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 learning models usually require a large amount of labeled data toachieve satisfactory performance. In multimedia analysis, domain adaptationstudies the problem of cross-domain knowledge transfer from a label rich sourcedomain to a label scarce target domain, thus potentially alleviates theannotation requirement for deep learning models. However, we find thatcontemporary domain adaptation methods for cross-domain image understandingperform poorly when source domain is noisy. Weakly Supervised Domain Adaptation(WSDA) studies the domain adaptation problem under the scenario where sourcedata can be noisy. Prior methods on WSDA remove noisy source data and align themarginal distribution across domains without considering the fine-grainedsemantic structure in the embedding space, which have the problem of classmisalignment, e.g., features of cats in the target domain might be mapped nearfeatures of dogs in the source domain. In this paper, we propose a novelmethod, termed Noise Tolerant Domain Adaptation, for WSDA. Specifically, weadopt the cluster assumption and learn cluster discriminatively with classprototypes in the embedding space. We propose to leverage the locationinformation of the data points in the embedding space and model the locationinformation with a Gaussian mixture model to identify noisy source data. Wethen design a network which incorporates the Gaussian mixture noise model as asub-module for unsupervised noise removal and propose a novel cluster-leveladversarial adaptation method which aligns unlabeled target data with the lessnoisy class prototypes for mapping the semantic structure across domains. Weconduct extensive experiments to evaluate the effectiveness of our method onboth general images and medical images from COVID-19 and e-commerce datasets.The results show that our method significantly outperforms state-of-the-artWSDA methods.

Towards Robust Cross-domain Image Understanding with Unsupervised Noise Removal