How do machines know when we will click, buy, lie or die?

I was talking with Eric Siegel, PhD., founder of Predictive Analytics World, a conference  about machine learning and predictive analytics. We began by talking about the importance of data and predictive analytics, while trying to separate what’s real from the hype. 

I liked how Eric was able to describe what makes it exciting.

“Machine learning is the most actionable form of data crunching… we’re using machine learning to render predictions to enable large scale operations and render more effective the things that companies do that make the world go round. Targeted marketing, fraud detection, credit risk management, assembly lines, manufacturing, supply chain… all of these areas where there’s lots of individual decisions that could be even improved even slightly has a tremendous impact on the bottom line. Those improvements are driven by per-individual decisions.” 

Machine learning is not the same as artificial intelligence -- which is one of those buzzwords that people use, without knowing what it is. 

“The meaning of the word intelligence is not just vague. There is no formal definition that could reasonably satisfy. And that’s why it’s not a reasonable goal to pursue with engineering.”

Both of us were in agreement that AI should be left to science fiction, which can help us to discover how to become better human beings. 

Machine learning is about getting very practical: for instance, if you have enough data points about a customer and can connect with them -- “will they purchase this other product if I send them a brochure?” 

However, if you want to know how machines predict when people die, you should watch my new Recurrent Patterns video conversation with Eric Siegel.