Cloud Haskell: First impressions and applications to processing large image datasets

摘要

Cloud Haskell is a new distributed processing domain-specific language embedded as a library written in the Haskell programming language. We present a short introduction to the challenges faced, some common patterns and our experiences through implementing image processing and machine learning tasks used in object recognition. We examine closely the task of namely batch processing images and demonstrate the scalability of our examples with benchmarks on a small add-hoc cluster of machines typical of many office networks.