Pons Puig, JordiSerrà Julià, JoanSerra, Xavier2019-05-222019Pons J, Serrà J, Serra X. Training neural audio classifiers with few data. In: 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP); 2019 May 12-17; Brighton, United Kingdom. New Jersey: Institute of Electrical and Electronics Engineers; 2019. p. 16-20. DOI: 10.1109/ICASSP.2019.86825912379-190Xhttp://hdl.handle.net/10230/37259Comunicació presentada a: ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing celebrat de 12 al 17 de maig de 2019 a Brighton, Regne Unit.We investigate supervised learning strategies that improve the training of neural network audio classifiers on small annotated collections. In particular, we study whether (i) a naive regularization of the solution space, (ii) prototypical networks, (iii) transfer learning, or (iv) their combination, can foster deep learning models to better leverage a small amount of training examples. To this end, we evaluate (i–iv) for the tasks of acoustic event recognition and acoustic scene classification, considering from 1 to 100 labeled examples per class. Results indicate that transfer learning is a powerful strategy in such scenarios, but prototypical networks show promising results when one does not count with external or validation data.application/pdfeng© 2019 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. http://dx.doi.org/10.1109/ICASSP.2019.8682591info:eu-repo/semantics/conferenceObjecthttp://dx.doi.org/10.1109/ICASSP.2019.8682591Prototypical networksTransfer learningAudio classificationSmall datainfo:eu-repo/semantics/openAccess