Michael Feindt on predictive analytics
From RAAIS 2017 - Michael Feindt, Founder of Blue Yonder · · TheResearchandAppliedAISummitRAAIS
“For us a prediction is a complete probability distribution for each possible future — once you have this you can do complete risk management of every single decision and optimize decisions including their uncertainty.”
On , Michael Feindt, Cofounder at Blue Yonder, spoke about predictive analytics during RAAIS 2017 - Michael Feindt, Founder of Blue Yonder on TheResearchandAppliedAISummitRAAIS.
About Michael Feindt
At the 2017 RAAIS conference, Michael Feindt, co-founder and chief scientific officer of Blue Yonder, discussed the company's work in vertical AI solutions for supply chain and pricing. Feindt stated that Blue Yonder's systems automate repeated decisions, with more than 99% of such decisions capable of being automated. He said that most of the company's clients hire more people after introducing AI systems because they become more effective and profitable. Feindt described Blue Yonder's approach as using a complete probability distribution for each possible future to optimize decisions including their uncertainty. Feindt cited a case with the UK retailer Morrison's, where 13 million automatic ordering decisions per day led to a more than 30% reduction in shelf-life gaps. He also noted that in one client example, the out-of-stock rate fell from about 7.5% to 0.5% after fully automated decision-making. Feindt emphasized that vertical end-to-end solutions combining domain expertise and machine learning are necessary to solve complicated business problems. He also discussed the company's background, noting that his NeuroBayes algorithm was developed during scientific research at CERN, where he said a decision algorithm implemented on a chip reached a world record of 8 billion intelligent decisions per second.