Search papers, labs, and topics across Lattice.
This paper introduces the Domain-Agnostic Incremental Learning for Audio Classification Task as part of the DCASE 2026 Challenge, focusing on the ability to learn sound classes across different acoustic domains without access to previous task data. The challenge emphasizes the need for systems to maintain performance on previously learned tasks while adapting to new domains, a significant hurdle in incremental learning. The baseline system achieves an accuracy of 44.9%, highlighting the difficulties in domain inference that impact overall performance.
Learning sound classes across varying acoustic domains without prior data access reveals a critical gap in current incremental learning approaches.
This paper presents the Domain-Agnostic Incremental Learning for Audio Classification Task of the DCASE 2026 Challenge. Incremental learning refers to sequentially learning new tasks with the same system while maintaining its knowledge and performance on the previously learned task. Domain-incremental learning for sound classification refers to learning the same sound classes but in different acoustic domains, and was formalized as a data challenge for the first time in DCASE 2026. Participants will train a system to learn ten sound classes in three different domains, with learning at each incremental task not having access to previous task data. Submitted systems will be ranked by the overall average accuracy calculated over the three domains. The provided baseline system obtains a modest performance of 44.9\% accuracy over the three domains, mostly due to erroneous inference of the domain for the test sample.
