홈 › Papers

Domain-incremental audio classification using domain-specific experts and prototype classifier

2026-06-22 · Jongyeon Park, Do-Hyeon Lim, Sang-won Park, Hong Kook Kim, Kyungdeuk Ko, Hyeongcheol Geum, Jeong Eun Lim arxiv

This technical report presents submission systems for Task 7(domain-incremental audio classification) of the DCASE 2026 Challenge. The main obstacle is that, the system is unable to access to past or future domain's data at once. We approached domain-incremental learning (DIL) as a frozen-feature replay problem. At each incremental stage, one or two compact experts are trained and then kept fixed; at the final stage, the penultimate features from all frozen experts are concatenated and used to train a lightweight per-class prototype classifier solely on cached features. This design prevents catastrophic forgetting by preserving each expert models at inference. To retain earlier-domain knowledge without storing raw audio, some experts were trained with DeepInversion-based generative replay. A cross-stage regression imputer was trained to fill the expert feature slots that did not yet exist at an ealier stage. We submit four fully DIL-compliant systems: three systems based on diverse frozen five-expert backbones and their cross-stack ensemble achieving 78.15% micro / 77.03% macro on the development set, outperforming every individual backbone on both evaluations.

📄 PDF Abstract BibTeX arXiv:2606.22952

Code (0)

등록된 구현이 없습니다.

Tasks

Incremental Learning Audio Classification

Domain-Incremental Learning for Audio Classification

2024-12-23 · Manjunath Mulimani, Annamaria Mesaros

In this work, we propose a method for domain-incremental learning for audio classification from a sequence of datasets recorded in different acoustic conditions. Fine-tuning a model on a sequence of evolving domains or d…

Audio Classification Classification Incremental Learning Multi-Label Classification+1

We Need a More Robust Classifier: Dual Causal Learning Empowers Domain-Incremental Time Series Classification

2026-01-15 · Zhipeng Liu, Peibo Duan, Xuan Tang, Haodong Jing 외 arxiv

The World Wide Web thrives on intelligent services that rely on accurate time series classification, which has recently witnessed significant progress driven by advances in deep learning. However, existing studies face c…

Time Series Classification Incremental Learning

Better Semi-supervised Learning for Multi-domain ASR Through Incremental Retraining and Data Filtering

2025-06-05 · Andres Carofilis, Pradeep Rangappa, Srikanth Madikeri, Shashi Kumar 외

Fine-tuning pretrained ASR models for specific domains is challenging when labeled data is scarce. But unlabeled audio and labeled data from related domains are often available. We propose an incremental semi-supervised …

named-entity-recognition Named Entity Recognition Named Entity Recognition (NER)NER

Unsupervised Improvement of Audio-Text Cross-Modal Representations

2023-05-03 · Zhepei Wang, Cem Subakan, Krishna Subramani, Junkai Wu 외

Recent advances in using language models to obtain cross-modal audio-text representations have overcome the limitations of conventional training approaches that use predefined labels. This has allowed the community to ma…

Acoustic Scene Classification Classification Scene Classification zero-shot-classification+1

FRIDA -- Generative Feature Replay for Incremental Domain Adaptation

2021-12-28 · Sayan Rakshit, Anwesh Mohanty, Ruchika Chavhan, Biplab Banerjee 외

We tackle the novel problem of incremental unsupervised domain adaptation (IDA) in this paper. We assume that a labeled source domain and different unlabeled target domains are incrementally observed with the constraint …

Domain Adaptation Generative Adversarial Network Unsupervised Domain Adaptation