As we build machines that listen not just to our words but to our emotions, the responsibility to use that power ethically ...

The goal of the SpeechHGT project is to develop a Transformer-based model that utilizes both linguistic and acoustic features to improve the early detection of Alzheimer's Disease. The model aims to ...

When performing emotion recognition on long audio such as speeches, cutting the entire audio into segments that contain only one emotion is the first step. However, current research usually needs a ...

A web application that uses machine learning to detect fake news in Malay language texts. The application provides user authentication, history tracking, and uses state-of-the-art machine learning ...

Scaling Zero-shot Text-to-speech (TTS) to large-scale datasets has been demonstrated as an effective method for improving the diversity and naturalness of synthesized speech. At the high level, ...

Methods: To address these limitations, we propose a novel framework named MCOT (M ultimodal Fake News Detection with C ontrastive Learning and O ptimal T ransport). MCOT integrates textual and visual ...