<ul><li>Researchers have developed low-complexity networks for sound event detection.</li><li>The goal is to achieve competitive performance with state-of-the-art models while reducing computational requirements.</li><li>Through adjustments to convolutional models, such as changing strides and adding sequence models, the performance of the low-complexity models was improved.</li><li>By optimizing training strategies, event detection performance comparable to state-of-the-art transformers was achieved with only 5% of the parameters.</li></ul>

Exploring Performance-Complexity Trade-Offs in Sound Event Detection

Discover more