Date: 2018-12-15 04:38 pm (UTC)
danieldwilliam: (Default)
From: [personal profile] danieldwilliam
I might see if I can persuade the lads at work to build a cocktail recipe system.

Date: 2018-12-20 10:48 am (UTC)
danieldwilliam: (Default)
From: [personal profile] danieldwilliam
I think there are two (of either two or many) options.

You can give it a database of ingredients and let it pick from that or you can give it a bunch of recipies and see if it can be trained to make acceptable cocktails. I think in either case you need to have some sort of feedback. Probably two layers. You can probably do layer one quickly with a generative adversarial network, see if one neural network can correctly guess which cocktail recipes have been generated by the other system and which are from a list of real cocktails.

Layer two is deciding if the cocktails that look like they are real are actually appealing to humans. You might be able to do this with training data if the training data included some sort of metric about popularity (although you would need to be careful what that popularity metric was actually measuring, taste, name, marketing, or price) or you might need to bring in actual humans.

Date: 2018-12-15 08:12 pm (UTC)
andlosers: (Default)
From: [personal profile] andlosers
I had a friend who worked for Tesla. One summer they got an email from Elon saying, again, that Tesla wasn’t a 9 to 5 company, and they should let their families know that they would see less of them for a while.

The email was sent from Hawaii.

He is a terrible person and I’ve never heard anyone who’s interacted with him say anything nice.

April 2019

 1 2 3 4 5 6
7 8 9 10 11 12 13
14 15 16 17 18 19 20

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags
Page generated Apr. 21st, 2019 12:46 am
Powered by Dreamwidth Studios