How Watson, Holmes and Automatic Statistician differ from each other?


How IBM’s Watson, Wipro’s Holmes and Google’s Automatic Statistician differ from each other ? Are these automated intelligence and cognitive computing systems going to thread jobs of Statisticians and Data Scientists ?



Google’s automatic statistician is a project which was awarded unconditional grant from Google. It is actually run by Cambridge Machine Learning group and is a way to automate machine learning models. The intent is to take datasets and then automate the process to come out with best machine learning algorithms. It would be meant to be used to take away hassle from machine learning process. There are a few other companies like Ayasdi and Nutonian who are trying to do similar thing. Even SAS would be running a similar project in house.

IBM Watson is a completely different beast. It would take questions from any end user or business user, collate data all by itself, analyse it and throw the outcome! So it looks far more evolved at this stage. But things can change rapidly!

I don’t know much about Wipro’s Holmes. From what I could read online, it would be similar to Watson but with different capabilities and use cases.