Exploring the Potential of Large Language Models in Graph Generation

Large language models (LLMs) have achieved great success in many fields, and recent works have studied exploring LLMs for graph discriminative tasks such as node classification. However, the abilities of LLMs for graph generation remain unexplored in the literature. Graph generation requires the LLM to generate graphs with given properties, which has valuable real-world applications such as drug discovery, while tends to be more challenging. In this paper, we propose LLM4GraphGen to explore the ability of LLMs for graph generation with systematical task designs and extensive experiments. Specifically, we propose several tasks tailored with comprehensive experiments to address key questions regarding LLMs' understanding of different graph structure rules, their ability to capture structural type distributions, and their utilization of domain knowledge for property-based graph generation. Our evaluations demonstrate that LLMs, particularly GPT-4, exhibit preliminary abilities in graph generation tasks, including rule-based and distribution-based generation. We also observe that popular prompting methods, such as few-shot and chain-of-thought prompting, do not consistently enhance performance. Besides, LLMs show potential in generating molecules with specific properties. These findings may serve as foundations for designing good LLMs based models for graph generation and provide valuable insights and further research.

[1]  Suqi Cheng,et al.  GraphGPT: Graph Instruction Tuning for Large Language Models , 2023, SIGIR.

[2]  Yi Qin,et al.  Graph Meets LLMs: Towards Large Graph Models , 2023, 2308.14522.

[3]  M. Bhattacharya,et al.  Artificial intelligence enabled ChatGPT and large language models in drug target discovery, drug discovery, and development , 2023, Molecular therapy. Nucleic acids.

[4]  Kai Sun,et al.  Head-to-Tail: How Knowledgeable are Large Language Models (LLM)? A.K.A. Will LLMs Replace Knowledge Graphs? , 2023, ArXiv.

[5]  Runhui Wang,et al.  Language is All a Graph Needs , 2023, FINDINGS.

[6]  Claus Stadler,et al.  LLM-assisted Knowledge Graph Engineering: Experiments with ChatGPT , 2023, AI Tomorrow.

[7]  Jiliang Tang,et al.  Exploring the Potential of Large Language Models (LLMs)in Learning on Graphs , 2023, SIGKDD Explor..

[8]  Xindong Wu,et al.  Unifying Large Language Models and Knowledge Graphs: A Roadmap , 2023, IEEE Transactions on Knowledge and Data Engineering.

[9]  Lun Du,et al.  GPT4Graph: Can Large Language Models Understand Graph Structured Data ? An Empirical Evaluation and Benchmarking , 2023, ArXiv.

[10]  Heng Wang,et al.  Can Language Models Solve Graph Problems in Natural Language? , 2023, NeurIPS.

[11]  Wayne Xin Zhao,et al.  StructGPT: A General Framework for Large Language Model to Reason over Structured Data , 2023, EMNLP.

[12]  Jiawei Zhang Graph-ToolFormer: To Empower LLMs with Graph Reasoning Ability via Prompt Augmented by ChatGPT , 2023, ArXiv.

[13]  Wayne Xin Zhao,et al.  A Survey of Large Language Models , 2023, ArXiv.

[14]  Marco Tulio Ribeiro,et al.  Sparks of Artificial General Intelligence: Early experiments with GPT-4 , 2023, ArXiv.

[15]  Dragomir R. Radev,et al.  LEVER: Learning to Verify Language-to-Code Generation with Execution , 2023, ICML.

[16]  V. Cevher,et al.  DiGress: Discrete Denoising diffusion for graph generation , 2022, ICLR.

[17]  Elena L. Glassman,et al.  Expectation vs. Experience: Evaluating the Usability of Code Generation Tools Powered by Large Language Models , 2022, CHI Extended Abstracts.

[18]  Jieyu Zhang,et al.  A Survey on Deep Graph Generation: Methods and Applications , 2022, LoG.

[19]  Amarda Shehu,et al.  Interpretable Molecular Graph Generation via Monotonic Constraints , 2022, SDM.

[20]  Sung Ju Hwang,et al.  Score-based Generative Modeling of Graphs via the System of Stochastic Differential Equations , 2022, ICML.

[21]  Xiangnan He,et al.  Discovering Invariant Rationales for Graph Neural Networks , 2022, ICLR.

[22]  Dale Schuurmans,et al.  Chain of Thought Prompting Elicits Reasoning in Large Language Models , 2022, NeurIPS.

[23]  Michael S. Bernstein,et al.  On the Opportunities and Risks of Foundation Models , 2021, ArXiv.

[24]  Davide Bacciu,et al.  Graphgen-redux: a Fast and Lightweight Recurrent Model for labeled Graph Generation , 2021, 2021 International Joint Conference on Neural Networks (IJCNN).

[25]  Shuiwang Ji,et al.  GraphDF: A Discrete Flow Model for Molecular Graph Generation , 2021, ICML.

[26]  Xiaojie Guo,et al.  A Systematic Survey on Deep Generative Models for Graph Generation , 2020, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[27]  Fei Wang,et al.  MoFlow: An Invertible Flow Model for Generating Molecular Graphs , 2020, KDD.

[28]  Tom B. Brown,et al.  Language Models are Few-Shot Learners , 2020, NeurIPS.

[29]  J. Leskovec,et al.  Open Graph Benchmark: Datasets for Machine Learning on Graphs , 2020, NeurIPS.

[30]  Olgica Milenkovic,et al.  Multi-MotifGAN (MMGAN): Motif-Targeted Graph Generation And Prediction , 2019, ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[31]  J. Leskovec,et al.  GNNExplainer: Generating Explanations for Graph Neural Networks , 2019, NeurIPS.

[32]  Krzysztof Rataj,et al.  Mol-CycleGAN: a generative model for molecular optimization , 2019, Journal of Cheminformatics.

[33]  Jure Leskovec,et al.  Graph Convolutional Policy Network for Goal-Directed Molecular Graph Generation , 2018, NeurIPS.

[34]  Regina Barzilay,et al.  Junction Tree Variational Autoencoder for Molecular Graph Generation , 2018, ICML.

[35]  David Weininger,et al.  SMILES, a chemical language and information system. 1. Introduction to methodology and encoding rules , 1988, J. Chem. Inf. Comput. Sci..

[36]  Q. Mei,et al.  Can LLMs Effectively Leverage Graph Structural Information: When and Why , 2023, ArXiv.

[37]  P. Xie,et al.  Graph Neural Architecture Search Under Distribution Shifts , 2022, ICML.