FLAMENCO – AI-Powered Early Diagnosis of Communication Disorders in Children.

The FLAMENCO project represents a pioneering step in personalized healthcare solutions, combining Artificial Intelligence (AI), Federated Learning (FL), and advanced data security mechanisms to revolutionize the early diagnosis of communication disorders in children. Funded under the Horizon 2020 TERMINET initiative, FLAMENCO introduces an innovative approach that enables speech therapists, educators, and healthcare professionals to detect developmental communication disorders early and accurately.
The project builds upon an existing AI-powered software suite, which collects and analyzes data from children through interactive gaming experiences and biometric monitoring. Using a combination of eye-tracking, voice analysis, motor response tracking, and heart rate monitoring via smartwatches, FLAMENCO evaluates a child’s communication and cognitive skills to assess potential developmental challenges.
A key innovation of FLAMENCO is the integration of Federated Learning (FL), a decentralized machine learning approach that allows AI models to be trained across multiple devices without transferring sensitive personal data. This ensures that privacy is preserved, as no raw data is shared or stored in centralized servers. To further enhance data security, the project employs Fully Homomorphic Encryption (FHE), a state-of-the-art encryption method that enables computations on encrypted data without compromising its confidentiality.
FLAMENCO also incorporates a clustering-based client selection mechanism that filters out unreliable data sources, ensuring the accuracy and robustness of the predictive model. By tackling challenges such as data imbalance, privacy concerns, and real-world variability in child responses, the project is set to deliver a highly reliable AI-driven tool that can be widely adopted in clinical and educational settings.