The individual who strolls a bustling city street, a blood cell that detects an invader, and a search engine on the Internet share many similar challenges: There is a vast amount of information and excessive environmental signals, but most of them are of little interest. How are these systems able to operate in the face of this constant flow of complex signals and information?
At our group, we study this issue using tools from statistical mechanics and machine learning. We focus specifically on two problems. One is what algorithms can be efficiently used to infer knowledge about the environment that an organism lives in. The other is how these algorithms can be implemented in biological systems. In most cases the algorithms and their implementations involve complex interactions between many elements, e.g. many neuron in the brain, and this is where statistical physics comes into the scene: it provides the tools and intuitions necessary for studying such complex systems.