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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

The black-box nature of deep learning models prevents them from beingcompletely trusted in domains like biomedicine. Most explainability techniquesdo not capture the concept-based reasoning that human beings follow. In thiswork, we attempt to understand the behavior of trained models that performimage processing tasks in the medical domain by building a graphicalrepresentation of the concepts they learn. Extracting such a graphicalrepresentation of the model's behavior on an abstract, higher conceptual levelwould unravel the learnings of these models and would help us to evaluate thesteps taken by the model for predictions. We show the application of ourproposed implementation on two biomedical problems - brain tumor segmentationand fundus image classification. We provide an alternative graphicalrepresentation of the model by formulating a concept level graph as discussedabove, which makes the problem of intervention to find active inference trailsmore tractable. Understanding these trails would provide an understanding ofthe hierarchy of the decision-making process followed by the model. [As well asoverall nature of model]. Our framework is available athttps://github.com/koriavinash1/BioExp

Abstracting Deep Neural Networks into Concept Graphs for Concept Level Interpretability