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/

Zendy Plus

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

Zendy Tools

Zendy Open

Deep semantic matching aims to discriminate the relationship betweendocuments based on deep neural networks. In recent years, it becomesincreasingly popular to organize documents with a graph structure, thenleverage both the intrinsic document features and the extrinsic neighborfeatures to derive discrimination. Most of the existing works mainly care abouthow to utilize the presented neighbors, whereas limited effort is made tofilter appropriate neighbors. We argue that the neighbor features could behighly noisy and partially useful. Thus, a lack of effective neighbor selectionwill not only incur a great deal of unnecessary computation cost, but alsorestrict the matching accuracy severely.  In this work, we propose a novel framework, Cascaded Deep Semantic Matching(CDSM), for accurate and efficient semantic matching on textual graphs. CDSM ishighlighted for its two-stage workflow. In the first stage, a lightweightCNN-based ad-hod neighbor selector is deployed to filter useful neighbors forthe matching task with a small computation cost. We design both one-step andmulti-step selection methods. In the second stage, a high-capacity graph-basedmatching network is employed to compute fine-grained relevance scores based onthe well-selected neighbors. It is worth noting that CDSM is a genericframework which accommodates most of the mainstream graph-based semanticmatching networks. The major challenge is how the selector can learn todiscriminate the neighbors usefulness which has no explicit labels. To copewith this problem, we design a weak-supervision strategy for optimization,where we train the graph-based matching network at first and then the ad-hocneighbor selector is learned on top of the annotations from the matchingnetwork.

CDSM: Cascaded Deep Semantic Matching on Textual Graphs Leveraging Ad-hoc Neighbor Selection