His name is Shane Legg. He earned his doctorate from USI's Faculty of Informatics (while researching at the Swiss AI Lab IDSIA) and is one of the founders of DeepMind, a London-based start-up concerned with technologies that bridge the gap between artificial intelligence and neuro-science, particularly in the field of learning techniques. On 26 January of this year, the company was acquired by Google for around 500 million dollars. According to analysts in this sector approached by the Financial Times, the Mountain View colossus offered better terms than its rival Facebook, which was also interested in making the acquisition.
As well as Shane Legg, three other members of DeepMind had been members of the research group led by Prof. Schmidhuber (Scientific director of the IDSIA, a joint USI/SUPSI institution): Alex Graves, Daan Wierstra and Tom Schaul. Prof. Schmidhuber, who is also leader of the Deep Learning research group, is proud of his former student, but not surprised by the news: "Since 1989, we have been working on artificial neural networks (NN), inspired by the human brain, which are able to learn complex tasks adaptively. Some have already been put into practice and provide the very best methods for recognizing the wording of handwritten French, Arabic and Chinese texts. The great thing about them is that they learn by themselves – thanks to sophisticated algorithms – on the basis of millions of training images, and then generalize what they have learned to test images they have never seen before. Because of their ability to recognize images, NNs are very important for the future of search engines, medical diagnosis and self-steering motor cars. Since 2009, our methods have won nine prizes in fiercely contested international competitions devoted to visual recognition. In some cases, they have been the first to out-perform human intelligence. I am particularly impressed by the way Shane has managed to combine entrepreneurial flair with his knowledge of the Deep Learning sector, to achieve such spectacular financial success".
According to Prof. Andrew Ng, director of Stanford Artificial intelligence and Lab leader of the Large-scale deep learning project at Google:"Schmidhuber's lab has been one of the pioneering groups in deep learning; for example, their work on GPUs (Graphics Processing Unit) and on scalability has substantially influenced my own (and many others') work, and his deep learning algorithms have also been knocking over benchmark results like crazy. For example, his team recently achieved amazing accomplishments on recognizing traffic signs from vehicles”.