The T12 System for AudioMOS Challenge 2025: Audio Aesthetics Score Prediction System Using KAN- and VERSA-based Models

TL;DR

The T12 system for AudioMOS Challenge 2025 introduces a novel ensemble of KAN- and VERSA-based models for predicting audio aesthetics scores, achieving top correlation results across multiple evaluation axes.

Contribution

It presents a new ensemble approach combining KAN-based and VERSA-based models for improved audio aesthetics score prediction.

Findings

Achieved the highest correlation scores among submissions.

Ensemble model improved prediction accuracy.

Released inference models for practical use.

Abstract

We propose an audio aesthetics score (AES) prediction system by CyberAgent (AESCA) for AudioMOS Challenge 2025 (AMC25) Track 2. The AESCA comprises a Kolmogorov--Arnold Network (KAN)-based audiobox aesthetics and a predictor from the metric scores using the VERSA toolkit. In the KAN-based predictor, we replaced each multi-layer perceptron layer in the baseline model with a group-rational KAN and trained the model with labeled and pseudo-labeled audio samples. The VERSA-based predictor was designed as a regression model using extreme gradient boosting, incorporating outputs from existing metrics. Both the KAN- and VERSA-based models predicted the AES, including the four evaluation axes. The final AES values were calculated using an ensemble model that combined four KAN-based models and a VERSA-based model. Our proposed T12 system yielded the best correlations among the submitted systems, in three axes at the utterance level, two axes at the system level, and the overall average. We also released the inference model of the proposed KAN-based predictor (KAN #1-#4).