The Redwood Center is proud to announce that Ph.D. student Yubei Chen has graduated (actually back in December) and has continued on as a postdoc, as he works on various unsupervised learning topics. Yubei plans to stay with us until the Fall when he will be joining Yann LeCun’s group at Facebook AI Research, where he intends to continue his work on unsupervised learning and energy-based models. After he finished his undergraduate studies at Beijing’s Tsinghua University, Yubei came to Berkeley as an electrical engineer studying high-performance computing and computer architectures. Eventually, he gravitated toward statistical signal processing and unsupervised learning topics, joining the Redwood Center in early 2013. His thesis work proposed a framework for signal representation called The Sparse Manifold Transform, which, along with his other work on unsupervised learning, enriches our current perspective on signal representation in the brain. The Sparse Manifold transform unifies ideas from sparse coding, manifold learning, and slow feature analysis to build a sparse signal representation with certain desirable topological properties. In addition to its general mathematical motivation, the Sparse Manifold Transform may explain certain properties of complex cells in the visual cortex. Yubei recently gave a Redwood Seminar summarizing some of this work, which you can view here. Congratulations Yubei!