Unstructured Pruning and Low Rank Factorisation of Self-Supervised Pre-Trained Speech Models
H Wang, WQ Zhang - IEEE Journal of Selected Topics in Signal …, 2024 - ieeexplore.ieee.org
Self-supervised pre-trained speech models require significant memory and computational
resources, limiting their applicability to many speech tasks. Unstructured pruning is a …
resources, limiting their applicability to many speech tasks. Unstructured pruning is a …
Full-Rank No More: Low-Rank Weight Training for Modern Speech Recognition Models
This paper investigates the under-explored area of low-rank weight training for large-scale
Conformer-based speech recognition models from scratch. Our study demonstrates the …
Conformer-based speech recognition models from scratch. Our study demonstrates the …
Sla-former: conformer using shifted linear attention for audio-visual speech recognition
Y **ao, J Huang, X Liu, A Zhu - Complex & Intelligent Systems, 2024 - Springer
Conformer-based models have proven highly effective in Audio-visual Speech Recognition,
integrating auditory and visual inputs to significantly enhance speech recognition accuracy …
integrating auditory and visual inputs to significantly enhance speech recognition accuracy …
Enhancing Quantised End-to-End ASR Models Via Personalisation
Recent end-to-end automatic speech recognition (ASR) models have become increasingly
larger, making them particularly challenging to be deployed on resource-constrained …
larger, making them particularly challenging to be deployed on resource-constrained …
[PDF][PDF] Leveraging Adapter for Parameter-Efficient ASR Encoder
K Shim, J Lee, H Kim - Proc. Interspeech 2024, 2024 - isca-archive.org
The expansion of speech models emphasizes the importance of parameter efficiency in
practical automatic speech recognition (ASR) systems. Parameter sharing, which reuses the …
practical automatic speech recognition (ASR) systems. Parameter sharing, which reuses the …
Speaker Adaptation for Quantised End-to-End ASR Models
End-to-end models have shown superior performance for automatic speech recognition
(ASR). However, such models are often very large in size and thus challenging to deploy on …
(ASR). However, such models are often very large in size and thus challenging to deploy on …
Conformer-Based Audio Visual Speech Recognition with Taylor Attention
Y **ao, J Huang, X Liu, A Zhu - International Conference on Pattern …, 2024 - Springer
Abstract Audio Visual Speech Recognition (AVSR) has witnessed significant advancements
by deploying sophisticated neural networks, particularly Convolutional Neural Networks …
by deploying sophisticated neural networks, particularly Convolutional Neural Networks …
Residualtransformer: Residual Low-Rank Learning With Weight-Sharing For Transformer Layers
Memory constraint of always-on devices is one of the major concerns when deploying
speech processing models on these devices. While larger models trained with sufficiently …
speech processing models on these devices. While larger models trained with sufficiently …
WiFi Sensing at the Edge Towards Scalable On-Device Wireless Sensing Systems
SM Hernandez - 2023 - scholarscompass.vcu.edu
WiFi sensing offers a powerful method for tracking physical activities using the radio-
frequency signals already found throughout our homes and offices. This novel sensing …
frequency signals already found throughout our homes and offices. This novel sensing …