The development of AI-related technologies often requires resources from large technology companies to proceed smoothly. Facebook, for one, has invested heavily in this area and recently opened up its AI tools to provide practical help to the open source development community.
While recommendation and personal systems are driving the successful adoption of deep learning in the industry, these networks remain under the radar in academia, according to Facebook’s paper. By providing a detailed description of state-of-the-art recommendation systems and open source, they hope to expose more people to the unique challenges these networks face for further ZBO shelters experimentation, modeling algorithms, system collaboration design and benchmarking.
In addition to DLRM, Facebook has recently opened up other AI models or frameworks, including PyRobot, PyTorch Robot Framework and PyTorch Hub, which enhance the repeatability of AI models and hopefully provide the development community with more tools to drive technology development.