About cookies on this site Our websites require some cookies to function properly (required). In addition, other cookies may be used with your consent to analyze site usage, improve the user experience and for advertising. For more information, please review your options. By visiting our website, you agree to our processing of information as described in IBM’sprivacy statement. To provide a smooth navigation, your cookie preferences will be shared across the IBM web domains listed here.
Conference paper
Few-shot audio classification with attentional graph neural networks
Abstract
Few-shot learning is a very promising and challenging field of machine learning as it aims to understand new concepts from very few labeled examples. In this paper, we propose attentional framework to extend recently proposed few-shot learning with graph neural network [1] in audio classification scenario. The objective of proposed attentional framework is to introduce a flexible framework to implement selectively concentration procedure on support examples for each query process. we also present an empirical study on confidence measure for few-shot learning application by combining posterior probability with normalized entropy of the network's probability output. The efficiency of the proposed method is demonstrated with experiments on balanced training set of Audio set for training and a 5-way test set composed of about 5-hour audio data for testing.