The AI 'Brewing' Process
by Russ Loignon
For nearly 12,000 years the beer making process has strived for delivering the brewmaster’s desired outcome. In sampling yesterday’s batch, it was plainly evident that there was creativity, effort, vision and personality that went into that process.
That brewing process, while not overly complicated in its basics, does require some keenly learned and developed steps. Some of these steps include sanitization, the “mash”, extract and boil just to name a few. The sanitization is important so that the process and the equipment can be clean and not influenced or contaminated by anything other than the introduced ingredients. The mash enables the enzymes to convert the starch from the grain to sugars. These sugars will provide food for the yeast. The extract and boil steps allow for a continuation to reach desired goal of a fine brew. In simple terms, this step continues the assurance that unwanted elements are removed from the process.
Similar lines can be drawn with AI. While sophisticated AI tools are able to blend disparate data types and sources. Understanding the type, source and cleanliness of the data can aid in achieving the desired learned outcome. As with the brewing process, extraction is important in AI. Data extraction and classification allows the data to begin its work. The data needs to be stored so that the learning process to begin. Tools such as Tensorflow and Accumulo working with a unified data space help the models and can accelerate the learning process.
Interestingly enough, both the brewing process and the AI process leverage experts in their field in order to oversee the process. The brewmaster brings his experience and guidance to oversee the process to ensure the desired taste are reached. The data scientist provides a similar function in AI by governing, overseeing and bringing their expertise to ensure desired results are achieved.
Please stay tuned for tomorrow's AI 'brew thoughts'….