Evaluating Balanced Domain Regularizations for Multi-Domain Learning in Audio Classification Tasks

Abstract

Collections of data obtained or generated under similar conditions are called do mains or data sources. The distinct data acquisition or generation conditions are often neglected, but understanding them is vital to address any phenomena emerging from these differences that might hinder model generalization. Multi-domain learning seeks the best way to train a model to perform adequately in all domains used during train ing. This work explores multi-domain learning techniques that use explicit information about the domain of an example in addition to its class. This study evaluates a general approach (Stew) by mixing all available data and also two batch domain-regularization methods: Balanced Domains and Loss Sum. We train machine learning models with the listed approaches using datasets with multiple data sources for audio classification tasks. The results suggest that training a model using the Loss Sum method improves the performance of models otherwise trained in a mix of all available data.