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

Numerous deep learning applications benefit from multi-task learning withmultiple regression and classification objectives. In this paper we make theobservation that the performance of such systems is strongly dependent on therelative weighting between each task's loss. Tuning these weights by hand is adifficult and expensive process, making multi-task learning prohibitive inpractice. We propose a principled approach to multi-task deep learning whichweighs multiple loss functions by considering the homoscedastic uncertainty ofeach task. This allows us to simultaneously learn various quantities withdifferent units or scales in both classification and regression settings. Wedemonstrate our model learning per-pixel depth regression, semantic andinstance segmentation from a monocular input image. Perhaps surprisingly, weshow our model can learn multi-task weightings and outperform separate modelstrained individually on each task.

Multi-Task Learning Using Uncertainty to Weigh Losses for Scene Geometry and Semantics