A conceptual framework for learning to listen by reward: Curiosity-driven search for novel sources

2026-05-19

AI Summary

This study addresses the challenge of enabling an agent to autonomously develop auditory perception capabilities in environments lacking fine-grained audio annotations, relying solely on reward signals. To this end, the work proposes a reinforcement learning framework that integrates intrinsic curiosity with audio representation learning, introducing novelty search into the audio domain to encourage the agent to actively explore and discover novel sound sources. This approach establishes, for the first time, a systematic reward-driven paradigm for “learning to listen,” thereby filling a methodological gap in exploration strategies within audio-focused reinforcement learning. Preliminary experiments demonstrate the feasibility of agents effectively responding to previously unheard sound sources under fully unsupervised conditions.

Abstract

Reinforcement learning is a powerful learning paradigm that has spearheaded progress in numerous domains. Its core promise lies in learning through high-level goals without the need for granular labels. However, it still remains elusive in the realm of audio, where it has received substantially less attention than in computer vision or other domains. The key question remains: how can agents learn to listen purely via reward-driven exploration? In this contribution, we present an overview of previous attempts and a new conceptual framework for learning to listen by reward. Our approach depends on the continuous search for novel sound sources. We formulate our framework, discuss open technical challenges, and present a first proof-of-concept implementation that showcases the feasibility of our approach.

Problem

• reinforcement learning
• audio perception
• reward-driven exploration
• curiosity-driven learning
• novel sound sources

Innovation

• reinforcement learning
• curiosity-driven exploration
• novel sound sources
• learning to listen
• audio perception