HyperMapper black-box optimizer based on Bayesian Optimization.
SLAMBench presents a foundation for quantitative, comparable and validatable experimental research in the form of a benchmark to investigate the performance, accuracy and power consumption of a dense, 3D scene understanding application. SLAMBench, a publicly-available software framework which supports both research in hardware accelerators, software tools, and also, by enabling quantitative evaluation of solution accuracy, research in novel algorithms. It provides clean KinectFusion implementations in most common languages, e.g. C++, CUDA and OpenCL, making this benchmark highly portable. SLAMBench is an open source project, have a try here.
YAO is a Domain-specific Language (DSL) for variational data assimilation. Check out the open source code here and have a look here if you are interested in an introduction of YAO.
