Efforts to develop AI are gathering steam fast. On Monday, the White House issued a federal plan to help develop technical standards for AI following up on a mandate contained in the Administration’s AI Executive Order of last February. Last week the Computing Community Consortium, associated with NSF, issued a 20-year roadmap for AI research. Three weeks ago, The Department of Energy held the first of four AI for Science Town Halls that could inform an AI strategy and implementation plan not unlike the ongoing Exascale Initiative.
“The Trump Administration continues to deliver on the American AI Initiative, the national strategy for U.S. leadership in AI. Public trust, security, and privacy considerations remain critical components of our approach to setting AI technical standards. As put forward by NIST, Federal guidance for AI standards development will support reliable, robust and trustworthy systems and ensure AI is created and applied for the benefit of the American people,” said Michael Kratsios, chief technology officer of the U.S. in the announcement.
Here’s a brief excerpt from the standards document (A Plan for Federal Engagement in Developing Technical Standards and Related Tools) which is being overseen by the National Institute of Standards and Technology (NIST).
“Standards should be complemented by an array of related tools to advance the development and adoption of effective, reliable, robust, and trustworthy AI technologies. These tools—which often have overlap- ping applications—include, but are not limited to:
- Data sets in standardized formats, including metadata for training, validation and testing of AI systems. Data standards are vital in measuring and sharing information relating to the quality, utility and access of data sets.26 They can preserve privacy, ensure accessibility, assist potential users in making informed decisions about the data’s applicability to their purpose, and help prevent misuse.
- Tools for capturing and representing knowledge and reasoning in AI systems to promote consistent formulation of, reasoning with, and sharing of knowledge, thereby promoting interoperability of AI systems and minimizing their misunderstandings and inferential errors.
- Fully documented use cases that provide a range of data and information about specific applications of AI technologies and any standards or best practice guides used in making decisions about deployment of these applications. For the use cases to be of real value, they must be accompanied not only by explicit information about the parameters of use, but also by the practical implications of such uses for persons who may be affected by AI deployments.”
The CCC roadmap (A 20-Year Community Roadmap for Artificial Intelligence Research in the U.S.) is similarly ambitious and sets three central recommendations: 1) create and operate a national AI Infrastructure to serve academia, industry, and government through four interlocking capabilities; 2) re-conceptualize and train an all-encompassing AI workforce; and 3) [invest in needed] core programs for basic AI Research – “The new resources and initiatives described in this Roadmap cannot come at the expense of existing programs for funding AI research…All of this will require substantial, sustained federal investment over the course of the 20-year period covered by this Roadmap, but the outcomes will be transformative.”
Under national infrastructure recommendations there are four substantial objectives:
- Open AI platforms and resources: a vast interlinked distributed collection of “AI-ready” resources (such as curated high- quality datasets, software, knowledge repositories, testbeds for personal assistants and robotics environments) contributed by and available to the academic research community, as well as to industry and government.
- Sustained community-driven AI challenges: organized sequences of challenges that build on one another, posed by AI and domain experts to drive research in key areas, building upon—and adding to—the shared resources in the Open AI Platforms and Facilities.
- National AI Research Centers: multi-university centers with affiliated institutions, focused on pivotal areas of long-term AI research (e.g., integrated intelligence, trust, and responsibility), with decade-long funding to support on the order of 100 faculty, 200 AI engineers, 500 students, and necessary computing infrastructure. These centers would offer rich training for students at all levels. Visiting fellows from academia, industry, and government will enable cross-cutting research and technology transition.
- Mission-Driven AI Laboratories: living laboratories for AI development in targeted areas of great potential for societal impact. These would be “AI-ready” facilities, designed to allow AI researchers to access unique data and expertise, such as AI-ready hospitals, AI-ready homes, or AI-ready schools. They would work closely with the National AI Research Centers to provide requirements, facilitate applied research, and transition research results. These laboratories would be crucial for R&D, dissemination, and workforce training. They would have decade-long funding to support on the order of 50 permanent AI researchers, 50 visitors from AI Research Centers, 100-200 AI engineers and technicians, and 100 domain experts and staff.
The proof of these efforts, of course, will be in the pudding but it does seem that the push to develop AI as the Next Big Thing is becoming increasingly concrete. (For an overview of AI’s potential in science see HPCwire article, AI is the Next Exascale – Rick Stevens on What that Means and Why It’s Important)
Link to CCC announcement: https://www.cccblog.org/2019/08/12/trump-administration-issues-plan-for-federal-engagement-in-ai-technical-standards/?utm_source=feedblitz&utm_medium=FeedBlitzRss&utm_campaign=cccblog
Link to NIST AI standards plan: https://www.nist.gov/sites/default/files/documents/2019/08/10/ai_standards_fedengagement_plan_9aug2019.pdf
Link to CCC 20-year roadmap: https://cra.org/ccc/wp-content/uploads/sites/2/2019/08/Community-Roadmap-for-AI-Research.pdf