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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Deep learning architectures have shown remarkable results in sceneunderstanding problems, however they exhibit a critical drop of performanceswhen they are required to learn incrementally new tasks without forgetting oldones. This catastrophic forgetting phenomenon impacts on the deployment ofartificial intelligence in real world scenarios where systems need to learn newand different representations over time. Current approaches for incrementallearning deal only with image classification and object detection tasks, whilein this work we formally introduce incremental learning for semanticsegmentation. We tackle the problem applying various knowledge distillationtechniques on the previous model. In this way, we retain the information aboutlearned classes, whilst updating the current model to learn the new ones. Wedeveloped four main methodologies of knowledge distillation working on bothoutput layers and internal feature representations. We do not store any imagebelonging to previous training stages and only the last model is used topreserve high accuracy on previously learned classes. Extensive experimentalresults on the Pascal VOC2012 and MSRC-v2 datasets show the effectiveness ofthe proposed approaches in several incremental learning scenarios.

Knowledge Distillation for Incremental Learning in Semantic Segmentation