Unveiling and harnessing hidden attention sinks: Enhancing large language models without training through attention calibration
Attention is a fundamental component behind the remarkable achievements of large
language models (LLMs). However, our current understanding of the attention mechanism …
language models (LLMs). However, our current understanding of the attention mechanism …
[HTML][HTML] Automatic Speech Recognition Advancements for Indigenous Languages of the Americas
Indigenous languages are a fundamental legacy in the development of human
communication, embodying the unique identity and culture of local communities in America …
communication, embodying the unique identity and culture of local communities in America …
Exploration of Adapter for Noise Robust Automatic Speech Recognition
H Shi, T Kawahara - arXiv preprint arXiv:2402.18275, 2024 - arxiv.org
Adapting a robust automatic speech recognition (ASR) system to tackle unseen noise
scenarios is crucial. Integrating adapters into neural networks has emerged as a potent …
scenarios is crucial. Integrating adapters into neural networks has emerged as a potent …
Continual learning optimizations for auto-regressive decoder of multilingual asr systems
Continual Learning (CL) involves fine-tuning pre-trained models with new data while
maintaining the performance on the pre-trained data. This is particularly relevant for …
maintaining the performance on the pre-trained data. This is particularly relevant for …
Wav2Gloss: Generating Interlinear Glossed Text from Speech
Thousands of the world's languages are in danger of extinction--a tremendous threat to
cultural identities and human language diversity. Interlinear Glossed Text (IGT) is a form of …
cultural identities and human language diversity. Interlinear Glossed Text (IGT) is a form of …
Continual Learning With Embedding Layer Surgery and Task-Wise Beam Search Using Whisper
Current Multilingual ASR models only support a fraction of the world's languages. Continual
Learning (CL) aims to tackle this problem by adding new languages to pre-trained models …
Learning (CL) aims to tackle this problem by adding new languages to pre-trained models …
A Parameter-efficient Language Extension Framework for Multilingual ASR
Covering all languages with a multilingual speech recognition model (MASR) is very difficult.
Performing language extension on top of an existing MASR is a desirable choice. In this …
Performing language extension on top of an existing MASR is a desirable choice. In this …
AutoAI2C: An Automated Hardware Generator for DNN Acceleration On Both FPGA and ASIC
Recent advancements in Deep Neural Networks (DNNs) and the slowing of Moore's law
have made domain-specific hardware accelerators for DNNs (ie, DNN chips) a promising …
have made domain-specific hardware accelerators for DNNs (ie, DNN chips) a promising …
Low Resource Language Adaptation using Two-stage Regularization for Multilingual ASR
A significant portion of the global population speaks multiple languages, including many low-
resource languages on which current multilingual ASR models perform poorly. To improve …
resource languages on which current multilingual ASR models perform poorly. To improve …
[PDF][PDF] Low-resource Language Adaptation with Ensemble of PEFT Approaches
Despite recent advances in automatic speech recognition (ASR) performance on common
languages, a large fraction of the world's languages remain unsupported. Parameter …
languages, a large fraction of the world's languages remain unsupported. Parameter …