English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Contrastive Learning (CL) has emerged as a dominant technique forunsupervised representation learning which embeds augmented versions of theanchor close to each other (positive samples) and pushes the embeddings ofother samples (negatives) apart. As revealed in recent studies, CL can benefitfrom hard negatives (negatives that are most similar to the anchor). However,we observe limited benefits when we adopt existing hard negative miningtechniques of other domains in Graph Contrastive Learning (GCL). We performboth experimental and theoretical analysis on this phenomenon and find it canbe attributed to the message passing of Graph Neural Networks (GNNs). Unlike CLin other domains, most hard negatives are potentially false negatives(negatives that share the same class with the anchor) if they are selectedmerely according to the similarities between anchor and themselves, which willundesirably push away the samples of the same class. To remedy this deficiency,we propose an effective method, dubbed \textbf{ProGCL}, to estimate theprobability of a negative being true one, which constitutes a more suitablemeasure for negatives' hardness together with similarity. Additionally, wedevise two schemes (i.e., \textbf{ProGCL-weight} and \textbf{ProGCL-mix}) toboost the performance of GCL. Extensive experiments demonstrate that ProGCLbrings notable and consistent improvements over base GCL methods and yieldsmultiple state-of-the-art results on several unsupervised benchmarks or evenexceeds the performance of supervised ones. Also, ProGCL is readily pluggableinto various negatives-based GCL methods for performance improvement. Werelease the code at\textcolor{magenta}{\url{https://github.com/junxia97/ProGCL}}.

ProGCL: Rethinking Hard Negative Mining in Graph Contrastive Learning