Machine learning (ML) is poised for explosive growth over the next two years with an increasing number of projects moving into production by 2020, based on a recent survey titled 'The Future of Machine Learning', conducted by technology marketing organisation Dimensional Research.
Though a diverse set of ML projects are currently initiated by 93% of the respondents, only 22% of these projects have actually moved into production, citing migration as the top technical challenge.
Univa, a specialist in on-premise and hybrid cloud workload management solutions for enterprise high-performance computing (HPC), sponsored the survey that polled 344 technology and IT professionals across the globe and across 17 industries, with technology, financial services and healthcare leading the charge in ML adoption.
“Our customers are already asking for guidance with migrating their HPC and machine learning workloads to the cloud or hybrid environment,” says Univa Navops vice president and general manager Rob Lalonde.
“As a result, we decided to conduct this survey to better understand the type of projects driving value in machine learning, as well as better assess what key challenges users are currently facing that are preventing them from moving their projects into production. We look forward to utilising this data to help guide our customers and recommend the right set of tools and migration options needed to accelerate ML value.”
Survey participants highlighted some key components driving their ability to successfully move ML projects into production:
There is a direct correlation between HPC and ML, with more than 88% of respondents indicating that they are working with HPC in their jobs.
Nearly 9 out of 10 companies surveyed expect to use GPUs as part of their ML infrastructure.
More than 80% of respondents plan to use hybrid cloud for ML projects while keeping costs down.
Though 69% of companies surveyed have three or more teams requesting ML projects, only 1 in 5 companies have ML projects running in production.
The biggest technical challenges cited with current ML projects include the migration of workloads, data and applications.
Yet experts surveyed expect the number of tools used for running ML projects to increase with Amazon and Microsoft benefiting from increased market share.
“We see a tremendous opportunity to help our customers move their ML projects into production,” adds Lalonde.
“This survey revealed that there are a diverse number of projects for ML learning, indicating numerous areas of value. We look forward to working with our customers to help them fully utilize and scale these projects and resources across their on-premise, hybrid and cloud infrastructures.”