Summary of Disentangling Textual and Acoustic Features Of Neural Speech Representations, by Hosein Mohebbi et al.
Disentangling Textual and Acoustic Features of Neural Speech Representations
by Hosein Mohebbi, Grzegorz Chrupała, Willem Zuidema, Afra Alishahi, Ivan Titov
First submitted to arxiv on: 3 Oct 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Machine Learning (cs.LG); Sound (cs.SD); Audio and Speech Processing (eess.AS)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary In this paper, researchers tackle the challenge of separating complex neural speech models into distinct components that can be used in various applications. They propose a disentanglement framework based on the Information Bottleneck principle to separate textual and acoustic features in speech representations. The framework is applied and evaluated for emotion recognition and speaker identification tasks, quantifying the contribution of each feature at different model layers. Additionally, the authors explore using their approach as an attribution method to identify the most important speech frames from both textual and acoustic perspectives. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper develops a way to break down complex neural models that understand speech into two parts: one focused on what’s being said (textual) and another on how it sounds (acoustic). This helps in real-world applications where privacy is a concern, as it can suppress the encoding of personal details like gender or speaker identity. The method is tested for recognizing emotions and identifying speakers, showing how different features contribute to these tasks. |
