Brewing Beer with Python

Lightning talk on using Artificial Intelligence to generate beer recipes

Transcript

1. Brewing Beer
 with Python @MarcoBonzanini 
 @PyDataLondon
2. Python + Beer = Over-engineering
3. MALT WATER HOPS YEAST
4. 1.Mashing (grains + water) 2.Boiling (+ hops) 3.Cooling 4.Fermentation (+ yeast)
5. Grain bill: 2Kg Pilsner malt 1Kg Pale malt 1Kg Wheat malt 1Kg Wheat flakes 0.5Kg Munich malt 0.5Kg Oat flakes Mash: 30m at 55C 60m at 67C 15m at 75C Boil: 40g Magnum @ 60m 40g Mosaic @ 10m 20g Coriander seeds @ 10m In fermenter: 5 gallons Fermentation: 2 weeks at 20C Yeast: M21 OG: 1.059 FG: 1.015 IBU: 64
6. Grain bill: 2Kg Pilsner malt 1Kg Pale malt 1Kg Wheat malt 1Kg Wheat flakes 0.5Kg Munich malt 0.5Kg Oat flakes Mash: 30m at 55C 60m at 67C 15m at 75C Boil: 40g Magnum @ 60m 40g Mosaic @ 10m 20g Coriander seeds @ 10m In fermenter: 5 gallons Fermentation: 2 weeks at 20C Yeast: M21 OG: 1.059 FG: 1.015 IBU: 64
9. Recipe
 URLs XML
 Recipes Text
 Recipes requests pybeerxml
10. Neural Networks
11. Recurrent Neural Networks (RNN) http://colah.github.io/posts/2015-08-Understanding-LSTMs/
12. RNN unrolled http://colah.github.io/posts/2015-08-Understanding-LSTMs/
13. Long Short Term Memory (LSTM) http://colah.github.io/posts/2015-08-Understanding-LSTMs/