Schedule as of Oct 11, 2022 - subject to change

Default Time Zone is EDT - Eastern Daylight Time

Back To Schedule
Thursday, October 20 • 4:00pm - 4:20pm
Improving Domain Generalization Via Event-Based Acoustic Scene Classification

Sign up or log in to save this to your schedule, view media, leave feedback and see who's attending!

Acoustic Scene Classification (ASC) has been typically addressed by feeding raw audio features to deep neural networks. However, such an audio-based approach has consistently proved to result in poor model generalization across different recording devices. In fact, device-specific transfer functions and nonlinear dynamic range compression highly affect spectro-temporal features, resulting in a deviation from the learned data distribution known as domain shift. In this paper, we present an alternative ASC paradigm that involves ditching the classic end-to-end audio-based training in favor of gathering an intermediate event-based representation of the acoustic scenes using large-scale pretrained models. Performance evaluation on the TAU Urban Acoustic Scenes 2020 Mobile Development dataset shows that the proposed event-based approach is up to 160% more robust than corresponding audio-based methods in the face of mismatched recording devices.


Thursday October 20, 2022 4:00pm - 4:20pm EDT