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

A massive number of well-trained deep networks have been released bydevelopers online. These networks may focus on different tasks and in manycases are optimized for different datasets. In this paper, we study how toexploit such heterogeneous pre-trained networks, known as teachers, so as totrain a customized student network that tackles a set of selective tasksdefined by the user. We assume no human annotations are available, and eachteacher may be either single- or multi-task. To this end, we introduce adual-step strategy that first extracts the task-specific knowledge from theheterogeneous teachers sharing the same sub-task, and then amalgamates theextracted knowledge to build the student network. To facilitate the training,we employ a selective learning scheme where, for each unlabelled sample, thestudent learns adaptively from only the teacher with the least predictionambiguity. We evaluate the proposed approach on several datasets andexperimental results demonstrate that the student, learned by such adaptiveknowledge amalgamation, achieves performances even better than those of theteachers.

Customizing Student Networks From Heterogeneous Teachers via Adaptive Knowledge Amalgamation