# Semi-Supervised Speech Confidence Detection using Pseudo-Labelling and Whisper Embeddings

arXiv: https://arxiv.org/abs/2606.16505 · HTML: https://mlsanity.com/paper/2606.16505/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 31 references that are in the map.
- First posted on arXiv 15 Jun 2026.

## Builds on (most cited first, 8 of 8)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [TED-LIUM 3: twice as much data and corpus repartition for experiments on speaker adaptation](https://mlsanity.com/paper/1805.04699/) (2018)
- [Emotion Recognition from Speech Using Wav2vec 2.0 Embeddings](https://mlsanity.com/paper/2104.03502/) (2021)
- [SEP-28k: A Dataset for Stuttering Event Detection From Podcasts With People Who Stutter](https://mlsanity.com/paper/2102.12394/) (2021)
- [SPICE: Self-supervised Pitch Estimation](https://mlsanity.com/paper/1910.11664/) (2019)
- [Detecting Multiple Speech Disfluencies using a Deep Residual Network with Bidirectional Long Short-Term Memory](https://mlsanity.com/paper/1910.12590/) (2019)
- [Whisper in Focus: Enhancing Stuttered Speech Classification with Encoder Layer Optimization](https://mlsanity.com/paper/2311.05203/) (2023)
- [Exploring Multilingual Unseen Speaker Emotion Recognition: Leveraging Co-Attention Cues in Multitask Learning](https://mlsanity.com/paper/2406.08931/) (2024)

## Abstract (arXiv)

Understanding speaker confidence is crucial in educational settings, as it can enhance personalised feedback and improve learning outcomes. This study introduces a novel framework for detecting speaker confidence by integrating human-engineered features with embeddings from the Whisper encoder. To address data limitations, a pseudo-labelling technique is employed to expand the labelled dataset, allowing the model to learn from both human-annotated and model-generated labels. The framework combines traditional speech features including pitch, volume, rate of speech, and the presence of disfluencies and stress, with Whisper embeddings, and uses a co-attention mechanism to fuse these representations and achieve an overall accuracy of 75%. This study contributes to advancing speech analysis, enabling applications that support personalised learning and speaking skill development.