Cambridge Machine Learning Systems Lab (CaMLSys)
Our lab investigates a variety of open problems that sit at the intersection of machine learning and various forms of computational systems (viz. embedded, cloud, mobile). The scientific contributions of the lab often take one of two forms. First, the development of innovative theoretically principled machine learning methods — especially those with applications to the modeling of data such as image, audio, spatial and inertial information. Second, the design and architecture of algorithms, system software and hardware that treat machine learning computation as a first-class citizen — this often results in transformative increases in training and inference efficiency. Our unifying aim is to invent the next-generation of device– and cloud-based systems able to perceive, reason and react to complex real-world environments and users with high levels of precision and efficiency. We seek to achieve this impact through holistic full-stack approaches that encourage lab members with skills in algorithms, hardware, statistics, mathematics and software to work closely together to solve critical challenges in this area. Our cross-discipline lab is based in the Department of Computer Science and Technology at the University of Cambridge.
Launching the Flower Federated Learning Tool
At this week’s google FL workshop, we are announcing our open source tool that makes it easier to devise and evaluate FL algorithms under real-world conditions: http://flower.dev. The paper can be found on arxiv.
- Research Group is Moving to Cambridge
Talk by Shangzhe Wu
Shangzhe Wu presented their work on “Unsupervised Learning of Probably Symmetric Deformable 3D Objects from Images in the Wild” which was awarded Best Paper at CVPR 2020!