In this article we explain the challenges with deploying ML/DL models in production and how USENIX OpML can help bring participants for different disciplines to address the herculean task of safely managing the model life cycle in production.
Machine learning (ML) and its variants such as deep learning (DL) and reinforcement learning are starting to impact every commercial industry. The 2019 USENIX Conference on Operational Machine Learning (OpML ‘19), dedicated to operational machine learning and its variants, will focus on the full life cycle of deploying and managing ML into production. The goal of the conference is to help develop robust practices for scaling the management of models (i.e., artifact of learning from big data) throughout their life cycle. Through such practices, we can help organizations transition from manually hand-holding to automated management of ML models in production (i.e., ML version of the move in server operations from “pets to cattle”).
