New method improves speech recognition models while reducing knowledge loss.
― 4 min read
Electrical Engineering and Systems Science - Audio and Speech Processing
RSSLatest Articles
Latest Articles
Discover new methods to enhance hearing aid performance and speech clarity.
― 5 min read
A novel method improves speech recognition tasks using less labeled data.
― 5 min read
This article examines recent improvements in creating written audio descriptions.
― 5 min read
Efficient audio recognition technology designed for low-power television devices.
― 4 min read
SCHmUBERT offers a fresh approach to creating symbolic music with AI.
― 6 min read
Using AI to identify invasive pink snail eggs for better management.
― 5 min read
A new model enhances confidence scores in speech recognition systems.
― 5 min read
New techniques improve understanding of dysarthric speech in communication systems.
― 5 min read
A novel unsupervised approach enhances voice isolation in audio mixtures.
― 4 min read
A new benchmark for evaluating machine learning models in understanding speech across languages.
― 6 min read
This article discusses methods to enhance phone classification using audio features.
― 6 min read
A new model enhances audio perception and reasoning capabilities in AI.
― 6 min read
NASS improves voice isolation in noisy environments, outperforming traditional methods.
― 4 min read
A novel approach to enhance audio quality for synthetic voice creation.
― 6 min read
New techniques improve sound recognition efficiency and reduce labeling costs.
― 6 min read
Enhancing sound quality metrics using new loudness calculation methods.
― 5 min read
AlignAtt enhances simultaneous speech translation with improved speed and quality.
― 5 min read
A new method ensures privacy in speech classification without sacrificing performance.
― 6 min read
This study shows how to adapt TTS technology to different accents efficiently.
― 5 min read
AMII model enhances communication for socially interactive agents through improved non-verbal behavior.
― 5 min read
Using federated learning to enhance speech analysis for Parkinson's diagnosis across languages.
― 5 min read
This study focuses on recognizing Arabic dialects using advanced methods and limited data.
― 4 min read
Introducing a model that integrates various data types for complex tasks.
― 6 min read
Researchers are improving how we detect animal sounds automatically.
― 6 min read
Discover how Whisper adapts to various speech tasks using prompt engineering.
― 5 min read
This study examines ways to enhance ASR for low-resource languages using data techniques.
― 4 min read
FastFit improves speech generation speed without losing sound quality.
― 5 min read
A new method improves keyword detection in audio recordings.
― 5 min read
This study introduces a method to better measure tongue movement during speech using X-ray data.
― 6 min read
AED-EEND system enhances speaker diarization by integrating advanced techniques for better accuracy.
― 5 min read
Pengi merges audio understanding and text generation into a single model.
― 7 min read
A new approach aims to minimize delays in speech recognition systems while maintaining accuracy.
― 4 min read
A new method enhances keyword spotting systems for better performance in changing audio.
― 4 min read
A new TTS system enhances speech generation across multiple languages with limited data.
― 6 min read
CoDi enables simultaneous generation of diverse content types from various inputs.
― 4 min read
New techniques improve sound separation from Ambisonics mixes for better audio experiences.
― 6 min read
A new method improves speech models while reducing resource needs.
― 6 min read
New methods using speech show promise in identifying breathing patterns and health conditions.
― 4 min read
MIDI-Draw allows anyone to make music by drawing melodies intuitively.
― 5 min read
New techniques borrowing from image processing enhance audio quality evaluation.
― 6 min read