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

Multitask learning (MTL) has gained prominence for its ability to jointlypredict multiple tasks, achieving better per-task performance while using fewerper-task model parameters than single-task learning. More recently,decoder-focused architectures have considerably improved multitask performanceby refining task predictions using the features of other related tasks.However, most of these refinement methods fail to simultaneously capture localand global task-specific representations, as well as cross-task patterns in aparameter-efficient manner. In this paper, we introduce the Efficient MultitaskAffinity Learning Network (EMA-Net), which is a lightweight framework thatenhances the task refinement capabilities of multitask networks. EMA-Netadeptly captures local, global, and cross-task interactions using our novelCross-Task Affinity Learning (CTAL) module. The key innovation of CTAL lies inits ability to manipulate task affinity matrices in a manner that is optimallysuited to apply parameter-efficient grouped convolutions without worrying aboutinformation loss. Our results show that we achieve state-of-the-art MTLperformance for CNN-based decoder-focused models while using substantiallyfewer model parameters. Our code is publicly available athttps://github.com/Armanfard-Lab/EMA-Net.

EMA-Net: Efficient Multitask Affinity Learning for Dense Scene Predictions