We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations

DBLP - CS Bibliography


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Sound

Title: CAT: Causal Audio Transformer for Audio Classification

Abstract: The attention-based Transformers have been increasingly applied to audio classification because of their global receptive field and ability to handle long-term dependency. However, the existing frameworks which are mainly extended from the Vision Transformers are not perfectly compatible with audio signals. In this paper, we introduce a Causal Audio Transformer (CAT) consisting of a Multi-Resolution Multi-Feature (MRMF) feature extraction with an acoustic attention block for more optimized audio modeling. In addition, we propose a causal module that alleviates over-fitting, helps with knowledge transfer, and improves interpretability. CAT obtains higher or comparable state-of-the-art classification performance on ESC50, AudioSet and UrbanSound8K datasets, and can be easily generalized to other Transformer-based models.
Comments: Accepted to ICASSP 2023
Subjects: Sound (cs.SD); Multimedia (cs.MM); Audio and Speech Processing (eess.AS)
Cite as: arXiv:2303.07626 [cs.SD]
  (or arXiv:2303.07626v1 [cs.SD] for this version)

Submission history

From: Xiaoyu Liu [view email]
[v1] Tue, 14 Mar 2023 04:50:52 GMT (1180kb,D)

Link back to: arXiv, form interface, contact.