Study of Positional Encoding Approaches for Audio Spectrogram Transformers

LEONARDO PEPINO; PABLO RIERA; LUCIANA FERRER

Congreso; ICASSP 2022; 2022

IEEE

Transformers have revolutionized the world of deep learning, specially in the field of natural language processing. Recently, the Audio Spectrogram Transformer (AST) was proposed for audio classification, leading to state of the art results in several datasets. However,  in order for ASTs to outperform  CNNs, pretraining with ImageNet is needed. In this paper, we study one component of the AST,  the positional encoding, and propose several variants  to improve the performance of ASTs trained from scratch, without ImageNet pretraining. Our best model, which incorporates conditional positional encodings, significantly improves performance on Audioset and ESC-50 compared to the original AST.