Strategies for Pre-training Graph Neural Networks

ICLR 2020 Weihua HuBowen LiuJoseph GomesMarinka ZitnikPercy LiangVijay PandeJure Leskovec

Many applications of machine learning require a model to make accurate pre-dictions on test examples that are distributionally different from training ones, while task-specific labels are scarce during training. An effective approach to this challenge is to pre-train a model on related tasks where data is abundant, and then fine-tune it on a downstream task of interest... (read more)

PDF Abstract

Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT LEADERBOARD
Drug Discovery BACE ContextPred AUC 0.845 # 1
Drug Discovery BBBP ContextPred AUC 0.687 # 1
Drug Discovery ClinTox ContextPred AUC 0.726 # 1
Drug Discovery HIV dataset ContextPred AUC 0.799 # 3
Drug Discovery MUV ContextPred AUC 0.813 # 3
Drug Discovery SIDER ContextPred AUC 0.627 # 1
Drug Discovery Tox21 ContextPred AUC 0.781 # 6
Drug Discovery ToxCast ContextPred AUC 0.657 # 3