CogniVoice: Multimodal and Multilingual Fusion Networks for Mild Cognitive Impairment Assessment from Spontaneous Speech
This addresses the need for automated, multilingual assessment tools for MCI, which is an incremental advancement in medical AI.
The paper tackled the problem of detecting Mild Cognitive Impairment (MCI) and estimating MMSE scores from speech data, achieving improvements of 2.8 points in F1 for classification and 4.1 points in RMSE for regression over baselines.
Mild Cognitive Impairment (MCI) is a medical condition characterized by noticeable declines in memory and cognitive abilities, potentially affecting individual's daily activities. In this paper, we introduce CogniVoice, a novel multilingual and multimodal framework to detect MCI and estimate Mini-Mental State Examination (MMSE) scores by analyzing speech data and its textual transcriptions. The key component of CogniVoice is an ensemble multimodal and multilingual network based on ``Product of Experts'' that mitigates reliance on shortcut solutions. Using a comprehensive dataset containing both English and Chinese languages from TAUKADIAL challenge, CogniVoice outperforms the best performing baseline model on MCI classification and MMSE regression tasks by 2.8 and 4.1 points in F1 and RMSE respectively, and can effectively reduce the performance gap across different language groups by 0.7 points in F1.