Project Submission: Decoding

[WonkyWill99]

Group name: Lissencephaly
Group members: Brian, Heidi, Anakin, and Will

Project description:
Take the brain signals and turn them back into audio.

AudioLM
Decoding
Language Model

[snastase]

Might be worth checking out as an easy starting point? https://huggingface.co/tasks/text-to-speech

[snastase]

I think it would be relatively straightforward to fit a linear model that maps from electrode data onto text embeddings, then use a text-to-speech model to turn those brain-predicted embeddings into audio.