Everyone will understand: scientists are creating a programme to help recognise speech difficulties in patients with Parkinson's, dysarthria and other conditions

Artificial intelligence is increasingly being used for medical purposes. Thus, the Speech Accessibility Project initiative launched by the University of Illinois in cooperation with technology giants such as Google, Apple, Meta, Amazon and Microsoft will help to better understand people with various speech disorders.

Here's What We Know

The programme aims to "decipher" the slurred speech of people suffering from, for example, Parkinson's disease, dysarthria and other serious neurological conditions.

To do this, speech samples are collected from people with different features in order to train a neural network to recognise and correctly interpret it.

How it works

• At the beginning, speech samples are collected from people with different speech patterns and disorders. These can be people with accents, stuttering, dyslexia, dysarthria and other features;
• the collected speech samples are analysed and processed to create datasets that are used to train the AI;
• The AI learns to recognise and interpret patients' unique patterns and, once tested, can be implemented into various services (voice assistants, dictation systems and other speech recognition technologies).

So far, the Speech Accessibility Project recognises only 30% of difficult speech, but the scientists are optimistic that in the near future they will be able to increase the efficiency of the programme, which in turn will help improve communication between people.

Source: MedicalXpress