Recognizing Sound Events - 2019

Explore the cutting-edge developments in automatic sound classification through this comprehensive lecture by Dan Ellis from Google's Sound Understanding team. Delve into the application of vision-inspired deep neural networks for classifying the 'AudioSet' ontology of approximately 600 sound events, encompassing speech, music, and environmental sounds. Learn about related applications in bioacoustics and cross-modal learning, and discover insights from a recent Kaggle competition run in collaboration with UPF Barcelona. Gain knowledge about the upcoming release of a pretrained model aimed at making state-of-the-art generic sound recognition widely accessible. This talk, presented at the Center for Language & Speech Processing (CLSP) at JHU, offers valuable insights into the latest advancements in sound event recognition and its practical applications.