3 ms·
I'm reminded of why Watson failed, and the problem with ml and ai in general- you can't peek under the hood to see why something happened, or how to keep it fro
by zdragnar 8y ago
I'm reminded of why Watson failed, and the problem with ml and ai in general- you can't peek under the hood to see why something happened, or how to keep it from happening without a lot of time, a lot of hard work, and a whole lot of carefully groomed data.
- EeThahs2 8y ago> you can't peek under the hood to see why something happened That's myth. There are approaches to analyze and debug NNs, deep dream basically fell out from one of those.
- bglusman 8y agoThis is especially memorable/well treated in the Cory Doctorow novella/short story "Human Readable" https://craphound.com/stories/2005/10/12/human-readable/ https://craphound.com/stories/2005/10/12/human-readable/