Hacker News new | ask | show | jobs
by orbital-decay 1189 days ago
The disk is unable to extract the correlations, nor is it able to apply the knowledge; it transparently stores the data verbatim. The model doesn't store the training set, it extracts the complex correlations from it, and is able to make actual predictions based on the knowledge it extracted.

But yeah, the "knowledge" and "understanding" are hard to define formally, so this discussion can be endless. Common well-defined terms are required.
1 comments
2-718-281-828 1189 days ago
the model does not extract knowledge. an external algorithm trains the models parameters and then the model is fed a string that is also evaluated externally based on the models configuration.
link
orbital-decay 1188 days ago
Semantics. You could say the same about the disk - the data doesn't get magically teleported from the magnetic plates to the RAM, it needs a lot of underlying hardware to read and transfer it.

Model is not just a set of weights, it's inseparable from the underlying architecture, the way to train and to apply them in practice.

link
2-718-281-828 1188 days ago
> You could say the same about the disk

that's exactly my point

link
orbital-decay 1188 days ago
Yeah, and that's arguing about semantics. We could do that in a loop endlessly, ignoring the fact they're fundamentally different.
link