Transfer Learning of Graph Neural Networks with Ego-graph Information Maximization
Dez 6, 2021
Sprecher:innen
Haonan Wang
Speaker · 0 followers
Chao Zhang
Speaker · 1 follower
Jiawei Han
Speaker · 1 follower
Über
Graph neural networks (GNNs) have achieved superior performance in various applications, but training dedicated GNNs can be costly for large-scale graphs. Some recent work started to study the pre-training of GNNs. However, none of them provide theoretical insights into the design of their frameworks, or clear requirements and guarantees towards their transferability. In this work, we establish a theoretically grounded and practically useful framework for the transfer learning of GNNs. Firstly, we propose a novel view towards the essential graph information and advocate the capturing of it as the goal of transferable GNN training, which motivates the design of EGI (Ego-Graph Information maximization) to analytically achieve this goal. Secondly,when node features are structure-relevant, we conduct an analysis of EGI transferability regarding the difference between the local graph Laplacians of the source and target graphs. We conduct controlled synthetic experiments to directly justify our theoretical conclusions. Comprehensive experiments on two real-world network datasets show consistent results in the analyzed setting of direct-transfering, while those on large-scale knowledge graphs show promising results in the more practical setting of transfering with fine-tuning.Graph neural networks (GNNs) have achieved superior performance in various applications, but training dedicated GNNs can be costly for large-scale graphs. Some recent work started to study the pre-training of GNNs. However, none of them provide theoretical insights into the design of their frameworks, or clear requirements and guarantees towards their transferability. In this work, we establish a theoretically grounded and practically useful framework for the transfer learning of GNNs. Firstly,…
Organisator
NeurIPS 2021
Account · 1.9k followers
Über NeurIPS 2021
Neural Information Processing Systems (NeurIPS) is a multi-track machine learning and computational neuroscience conference that includes invited talks, demonstrations, symposia and oral and poster presentations of refereed papers. Following the conference, there are workshops which provide a less formal setting.
Gefällt euch das Format? Vertraut auf SlidesLive, um euer nächstes Event festzuhalten!
Professionelle Aufzeichnung und Livestreaming – weltweit.
Freigeben
Empfohlene Videos
Präsentationen, deren Thema, Kategorie oder Sprecher:in ähnlich sind
Data vast and low in variance: Augment machine learning pipelines with dataset profiles to improve data quality without sacrificing scale
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Deep Molecular Representation Learning via Fusing Physical and Chemical Information
Watch later
Shuwen Yang, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Differentially Private n-gram Extraction
Watch later
Kunho Kim, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Modeling the human brain from invariance and robustness to clutter towards multimodal, multi-task and continuous learning models
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Fault-Tolerant Federated Reinforcement Learning with Theoretical Guarantee
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Compositional Transformers for Scene Generation
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%