Feature-Rich Audio Model Inversion for Data-Free Knowledge Distillation Towards General Sound Classification

2023-03-14 · Zuheng Kang, Yayun He, Jianzong Wang, Junqing Peng, Xiaoyang Qu, Jing Xiao

Data-Free Knowledge Distillation (DFKD) has recently attracted growing attention in the academic community, especially with major breakthroughs in computer vision. Despite promising results, the technique has not been well applied to audio and signal processing. Due to the variable duration of audio signals, it has its own unique way of modeling. In this work, we propose feature-rich audio model inversion (FRAMI), a data-free knowledge distillation framework for general sound classification tasks. It first generates high-quality and feature-rich Mel-spectrograms through a feature-invariant contrastive loss. Then, the hidden states before and after the statistics pooling layer are reused when knowledge distillation is performed on these feature-rich samples. Experimental results on the Urbansound8k, ESC-50, and audioMNIST datasets demonstrate that FRAMI can generate feature-rich samples. Meanwhile, the accuracy of the student model is further improved by reusing the hidden state and significantly outperforms the baseline method.

📄 PDF Abstract BibTeX arXiv:2303.07643

Code (0)

등록된 구현이 없습니다.

Tasks

Data-free Knowledge Distillation Knowledge DistillationSound Classification

Methods 이 논문이 사용한 방법론

Knowledge Distillation A very simple way to improve the performance of almost any machine learning algorithm is to train many different models on the same data and then to average their predictions.…

DirectAudioEdit: Inversion-Free Text-Guided Audio Editing via Diffusion Prediction Contrast

2026-06-05 · Zhengkun Ge, Xiaoqian Liu, Haoran Zhang, Yuan Ge 외 arxiv

Text-guided audio editing aims to modify the language-specified acoustic content while preserving edit-irrelevant source components. Existing training-free methods typically rely on inversion-based editing. While inversi…

Virtual Consistency for Audio Editing

2025-09-21 · Matthieu Cervera, Francesco Paissan, Mirco Ravanelli, Cem Subakan arxiv

Free-form, text-based audio editing remains a persistent challenge, despite progress in inversion-based neural methods. Current approaches rely on slow inversion procedures, limiting their practicality. We present a virt…

FreeSonic: Training-Free Temporal-Aware Decoupled Attention for Precise Audio Editing

2026-06-13 · Yuxuan Jiang, Mingyang Han, Yusheng Dai, Andong Wang 외 arxiv

Text-to-audio (TTA) generation has made significant strides, yet achieving precise and consistent audio editing remains a major challenge. However, existing methods struggle to balance temporal consistency with backgroun…

Introducing Model Inversion Attacks on Automatic Speaker Recognition

2023-01-09 · Karla Pizzi, Franziska Boenisch, Ugur Sahin, Konstantin Böttinger

Model inversion (MI) attacks allow to reconstruct average per-class representations of a machine learning (ML) model's training data. It has been shown that in scenarios where each class corresponds to a different indivi…

model Speaker Recognition

SAO-Instruct: Free-form Audio Editing using Natural Language Instructions

2025-10-26 · Michael Ungersböck, Florian Grötschla, Luca A. Lanzendörfer, June Young Yi 외 arxiv

Generative models have made significant progress in synthesizing high-fidelity audio from short textual descriptions. However, editing existing audio using natural language has remained largely underexplored. Current app…