XAI-SA 2024 : Explainable Machine Learning for Speech and Audio

We invite paper contributions for our workshop on Explainable AI for Speech and Audio which will be co-located with ICASSP 2024. We invite applied and methodological papers on explainable models, explanation methods, visualization methods, or any technique that would increase the interpretability of black-box AI systems that are deployed on speech and audio problems. We will also welcome paper submissions from domains other than speech and audio provided that the submission contains interesting contributions on interpretability. More specific topics for paper contributions include (but are not limited to),


- Methodological Contributions to improve Posthoc Interpretation Methods for Speech and Audio

- Evaluation Methods for Explanations of Audio Classifiers

- Glassbox Models for Speech and Audio

- Explaining Self-Supervised Models

- Explanation/Interpretability Methods for Multi-Modal Methods

- Explanation/Interpretability Methods for Time Series Models

- Applications of Explanation Methods and Glassbox Models


The paper format is the same as ICASSP, and the review format is single-blind. We will have two tracks for paper submissions.


IEEEXplore track: This track will accept submissions that are novel contributions, thus, that have yet to be published in a conference with proceedings or a journal. These contributions will appear on IEEEXplore. The deadline for submissions is January 20, 2024. Submission portal.


Workshop track: This track will accept submissions of work that are works-in progress, or papers which have already appeared in other venues with proceedings / journals. The deadline for submissions is February 20, 2024. Submission portal.

Organizers:
Cem Subakan, Laval University/Mila
Francesco Paissan, University of Trento/FBK
Mirco Ravanelli, Concordia University/Mila
Shubham Gupta, Laval University/Mila
Pascal Germain, Laval University/Mila
Paris Smaragdis, University of Illinois at Urbana-Champaign