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

Deep neural networks trained for predicting cellular events from DNA sequencehave become emerging tools to help elucidate the biological mechanismunderlying the associations identified in genome-wide association studies. Toenhance the training, multi-task learning (MTL) has been commonly exploited inprevious works where trained networks were needed for multiple profilesdiffering in either event modality or cell type. All existing works adopted asimple MTL framework where all tasks share a single feature extraction network.Such a strategy even though effective to certain extent leads to substantialnegative transfer, meaning the existence of large portion of tasks for whichmodels obtained through MTL perform worse than those by single task learning.There have been methods developed to address such negative transfer in otherdomains, such as computer vision. However, these methods are generallydifficult to scale up to handle large amount of tasks. In this paper, wepropose a highly scalable task grouping framework to address negative transferby only jointly training tasks that are potentially beneficial to each other.The proposed method exploits the network weights associated with task specificclassification heads that can be cheaply obtained by one-time joint training ofall tasks. Our results using a dataset consisting of 367 epigenetic profilesdemonstrate the effectiveness of the proposed approach and its superiority overbaseline methods.

Highly Scalable Task Grouping for Deep Multi-Task Learning in Prediction of Epigenetic Events