Google launches new research group in Zurich
On Thursday, Google unveiled its new research group, «Google Research Europe,» which will focus on machine learning. The team, led by Engineering Director Emmanuel Mogenet, will be based in Zurich.

Google Research Europe supports software developers and researchers who specialize in machine learning. This will give them the opportunity to drive Google's global product development in the heart of Europe and carry out their research work as part of a broader, global approach. The new team will work closely with Google researchers around the world, all of whom will contribute their specific knowledge to the cutting edge of ML technology. They will share their ideas on ML technologies and techniques so that useful products and tools can be developed for global application.
In addition to this collaboration, the team will focus on three core areas of machine learning:
Machine learning is the science of making things smart. An important part of this is machine learning (ML) - for example, teaching a computer to recognize patterns instead of programming it to follow certain rules. The adaptive software is called a neural network and remotely resembles a real brain. It consists of millions or billions of "neurons", small computing units, each of which performs a simple calculation and then passes the information on to the neurons to which it is connected. As a whole, the neural network can learn to recognize relatively complex patterns. In this "deep learning", the neurons are arranged on different levels so that each level learns from the level below. It therefore learns the pattern of the pattern of the pattern. Smart Reply in Inbox, for example, uses the latest neural networks to generate suggestions for short replies to incoming messages.
Language processing and comprehension is a field of computer science, artificial intelligence and computational linguistics that deals with the interaction between computers and human language. For example, when you speak to the Google app on your smartphone, speech recognition turns sounds into words (Google has used deep learning to improve speech recognition by around 25 percent), natural language processing helps to understand what you mean. In addition, a version of deep learning helps to rank the search results.
Machine Perception takes on the big challenge of understanding images, sounds, music and videos, e.g. you can search for everything from "hug" to "dogs" with Google Photos because the system uses Google's latest image recognition system.
Within these core areas, the team will look for ways to optimize the ML infrastructure. In addition, the research results will be made available to the community, enabling them to be applied in practice.
Emmanuel Mogenet on the announcement: "Some of the world's best technical universities are based in Europe. It therefore makes sense to build a team here. We look forward to working with the excellent local computer science institutions and hope to make further valuable contributions to the academic community, whether through our publications or academic support."
Behind many of the products that hundreds of millions of people use every day, from Translate and image search to Smartreply for Inbox , lies Google's ongoing research into machine learning. The offices in Zurich are now home to Google's largest development site outside the United States, with over 1800 employees from 75 nations. The technology on which the Knowledge Graph is based was developed in Zurich and engineers developed the conversation technology for the Google Assistant in Allo here.













