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

Humans use abstract concepts for understanding instead of hard features.Recent interpretability research has focused on human-centered conceptexplanations of neural networks. Concept Activation Vectors (CAVs) estimate amodel's sensitivity and possible biases to a given concept. In this paper, weextend CAVs from post-hoc analysis to ante-hoc training in order to reducemodel bias through fine-tuning using an additional Concept Loss. Concepts weredefined on the final layer of the network in the past. We generalize it tointermediate layers using class prototypes. This facilitates class learning inthe last convolution layer, which is known to be most informative. We alsointroduce Concept Distillation to create richer concepts using a pre-trainedknowledgeable model as the teacher. Our method can sensitize or desensitize amodel towards concepts. We show applications of concept-sensitive training todebias several classification problems. We also use concepts to induce priorknowledge into IID, a reconstruction problem. Concept-sensitive training canimprove model interpretability, reduce biases, and induce prior knowledge.Please visit https://avani17101.github.io/Concept-Distilllation/ for code andmore details.

Concept Distillation: Leveraging Human-Centered Explanations for Model Improvement