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

Training a deep neural model for semantic segmentation requires collecting alarge amount of pixel-level labeled data. To alleviate the data scarcityproblem presented in the real world, one could utilize synthetic data whoselabel is easy to obtain. Previous work has shown that the performance of asemantic segmentation model can be improved by training jointly with real andsynthetic examples with a proper weighting on the synthetic data. Suchweighting was learned by a heuristic to maximize the similarity betweensynthetic and real examples. In our work, we instead learn a pixel-levelweighting of the synthetic data by meta-learning, i.e., the learning ofweighting should only be minimizing the loss on the target task. We achievethis by gradient-on-gradient technique to propagate the target loss back intothe parameters of the weighting model. The experiments show that our methodwith only one single meta module can outperform a complicated combination of anadversarial feature alignment, a reconstruction loss, plus a hierarchicalheuristic weighting at pixel, region and image levels.

MetaPix: Domain Transfer for Semantic Segmentation by Meta Pixel Weighting