Lightning talk on using Artificial Intelligence to generate beer recipes
Brewing Beer with Python@MarcoBonzanini @PyDataLondon
View Slide
Python +Beer =Over-engineering
MALTWATER HOPS YEAST
1.Mashing (grains + water)2.Boiling (+ hops)3.Cooling4.Fermentation (+ yeast)
Grain bill:2Kg Pilsner malt1Kg Pale malt1Kg Wheat malt1Kg Wheat flakes0.5Kg Munich malt0.5Kg Oat flakesMash:30m at 55C60m at 67C15m at 75CBoil:40g Magnum @ 60m40g Mosaic @ 10m20g Coriander seeds @ 10mIn fermenter: 5 gallonsFermentation: 2 weeks at 20CYeast: M21OG: 1.059FG: 1.015IBU: 64
Recipe URLsXML RecipesText Recipesrequests pybeerxml
Neural Networks
Recurrent Neural Networks (RNN)http://colah.github.io/posts/2015-08-Understanding-LSTMs/
RNN unrolledhttp://colah.github.io/posts/2015-08-Understanding-LSTMs/
Long Short Term Memory (LSTM)http://colah.github.io/posts/2015-08-Understanding-LSTMs/