Our AI model provides your unique wellness scores by analysing your speech during ] different speech-based activities. Over time, the model learns from your voice to increase the accuracy of your scores
What data is used to train the AI model?
The AI model is trained on a diverse and extensive dataset, comprising hundreds of thousands of minutes of speech recordings from a wide array of individuals. This dataset is continuously evolving, with a focus on inclusivity and balance, representing a broad spectrum of ages, genders, backgrounds and accents. Notably, our datasets include individuals with a broad spectrum of experience with different mental health symptoms, including many users without mental health conditions. By leveraging this extensive and varied dataset, we ensure that our AI model is reliable, ethically sound and applicable across many different user contexts.
Final note:
Our AI model offers valuable insights into your mental wellness through a tailored analysis of your voice and speech patterns. At thymia, we prioritise accuracy and transparency. If you are currently suffering from any condition or symptom which may affect how your voice sounds, you may experience variability in your scores. For personalised support, please consult a qualified professional.