4 ms·
> Researchers from the University of Texas at Austin recently trained an AI model to decipher the gist of a limited range of sentences as individuals listened t
by dorkwood 3y ago
> Researchers from the University of Texas at Austin recently trained an AI model to decipher the gist of a limited range of sentences as individuals listened to them—gesturing toward a near future in which artificial intelligence might give us a deeper understanding of the human mind.
In what way could something like this give us a deeper understanding of the human mind? My understanding is that these ML-trained models are essentially black boxes.
- dinvlad 3y agoExactly this. They just fed more data into it and suddenly it’s “emergent” behavior, even though there’s no scientific basis behind why it would be emergent (and the fact that the human cognition does not appear to operate like this). I’m frankly horrified by all of science (including physics) jumping on the bandwagon of unscientific methods, simply because we are out of better ideas. It’s basically pseudoscience, at this point.