How to produce delicious beer with big data and NVIDIA GPU?

ByKaren Neoh

"Brewed beer" which is a bit different from major beer makers is increasing in popularity year by year in the United States,American Beer Brewing AssociationAccording to the company, sales of beer makers are growing rapidly to account for 19% of the annual sales of US beer industry of US $ 101.5 billion (approximately 12.2 trillion yen). Meanwhile, there are businessmen who developed an application to efficiently produce better beer using big data and NVIDIA GPU.

Beer maker uses big data and graphics chips to create a better brew | VentureBeat | Big Data | by Dean Takahashi
http://venturebeat.com/2015/09/02/brewer-uses-graphics-chips-to-create-better-beer/


Better Beer Through GPUs: How GPUs and Deep Learning Help Brewers Improve Their Suds | The Official NVIDIA Blog
http://blogs.nvidia.com/blog/2015/09/02/beer/

Jason Cohen is a professional tea appraiser with parents of an Olive Oil appraiser with a skill arm. Mr. Cohen inherited the taste with excellent sorting ability from his parents founded Tea Research Institute in Pennsylvania and became the world leader in tea culture worldwide. Cohen asked college students to record tea tasting and impressions in order to improve the taste of tea, but the data did not gather easily.

So Cohen asked for a free beer tasting tester, not tea, where the volunteers gathered enough to make a flock and enough data gathered in just a few weeks. Cohen developed an iOS / Android application that analyzes the taste of beer in real time based on the gathered data.


The application can analyze beer and other beverages with 24 flavors without being influenced by factors such as age, race, gender, past experience of eating and drinking. When a local brewer maker etc. inputs information of its own product into the application, the data is sent to the data center which manages artificial intelligence and machine learning algorithm. The data is analyzed by the NVIDIA GPU from the big data of the review that exists in real time, three times faster than usual, and what is missing in your beer is immediately displayed on the dashboard.


Also, since we make new product data every week, data will be added immediately even if there is no evaluation of my product. Cohen is making a new model to analyze more than 100,000 reviews with GPU and deep learning and the application can analyze not only beer but also coffee · alcohol · food data, It also has a function to receive notifications. Although tasting tends to be subjective to the subjectivity of the tester, it has become an unprecedented effort to obtain accurate evaluation of the taste of beer from big data.