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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Given two large lists of records, the task in entity resolution (ER) is tofind the pairs from the Cartesian product of the lists that correspond to thesame real world entity. Typically, passive learning methods on such tasksrequire large amounts of labeled data to yield useful models. Active Learningis a promising approach for ER in low resource settings. However, the searchspace, to find informative samples for the user to label, grows quadraticallyfor instance-pair tasks making active learning hard to scale. Previous works,in this setting, rely on hand-crafted predicates, pre-trained language modelembeddings, or rule learning to prune away unlikely pairs from the Cartesianproduct. This blocking step can miss out on important regions in the productspace leading to low recall. We propose DIAL, a scalable active learningapproach that jointly learns embeddings to maximize recall for blocking andaccuracy for matching blocked pairs. DIAL uses an Index-By-Committee framework,where each committee member learns representations based on powerfulpre-trained transformer language models. We highlight surprising differencesbetween the matcher and the blocker in the creation of the training data andthe objective used to train their parameters. Experiments on five benchmarkdatasets and a multilingual record matching dataset show the effectiveness ofour approach in terms of precision, recall and running time. Code is availableat https://github.com/ArjitJ/DIAL

Deep Indexed Active Learning for Matching Heterogeneous Entity Representations