation (RAG) methods, which fail to capture latent technical details hidden in referenced papers. Furthermore, previous approaches tend to overlook valuable implementation-level code signals and lack structured knowledge representations that support multi-granular retrieval and reuse. To overcome these challenges, we propose Executable Knowledge Graphs (xKG), a pluggable, paper-centric knowledge base that automatically integrates code snippets and technical insights extracted from scientific literature. When integrated into three agent frameworks with two different LLMs, xKG shows substantial performance gains (10.9% with o3-mini) on PaperBench, demonstrating its effectiveness as a general and extensible solution for automated AI research replication. Code is available at https://github.com/zjunlp/xKG.