Machine learning sublime influence

2 minutes estimated reading time

Scott Galloway talks about the way brands are using AI (machine learning) and the examples are very much in the background.  Welcome to the sublime world of machine learning where the impact on the customer experience won’t be apparent. In many respects this is similar to how fuzzy logic became invisible as it was introduced in the late 1980s.

The Japanese were particularly adept at putting an obscure form of mathematics to use. They made lifts that adapted to the traffic flows of people going in and out of a building and microwaves which knew how long to defrost whatever you put into it. Fuzzy logic compensated for blur in video camera movement in a similar manner to way smartphone manufacturers now use neural networks on images.

The Japanese promoted fuzzy logic inside products to the home market, but generally backed off from promoting it abroad. The features just were and consumers accepted them over time. In a quote that is now eerily reminiscent of our time a spokesperson for the American Electronics Association’s Tokyo office said to the Washington Post

“Some of the fuzzy concepts may be valid in the U.S.,”

“The idea of better energy efficiency, or more precise heating and cooling, can be successful in the American market,”

“But I don’t think most Americans want a vacuum cleaner that talks to you and says, ‘Hey, I sense that my dust bag will be full before we finish this room.’ “

This was also the case with the use technology companies made of Bayes Theory. This was used by the likes of Autonomy and Microsoft Research.

A second technique was rules, put simply IF then THAT. This kind of technology has been used to drive automated trading models and credit card approvals for decades. Pegasystems are one of the leaders in developing rules based processing. Rules based systems could even be built in an Excel macro and would still count as a form of machine learning. 

Finally machine learning needs to think about a number of things with regards the models being used:

  • The importance of accuracy in the use case
  • The level of precision required and ways to indicate that precision means
  • The cost of generation versus other methods, this is very important in terms of computing power and energy consumption 

More information
The Future of Electronics Looks Fuzzy | Washington Post (December 23, 1990)