AI Speech Analysis Achieves 92% Accuracy in Detecting Alzheimer’s Disease
Researchers at the Vector Institute for Artificial Intelligence have developed a novel approach to detect Alzheimer’s Disease (AD) using speech analysis and simple AI models. The study challenges the assumption that complex, context-based models are necessary for accurate detection, demonstrating that a Word2Vec-based model can achieve 92% accuracy. This method is significantly cheaper and less invasive than current diagnostic procedures. The research utilized the ADReSS dataset, comprising 156 balanced speech samples from participants describing the 'Cookie Theft' picture. By combining word embeddings with linguistic features like utterance length and type-token ratio, the team created a robust classification system. The findings highlight the potential of Natural Language Processing (NLP) in addressing the growing global dementia crisis, offering a scalable tool for early screening. As the population ages, such non-invasive technologies could play a critical role in managing the projected increase in dementia cases, providing a practical solution for healthcare systems worldwide.
Wire timeline
AI Speech Analysis Achieves 92% Accuracy in Detecting Alzheimer’s Disease
Researchers at the Vector Institute for Artificial Intelligence have developed a novel approach to detect Alzheimer’s Disease (AD) using speech analysis and simple AI models. The study challenges the assumption that complex, context-based models are necessary for accurate detection, demonstrating that a Word2Vec-based model can achieve 92% accuracy. This method is significantly cheaper and less invasive than current diagnostic procedures. The research utilized the ADReSS dataset, comprising 156 balanced speech samples from participants describing the 'Cookie Theft' picture. By combining word embeddings with linguistic features like utterance length and type-token ratio, the team created a robust classification system. The findings highlight the potential of Natural Language Processing (NLP) in addressing the growing global dementia crisis, offering a scalable tool for early screening. As the population ages, such non-invasive technologies could play a critical role in managing the projected increase in dementia cases, providing a practical solution for healthcare systems worldwide.
Vector Institute for Artificial Intelligence