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

Unsupervised representation learning aims at finding methods that learnrepresentations from data without annotation-based signals. Abstaining fromannotations not only leads to economic benefits but may - and to some extentalready does - result in advantages regarding the representation's structure,robustness, and generalizability to different tasks. In the long run,unsupervised methods are expected to surpass their supervised counterparts dueto the reduction of human intervention and the inherently more general setupthat does not bias the optimization towards an objective originating fromspecific annotation-based signals. While major advantages of unsupervisedrepresentation learning have been recently observed in natural languageprocessing, supervised methods still dominate in vision domains for most tasks.In this dissertation, we contribute to the field of unsupervised (visual)representation learning from three perspectives: (i) Learning representations:We design unsupervised, backpropagation-free Convolutional Self-OrganizingNeural Networks (CSNNs) that utilize self-organization- and Hebbian-basedlearning rules to learn convolutional kernels and masks to achieve deeperbackpropagation-free models. (ii) Evaluating representations: We build upon thewidely used (non-)linear evaluation protocol to define pretext- andtarget-objective-independent metrics for measuring and investigating theobjective function mismatch between various unsupervised pretext tasks andtarget tasks. (iii) Transferring representations: We contribute CARLANE, thefirst 3-way sim-to-real domain adaptation benchmark for 2D lane detection, anda method based on prototypical self-supervised learning. Finally, we contributea content-consistent unpaired image-to-image translation method that utilizesmasks, global and local discriminators, and similarity sampling to mitigatecontent inconsistencies.

Towards Unsupervised Representation Learning: Learning, Evaluating and Transferring Visual Representations