The Collective Knowledge project: making ML models more portable and reproducible with open APIs, reusable best practices and MLOps

This article provides an overview of the Collective Knowledge technology (CK\nor cKnowledge). CK attempts to make it easier to reproduce ML&systems research,\ndeploy ML models in production, and adapt them to continuously changing data\nsets, models, research techniques, software, and hardware. The CK concept is to\ndecompose complex systems and ad-hoc research projects into reusable\nsub-components with unified APIs, CLI, and JSON meta description. Such\ncomponents can be connected into portable workflows using DevOps principles\ncombined with reusable automation actions, software detection plugins, meta\npackages, and exposed optimization parameters. CK workflows can automatically\nplug in different models, data and tools from different vendors while building,\nrunning and benchmarking research code in a unified way across diverse\nplatforms and environments. Such workflows also help to perform whole system\noptimization, reproduce results, and compare them using public or private\nscoreboards on the CK platform (https://cKnowledge.io). For example, the\nmodular CK approach was successfully validated with industrial partners to\nautomatically co-design and optimize software, hardware, and machine learning\nmodels for reproducible and efficient object detection in terms of speed,\naccuracy, energy, size, and other characteristics. The long-term goal is to\nsimplify and accelerate the development and deployment of ML models and systems\nby helping researchers and practitioners to share and reuse their knowledge,\nexperience, best practices, artifacts, and techniques using open CK APIs.\n