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

We address the problem of ensemble selection in transfer learning: Given alarge pool of source models we want to select an ensemble of models which,after fine-tuning on the target training set, yields the best performance onthe target test set. Since fine-tuning all possible ensembles iscomputationally prohibitive, we aim at predicting performance on the targetdataset using a computationally efficient transferability metric. We proposeseveral new transferability metrics designed for this task and evaluate them ina challenging and realistic transfer learning setup for semantic segmentation:we create a large and diverse pool of source models by considering 17 sourcedatasets covering a wide variety of image domain, two different architectures,and two pre-training schemes. Given this pool, we then automatically select asubset to form an ensemble performing well on a given target dataset. Wecompare the ensemble selected by our method to two baselines which select asingle source model, either (1) from the same pool as our method; or (2) from apool containing large source models, each with similar capacity as an ensemble.Averaged over 17 target datasets, we outperform these baselines by 6.0% and2.5% relative mean IoU, respectively.

Transferability Metrics for Selecting Source Model Ensembles