Acumos Project Aims for Reusable AI Standard

The Linux Foundation, AT&T, and Tech Mahindra last week launched a new initiative to make it easier to use emerging artificial intelligence technology by creating a standard for reusing machine learning models. Dubbed the Acumos Project, the group aims to have code ready to use next year.

While the potential impact that AI and ML can have on businesses is considered immense, so is the learning curve to effectively wield this technology. While the technological barriers to entry are gradually coming down, the domain AI and ML is still dominated by data scientists and data engineers who have the requisite skills and training to incorporate the emerging technology in a meaningful manner.

And once a data scientist creates something using one of the many frameworks and languages availed today – from Spark, Tensorflow and H2O to Numpy, SAS, R and many others – that model is often not applicable to other uses.

The Acumos Project aims to accelerate the democratization of AI and ML technologies by creating an “industry standard for making AI applications and models reusable,” the group says. It also aims to create an ecosystem of technology providers around Acumos to help make AI and ML applications more accessible to regular companies.

AI is a “critical tool for growing our business,” AT&T Labs Vice President of Advanced Technology Mazin Gilbert says in a press release. “However, the current state of today’s AI environment is fractured, which creates a significant barrier to adoption. Acumos will expedite innovation and deployment of AI applications, and make them available to everyone.”

As an “open and connected” AI platform, the Acumos Project will empower those looking to “define the future of AI,” according to Jim Zemlin, the executive director of The Linux Foundation. “Because the platform is open source,” he says, “it will be accessible to anyone with an interest in AI and machine learning, and customizable to meet specific needs.”

Technical details around the Acumos are slim. The group says the Acumos platform “will be user-centric, with an initial focus on creating apps and microservices.” Zemlin says he expects organizations that are doing work with autonomous vehicles, drones, content curation, and analytics to be interested in the project.

Tech Mahindra is an Indian IT services and consulting firm that does a lot of work in the telecommunications industry. The company says it has worked on data analytic projects with 35 of the Fortune 500, including the top two communications service providers in the US (that would be AT&T and Verizon), as well as clients in automotive, pharmaceutical, and chemical manufacturing.

The group plans to have working code and a marketplace opened by early 2018.