One super interesting company that are only going to get more and more relevant in the world we live in today is Sentient Technologies. As well as being one of the most funded company in A.I. and being the brains behind some of the technology in Siri, these guys are building out a huge machine learning platform that others can use to make their own products that bit smarter.
Sentient’s first practical application for its AI system was financial, creating a “species” of AI trader that could compete, reproduce, die and eventually evolve to respond to changes in the markets on a level beyond that possible in human traders. A good analogy would be millions of monkeys beating their keyboards executing stock market trades and the system learning which monkeys were making the most money. Those monkeys then survived, and were used to breed more. Essentially a genetic algorithm.
This month, we have been treated to another live demo of exactly how some of the algorithms Sentient technologies are developing translate to real world consumer applications. They have launched an online intelligent system for online shopping that does not just provide a list of recommendations, like Amazon or Asos do, but lets you have a real-time dialogue in the virtual world, just as you might do with a particularly good salesperson in a bricks-and-mortar store. Their visual shopping tool reacts in real time to your choices, actively learning and determining what you are looking for based on previous decisions you have made as a shopper.
Take a look at it on Shoes.ca ( Visual Filter ). Browsing each shoe, you can see instantly just how powerful this is going to be at working out quicker what it is we are looking for online, and indeed increase sales as a result. Whereas Amazon use collaborative filters to work out ‘people who bought this also bought this’, the work involved in computing these relationships has to be done overnight and has it’s problems. This paper tells you just how Amazon are performing this (if you’ve worked on e-commerce applications before, you’ll be all to aware of some of the challenges in making this quick).
Although we are still years away from computers actually becoming sentient and self-aware, this stuff is fascinating.