# A Hybrid CNN-BiLSTM Voice Activity Detector

arXiv: https://arxiv.org/abs/2103.03529 · HTML: https://mlsanity.com/paper/2103.03529/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 5 Mar 2021.

## Built on by (most cited first, 3 of 3)

- [Attention Is Not Always the Answer: Optimizing Voice Activity Detection with Simple Feature Fusion](https://mlsanity.com/paper/2506.01365/) (2025)
- [WhaleVAD-BPN: Improving Baleen Whale Call Detection with Boundary Proposal Networks and Post-processing Optimisation](https://mlsanity.com/paper/2510.21280/) (2025)
- [Never Stop Speaking: a Denial-of-Service Attack on End-to-End Speech Language Models](https://mlsanity.com/paper/2608.10405/) (2026)

## Abstract (arXiv)

This paper presents a new hybrid architecture for voice activity detection (VAD) incorporating both convolutional neural network (CNN) and bidirectional long short-term memory (BiLSTM) layers trained in an end-to-end manner. In addition, we focus specifically on optimising the computational efficiency of our architecture in order to deliver robust performance in difficult in-the-wild noise conditions in a severely under-resourced setting. Nested k-fold cross-validation was used to explore the hyperparameter space, and the trade-off between optimal parameters and model size is discussed. The performance effect of a BiLSTM layer compared to a unidirectional LSTM layer was also considered. We compare our systems with three established baselines on the AVA-Speech dataset. We find that significantly smaller models with near optimal parameters perform on par with larger models trained with optimal parameters. BiLSTM layers were shown to improve accuracy over unidirectional layers by ≈2% absolute on average. With an area under the curve (AUC) of 0.951, our system outperforms all baselines, including a much larger ResNet system, particularly in difficult noise conditions.