2020/09/11 by Huy Phan, Phan, Huy, Lam Pham +9 · 1 citation
Computer Science · #Audio and Speech Processing (eess.AS) #Biology #Computer science #Event (particle physics) #FOS: Computer and information sciences #FOS: Electrical engineering #Function (biology) #Machine Learning (cs.LG) #Music and Audio Processing #Physics #Speech Recognition and Synthesis #Speech and Audio Processing #Speech recognition #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2009.05527
openalex publication_date 2020/09/11 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Audio event localization and detection (SELD) have been commonly tackled using multitask models. Such a model usually consists of a multi-label event classification branch with sigmoid cross-entropy loss for event activity detection and a regression branch with mean squared error loss for direction-of-arrival estimation. In this work, we propose a multitask regression model, in which both (multi-label) event detection and localization are formulated as regression problems and use the mean squared error loss homogeneously for model training. We show that the common combination of heterogeneous loss functions causes the network to underfit the data whereas the homogeneous mean squared error loss leads to better convergence and performance. Experiments on the development and validation sets of the DCASE 2020 SELD task demonstrate that the proposed system also outperforms the DCASE 2020 SELD baseline across all the detection and localization metrics, reducing the overall SELD error (the combined metric) by approximately 10% absolute.