Meaningfully Explaining Model Mistakes Using Conceptual Counterfactuals

1 code implementation • 24 Jun 2021 • Abubakar Abid, Mert Yuksekgonul, James Zou

In this paper, we propose a systematic approach, conceptual counterfactual explanations(CCE), that explains why a classifier makes a mistake on a particular test sample(s) in terms of human-understandable concepts (e. g. this zebra is misclassified as a dog because of faint stripes).

ImageNet performance correlates with pose estimation robustness and generalization on out-of-domain data

1 code implementation • ICML UDL 2020 • Alexander Mathis, Thomas Biasi, Mert Yuksekgonul, Byron Rogers, Matthias Bethge, Mackenzie Weygandt Mathis

Neural networks are highly effective tools for pose estimation.

Animal Pose Estimation Transfer Learning

Learning Maximally Predictive Prototypes in Multiple Instance Learning

1 code implementation • 2 Oct 2019 • Mert Yuksekgonul, Ozgur Emre Sivrikaya, Mustafa Gokce Baydogan

In this work, we propose a simple model that provides permutation invariant maximally predictive prototype generator from a given dataset, which leads to interpretability of the solution and concrete insights to the nature and the solution of a problem.

Multiple Instance Learning