arXiv Open Access 2025

Audio Flamingo Sound-CoT Technical Report: Improving Chain-of-Thought Reasoning in Sound Understanding

Zhifeng Kong Arushi Goel Joao Felipe Santos Sreyan Ghosh Rafael Valle +2 lainnya
Lihat Sumber

Abstrak

Chain-of-thought reasoning has demonstrated significant improvements in large language models and vision language models, yet its potential for audio language models remains largely unexplored. In this technical report, we take a preliminary step towards closing this gap. For better assessment of sound reasoning, we propose AF-Reasoning-Eval, a benchmark targeting common-sense reasoning and the ability to discriminate among closely related choices. To prepare training corpus for sound reasoning abilities, we propose automatic pipelines that transform existing audio question answering and classification data into explicit reasoning chains, yielding AF-CoT-Train with 1.24M samples. We study the effect of finetuning Audio Flamingo series on AF-CoT-Train and observe considerable improvements on several reasoning benchmarks, validating the effectiveness of chain-of-thought finetuning on advanced sound understanding.

Topik & Kata Kunci

Penulis (7)

Z

Zhifeng Kong

A

Arushi Goel

J

Joao Felipe Santos

S

Sreyan Ghosh

R

Rafael Valle

W

Wei Ping

B

Bryan Catanzaro

Format Sitasi

Kong, Z., Goel, A., Santos, J.F., Ghosh, S., Valle, R., Ping, W. et al. (2025). Audio Flamingo Sound-CoT Technical Report: Improving Chain-of-Thought Reasoning in Sound Understanding. https://arxiv.org/abs/2508.11818

Akses Cepat

Lihat di Sumber
Informasi Jurnal
Tahun Terbit
2025
Bahasa
en
Sumber Database
arXiv
Akses
Open Access ✓