Deep clustering machine learning enables AI to distinguish individual voices in a crowd

AI system can isolate individuals’ voices from other environmental noise, including other voices. Such a system has many potential uses, both benign and nefarious. The ability is rapidly improving to untangle signals from noise and identify which signals are from which sources. The approach should be able to apply to other kinds of signals too, not only sounds.

About Mark H

Information technologist, knowledge management expert, and writer. Academic background in knowledge management, social and natural sciences, information technologies, learning, educational technologies, and philosophy. Married with one adult child who's married and has a teenage daughter.

One thought on “Deep clustering machine learning enables AI to distinguish individual voices in a crowd

  1. Perhaps an AI-enabled solution can be devised for hidden hearing loss (HHL), an auditory deficit in which a person does well on the typical tone-recognition hearing tests, but has difficulty understanding another person’s speech in environments where many people are speaking in the background or there are other background noises.

    HHL is attributed to synaptic disruption in the auditory nerves, rather than to loss of cilia in the ear canal. Nerve demyelination may also contribute.

    I foresee a new type of hearing aid that uses AI-enabled signal separation and clarification to help people with HHL. Early versions might require the user to take some action, perhaps with a smartphone app, to choose which voice to focus on. More advanced versions might use eye-gaze tracking or other natural cues to focus.

    Test your own hearing in simulated noisy conditions:

    Recent research on causation:

Leave a Reply