The ICASSP 2026 Automatic Song Aesthetics Evaluation Challenge

2026-01-12

AI Summary

This work proposes a data-driven approach to automatically predict subjective aesthetic ratings of AI-generated music, aligning with human aesthetic preferences. The study introduces the first standardized evaluation benchmark specifically designed for AI-generated music, encompassing overall musicality and five fine-grained aesthetic dimensions. An end-to-end model is developed to predict these ratings directly from audio inputs. Experimental results demonstrate that the proposed system significantly outperforms baseline methods and effectively captures human aesthetic judgments. This framework establishes a reproducible, human-aligned paradigm for evaluating AI-generated music, offering a principled alternative to conventional metrics that often fail to reflect nuanced perceptual qualities valued by listeners.

Technology Category

• Humans and AI: Learning Human Values and Preferences
• Machine Learning: Learning Preferences or Rankings
• Cognitive Modeling & Cognitive Systems: Computational Creativity

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• User Modeling, Personalization and Recommendation: Metrics for user behavior and evaluating success
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications

Abstract

This paper summarizes the ICASSP 2026 Automatic Song Aesthetics Evaluation (ASAE) Challenge, which focuses on predicting the subjective aesthetic scores of AI-generated songs. The challenge consists of two tracks: Track 1 targets the prediction of the overall musicality score, while Track 2 focuses on predicting five fine-grained aesthetic scores. The challenge attracted strong interest from the research community and received numerous submissions from both academia and industry. Top-performing systems significantly surpassed the official baseline, demonstrating substantial progress in aligning objective metrics with human aesthetic preferences. The outcomes establish a standardized benchmark and advance human-aligned evaluation methodologies for modern music generation systems.

Problem

• Automatic Song Aesthetics Evaluation
• AI-generated songs
• subjective aesthetic scores
• musicality prediction
• human-aligned evaluation

Innovation

• Automatic Song Aesthetics Evaluation
• music generation
• aesthetic scoring
• human-aligned evaluation
• benchmark