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

Data poisoning attacks aim at manipulating model behaviors through distortingtraining data. Previously, an aggregation-based certified defense, DeepPartition Aggregation (DPA), was proposed to mitigate this threat. DPA predictsthrough an aggregation of base classifiers trained on disjoint subsets of data,thus restricting its sensitivity to dataset distortions. In this work, wepropose an improved certified defense against general poisoning attacks, namelyFinite Aggregation. In contrast to DPA, which directly splits the training setinto disjoint subsets, our method first splits the training set into smallerdisjoint subsets and then combines duplicates of them to build larger (but notdisjoint) subsets for training base classifiers. This reduces the worst-caseimpacts of poison samples and thus improves certified robustness bounds. Inaddition, we offer an alternative view of our method, bridging the designs ofdeterministic and stochastic aggregation-based certified defenses. Empirically,our proposed Finite Aggregation consistently improves certificates on MNIST,CIFAR-10, and GTSRB, boosting certified fractions by up to 3.05%, 3.87% and4.77%, respectively, while keeping the same clean accuracies as DPA's,effectively establishing a new state of the art in (pointwise) certifiedrobustness against data poisoning.

Improved Certified Defenses against Data Poisoning with (Deterministic) Finite Aggregation