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

Question Answering (QA), a popular and promising technique for intelligentinformation access, faces a dilemma about data as most other AI techniques. Onone hand, modern QA methods rely on deep learning models which are typicallydata-hungry. Therefore, it is expected to collect and fuse all the available QAdatasets together in a common site for developing a powerful QA model. On theother hand, real-world QA datasets are typically distributed in the form ofisolated islands belonging to different parties. Due to the increasingawareness of privacy security, it is almost impossible to integrate the datascattered around, or the cost is prohibited. A possible solution to thisdilemma is a new approach known as federated learning, which is aprivacy-preserving machine learning technique over distributed datasets. Inthis work, we propose to adopt federated learning for QA with the specialconcern on the statistical heterogeneity of the QA data. Here the heterogeneityrefers to the fact that annotated QA data are typically with non-identical andindependent distribution (non-IID) and unbalanced sizes in practice.Traditional federated learning methods may sacrifice the accuracy of individualmodels under the heterogeneous situation. To tackle this problem, we propose anovel Federated Matching framework for QA, named FedMatch, with abackbone-patch architecture. The shared backbone is to distill the commonknowledge of all the participants while the private patch is a compact andefficient module to retain the domain information for each participant. Tofacilitate the evaluation, we build a benchmark collection based on several QAdatasets from different domains to simulate the heterogeneous situation inpractice. Empirical studies demonstrate that our model can achieve significantimprovements against the baselines over all the datasets.

FedMatch: Federated Learning Over Heterogeneous Question Answering Data