HPE is throwing its weight behind a new way to shrink the time required to deploy machine learning (ML) and artificial intelligence (AI) workloads from months to mere days.
The company cites Gartner statistics that show enterprise AI adoption has doubled in the last four years. Much of this adoption takes place in areas such as fraud detection, personalised medicine, and predictive analytics.
However, operationalising AI and ML technologies in the ‘last mile' presents plenty of challenges – Gartner says that by 2020, at least 50% of ML projects won't be deployed fully because they don't have the required operationalisation.
HPE designed its new container-based software solution, HPE ML Ops, to support the machine learning lifecycle across on-premise, hybrid cloud, and public cloud environments. It addresses the entire machine learning lifecycle from data preparation and model building, through to training, deployment, monitoring, and collaboration.
The solution, HPE says, will bring a process similar to devops that can standardise machine learning workflows and speed up deployments.
The HPE ML Ops solution also expands capabilities of the BlueData EPIC container software platform, which provides data science teams with on-demand access to containerised environments for distributed AI, ML and analytics.
Additionally, the solution works with a wide range of open source machine learning and deep learning frameworks including Keras, MXNet, PyTorch, and TensorFlow as well as commercial machine learning applications from ecosystem software partners such as Dataiku and H2O.ai.
“Only operational machine learning models deliver business value,” says HPE's SVP and CTO of hybrid IT, Kumar Sreekanti.
“With HPE ML Ops, we provide the only enterprise-class solution to operationalise the end-to-end machine learning lifecycle for on-premises and hybrid cloud deployments. We're bringing DevOps speed and agility to machine learning, delivering faster time-to-value for AI in the enterprise.
The HPE ML Ops solution offers:
- Model Build: Pre-packaged, self-service sandbox environments for ML tools and data science notebooks
- Model Training: Scalable training environments with secure access to data
- Model Deployment: Flexible and rapid deployment with reproducibility
- Model Monitoring: End-to-end visibility across the ML model lifecycle
- Collaboration: Enable CI/CD workflows with code, model, and project repositories
- Security and Control: Secure multi-tenancy with integration to enterprise authentication mechanisms
- Hybrid Deployment: Support for on-premises, public cloud, or hybrid cloud.
“From retail to banking to manufacturing to healthcare and beyond, virtually all industries are adopting or investigating AI/ML to develop innovative products and services and gain a competitive edge,” comments IDC's program vice president of AI strategies, Ritu Jyoti.
“While most businesses are ramping up on the build and train phase of their AI/ML projects, they are struggling to operationalise the entire ML lifecycle from PoC to pilot to production deployment and monitoring.
“HPE is closing this gap by addressing the entire ML lifecycle with its container-based, platform-agnostic offering – to support a range of ML operational requirements, accelerate the overall time to insights, and drive superior business outcomes.
HPE ML Ops is generally available now as a software subscription, together with HPE Pointnext services and customer support.