The program will be available shortly. Please check back later.
ExMap: Leveraging Explainability Heatmaps for Unsupervised Group Robustness to Spurious Correlations
Presenter:
Rwiddhi Chakraborty, Doctoral Research Fellow at the Machine Learning Group, UiT The Arctic University of Norway
Abstract:
Group robustness strategies aim to mitigate learned biases in deep learning models that arise from spurious correlations present in their training datasets. However, most existing methods rely on the access to the label distribution of the groups, which is time-consuming and expensive to obtain. As a result, unsupervised group robustness strategies are sought. Based on the insight that a trained model's classification strategies can be inferred accurately based on explainability heatmaps, we introduce ExMap, an unsupervised two stage mechanism designed to enhance group robustness intraditional classifiers. ExMap utilizes a clustering module to inferpseudo-labels based on a model's explainability heat maps, which are then used during training in lieu of actual labels. Our empirical studies validate the efficacy of ExMap - We demonstrate that it bridges the performance gap with its supervised counterparts and outperforms existing partially supervised and unsupervised methods. Additionally, ExMap can be seamlessly integrated with existing group robustness learning strategies. Finally, we demonstrate its potential in tackling the emerging issue of multiple shortcut mitigation.
In compliance with GDPR consent requirements, presentations given in a Visual Intelligence context may be recorded with the consent of the speaker. All recordings are edited to remove all faces, names and voices of other participants. Questions and comments by the audience will hence be removed and will not appear in the recording. With the freely given consent from the speaker, recorded presentation may be posted on the Visual Intelligence YouTube channel.