Graphs are ubiquitous due to their flexibility in representing social and technological systems as networks of interacting elements. Graph representation learning methods, such as node embeddings, are powerful approaches to map nodes into a latent vector space, allowing their use for various graph tasks. Despite their success, only few studies have focused on explaining node embeddings locally. Moreover, global explanations of node embeddings remain unexplored, limiting interpretability and debugging potentials. We address this gap by developing human-understandable explanations for dimensions in node embeddings. Towards that, we first develop new metrics that measure the global interpretability of embedding vectors based on the marginal contribution of the embedding dimensions to predicting graph structure. We say that an embedding dimension is more interpretable if it can faithfully map to an understandable sub-structure in the input graph - like community structure. Having observed that standard node embeddings have low interpretability, we then introduce DINE (Dimension-based Interpretable Node Embedding), a novel approach that can retrofit existing node embeddings by making them more interpretable without sacrificing their task performance. We conduct extensive experiments on synthetic and real-world graphs and show that we can simultaneously learn highly interpretable node embeddings with effective performance in link prediction.
From GNNs to Trees: Multi-Granular Interpretability for Graph Neural Networks
Zhejiang University(浙江大学);State Key Laboratory of Blockchain and Data Security(区块链与数据安全国家重点实验室);Hangzhou High-Tech Zone (Binjiang) Institute of Blockchain and Data Security(杭州高新技术区(滨江区)区块链与数据安全研究院);Big Graph Center, Hangzhou City University(杭州城市大学大数据中心);Nanyang Technological University(南洋理工大学)
Department of Computer, Control and Management Engineering, Sapienza University of Rome(西帕辛扎大学罗马分校计算机、控制与管理工程系);Department of Life Science Informatics and Data Science, B-IT, LIMES Program Unit Chemical Biology and Medicinal Chemistry, Rheinische Friedrich-Wilhelms-Universität, Bonn, Germany(波恩德国莱茵-威斯巴登大学生命科学信息技术与数据科学系);Lamarr Institute for Machine Learning and Artificial Intelligence, Bonn, Germany(波恩德国拉马尔机器学习与人工智能研究所)
GnnXemplar: Exemplars to Explanations -- Natural Language Rules for Global GNN Interpretability
Dept. of CSE, IIT Delhi(印度德里理工学院计算机科学与工程系);Fujitsu Research of India, Bangalore(印度班加罗尔富士通印度研究机构);Dept. of CSE and Yardi ScAI, IIT Delhi(印度德里理工学院计算机科学与工程系)
Improving the interpretability of GNN predictions through conformal-based graph sparsification
Max Planck Institute for Intelligent Systems(马克斯·普朗克智能系统研究所);McGill University(麦吉尔大学);Saarland University(萨尔大学);Max Planck Institute for Software Systems(马克斯·普朗克软件系统研究所)
Learnable Structural Semantic Readout for Graph Classification
University of Illinois at Urbana-Champaign (UIUC)(伊利诺伊大学厄巴纳-香槟分校 (UIUC));Pohang University of Science and Technology (POSTECH)(浦项科技大学 (POSTECH));Korea Advanced Institute of Science and Technology (KAIST)(韩国科学技术院 (KAIST))
University of Michigan(密歇根大学);Alibaba Hangzhou(阿里巴巴(杭州));University of California, San Diego(加州大学圣地亚哥分校);Fudan University(复旦大学);Amazon Web Services(亚马逊网络服务)
The Snowflake Hypothesis: Training Deep GNN with One Node One Receptive field
University of Science and Technology of China (USTC)(中国科学技术大学);King Abdullah University of Science and Technology (KAUST)(阿卜杜拉国王科技大学);Tongji University(同济大学);National University of Singapore(新加坡国立大学);Shenzhen Technology University(深圳技术大学);The Hong Kong University of Science and Technology (Guangzhou)(香港科技大学(广州))
Heterogeneous Graph Alignment for Joint Reasoning and Interpretability
Department of Statistics, University of California, Irvine(统计学系,加州大学伊维特分校);Department of Neurobiology and Behavior, University of California, Irvine(神经生物学与行为系,加州大学伊维特分校)
Mixture of Message Passing Experts with Routing Entropy Regularization for Node Classification
Institute of Cyberspace Security, Zhejiang University of Technology(浙江工业大学网络安全学院);Binjiang Cyberspace Security Institute of ZJUT(浙江工业大学滨江区网络安全研究所);Faculty of Artificial Intelligence in Education, Central China Normal University(中央财经大学教育人工智能学院)
School of Artificial Intelligence, South China Normal University(人工智能学院,华南师范大学);State Key Laboratory of Mathematical Sciences, Academy of Mathematics and Systems Science, Chinese Academy of Sciences(数学科学国家重点实验室,数学与系统科学研究院,中国科学院)
GraphAU-Pain: Graph-based Action Unit Representation for Pain Intensity Estimation
Department of Computer Science, University of Cambridge, Cambridge, UK(剑桥大学计算机科学系);Center for Machine Vision and Signal Analysis, University of Oulu, Oulu, Finland(奥卢大学机器视觉与信号分析中心)
SurfGNN: A robust surface-based prediction model with interpretability for coactivation maps of spatial and cortical features
Sun Yat-sen University(中山大学);Ningbo Institute of Materials Technology and Engineering, Chinese Academy of Sciences(中国科学院宁波材料技术与工程研究所);Ningbo University of Technology(宁波工程学院);University of California San Francisco(加利福尼亚大学旧金山分校);University of Southern California(南加州大学);Shenzhen Maternity and Child Healthcare Hospital(深圳市妇幼保健院)
Mechanistic Interpretability for Neural TSP Solvers
University of Washington, Michael G. Foster School of Business(华盛顿大学,迈克尔·G·福斯特商学院);University of Washington, Paul G. Allen School of Computer Science & Engineering(华盛顿大学,保罗·G·阿伦计算机科学与工程学院)
Optimization and Interpretability of Graph Attention Networks for Small Sparse Graph Structures in Automotive Applications
CARISSMA Institute of Automated Driving, Technische Hochschule Ingolstadt(英戈尔施塔特应用技术大学CARISSMA自动驾驶研究所);AUDI AG(奥迪股份公司);Technical University of Munich(慕尼黑工业大学)