Learning to Discover: A Generalized Framework for Raga Identification without Forgetting

2026-01-26

AI Summary

This work addresses the challenge of identifying rare, unlabeled ragas in Indian classical music, where traditional classification models—constrained by a closed-set assumption—struggle to recognize unseen ragas and often suffer from catastrophic forgetting when learning new classes. To overcome these limitations, the paper proposes the first unified framework that integrates continual learning without forgetting and novel class discovery (NCD). By leveraging semi-supervised representation learning, the method jointly optimizes labeled and unlabeled audio data, enabling simultaneous retention of known raga recognition capabilities and discovery of previously unseen ragas. Experiments on standard raga datasets demonstrate that the proposed approach significantly outperforms existing NCD methods, achieving consistent improvements in classification and discovery performance across known, unknown, and overall categories.

Technology Category

• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Constraint Satisfaction and Optimization: Constraint Learning and Acquisition
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

Raga identification in Indian Art Music (IAM) remains challenging due to the presence of numerous rarely performed Ragas that are not represented in available training datasets. Traditional classification models struggle in this setting, as they assume a closed set of known categories and therefore fail to recognise or meaningfully group previously unseen Ragas. Recent works have tried categorizing unseen Ragas, but they run into a problem of catastrophic forgetting, where the knowledge of previously seen Ragas is diminished. To address this problem, we adopt a unified learning framework that leverages both labeled and unlabeled audio, enabling the model to discover coherent categories corresponding to the unseen Ragas, while retaining the knowledge of previously known ones. We test our model on benchmark Raga Identification datasets and demonstrate its performance in categorizing previously seen, unseen, and all Raga classes. The proposed approach surpasses the previous NCD-based pipeline even in discovering the unseen Raga categories, offering new insights into representation learning for IAM tasks.

Problem

• Raga identification
• Indian Art Music
• catastrophic forgetting
• unseen categories
• class discovery

Innovation

• Raga identification
• catastrophic forgetting
• generalized learning framework
• unseen category discovery
• Indian Art Music