Martin Szummer

About

The machine learning revolution is ON - and as a researcher and entrepreneur, I want to stay at its forefront. I have been building my machine learning expertise for 20 years at Google DeepMind, Microsoft Research, MIT and Cambridge University. I like to be broad: I have created "magic"​ systems that understand language, recognize visual objects and infer people's intentions. I like to be deep: I have published award-winning papers, in many areas including deep learning, kernel methods, and Bayesian methods. Designing the right algorithms makes a big difference: Microsoft's search engine was unable to match queries to relevant ads well. Through novel machine learning methods, we were able to match synonymous phrases, and this increased online advertising revenues by four million dollars, in the first two months. At DeepMind, we estimated and increased user clicks on Google ads over the long-term, more valuable than just increasing them short-term. I developed pioneering machine learning methods using causality for marketing that changes behaviors, not just follows it. I have gone all the way from research to business; I co-founded and served as the CTO for a spoken dialog company t

Experience

Education