The Linux Foundation is continuing to expand its scope, announcing the launch of the LF Deep Learning Foundation on March 26.
The goal of the LF Deep Learning Foundation is to make it easier to adopt and deploy artificial intelligence and machine learning methodologies for industry-specific use cases, including cyber-security threat detection, network automation and image recognition. The LF Deep Learning Foundation is backed by Amdocs, AT&T, B.Yond, Baidu, Huawei, Nokia, Tech Mahindra, Tencent, Univa and ZTE.
The initial project at the core of the LF Deep Learning Foundation is Acumos, which was announced in November 2017, though few details were publicly disclosed at the time. The Acumos project integrates code contributed by AT&T and Tech Mahindra to enable organizations to more easily deploy AI models.
There are multiple key components that will help to make AI use more widespread, Arpit Joshipura, general manager of networking and orchestration at the Linux Foundation, told eWEEK.
AI frameworks, such as TensorFlow, PaddlePaddle and Caffe, are needed but can be difficult to consume, Joshipura said. That’s why there are various forms of data wrappers available to help get both public and private data into the different AI models. He noted that understanding what data can be shared and how it can be used is also a challenge.
“All the different components need to come together with shared best practices for different verticals, and that’s where Acumos comes in,” Joshipura said. “Acumos as a project brings the algorithms, data and the compute together, creating templates in a marketplace along with an app store that is vertical specific for a given use case that can be shared by peers across a community.”
With Acumos, multiple industry verticals can benefit from AI templates that will enable more rapid adoption, according to Joshipura. Rather than each individual company re-creating the data flow model for a deep learning model, he said the Acumos templates will enable organizations to reuse approaches that have already been proven.
LF Deep Learning
The Linux Foundation doesn’t see Acumos as a stand-alone effort, but rather as the beginning of a broader AI initiative that will emerge over time. To help facilitate a governance framework for how different AI projects can join Acumos at the Linux Foundation, Joshipura said that a decision was made to create the LF Deep Learning Foundation. The LF Deep Learning Foundation is set to be an umbrella organization for multiple AI projects.
“So as more deep learning projects come in, we will already have a governance model that is scoped out,” he said.
The Linux Foundation in recent years has created multiple efforts that have become umbrella projects, or foundations of their own. One example is the LF Networking project, which was created as an umbrella effort for networking in January 2018. The Cloud Native Computing Foundation is another example; it was originally created with just the Kubernetes project and is now home to 17 projects.
Sean Michael Kerner is a senior editor at eWEEK and InternetNews.com. Follow him on Twitter @TechJournalist.