STEM Update #3, NSF as University, AI Everywhere All At Once

Tuesday, May 9, 2023


Context: As division director of IIS, I’ve started sending out a short message to the IIS mailing list on the Second Tuesday Every Month (STEM). Here’s the latest installment.

 

As an academic, I’ve been finding it useful to navigate the NSF organization by mapping it to something much more familiar to me, the organizational structure of a typical university. I’ll share this analogy with you in the hopes that it might make a complex entity a little more comprehensible (and maybe even friendlier!).

 

In place of a university president and provost, NSF has a director (Sethuraman Panchanathan, aka Dr. Panch) and a COO (Karen Marrongelle). Like their counterparts at universities, they oversee units having to do with various operational topics (finances, legal, human resources, information services, diversity, public affairs, etc.). Whereas the “content” part of a university is organized into Schools headed up by deans, NSF has directorates led by “Assistant Directors” (ADs). Directorates include Biological Sciences (BIO), Geosciences (GEO), Mathematical and Physical Sciences (MPS), and Social, Behavioral and Economic Sciences (SBE).

 

My AD (dean?) is Margaret Martonosi, the head of the Computer and Information Science and Engineering (CISE) directorate. She works closely with her Deputy AD (my DAD… yes, that’s as weird as it sounds) Joydip Kundu (aka JD). JD is kind of like a vice dean or an associate dean. (The DADs from all the directorates meet regularly. I really want them to call it the Council of DADs after the short-lived NBC series.) In most of the directorates, including CISE, the AD is a rotator (aka IPA, see STEM #1, Margaret comes from Princeton) and the DAD is a permanent federal employee (JD previously worked in IIS and also the White House Office of Management and Budget). They are both really smart and really dedicated. CISE has three divisions and one office (akin to departments and centers in the academic world). In my experience, universities that have schools of computer science have a difficult time dividing things neatly into departments. NSF is no different. The structure here, which is probably not perfect, is: Computing and Communication Foundations (CCF), which covers algorithms, theory, and hardware; Computer and Network Systems (CNS), which covers operating systems and networks, but also educational and diversity initiatives; Office of Advanced Cyberinfrastructure (OAC), which covers topics like numerical and high performance computing, often in support of research in the other directorates; and, of course, Information and Intelligent Systems (IIS), which covers AI, ML, HCI, graphics, databases, visualization, computing for health, data science, computational neuroscience, and various other topics that often involve the boundary between computers and people. IIS is arguably the most intellectually coherent of the divisions, but that’s because the bar is low. Like I said, it’s hard to split things up neatly.

 

In this extended analogy, as division director, I correspond to a university department chair and Wendy Nilsen (see STEM #2), IIS’s deputy division director, is my vice chair. I’ve been a department chair before and didn’t realize this job would have so many parallels. Just like a university department has office staff and faculty, IIS has operational staff and program officers. (You sometimes hear program officers, sometimes program managers, and sometimes program directors. There may be subtle distinctions between these terms, but mostly they are used interchangeably.) Just as faculty in a department are often loosely organized into topic areas, IIS has a cluster structure consisting of Human-Centered Computing (HCC), Information Integration and Informatics (III), and Robust Intelligence (RI) as well as other programmatic initiatives such as Smart Health and the AI Institutes.

 

Whew. That’s NSF in a nutshell. I’ll leave it as an exercise to the reader to decide what the research community corresponds to in this analogy…

 

In other news, CISE leadership has been spending an inordinate amount of time on topics surrounding AI lately. We have taken to calling this period of time “AI Everywhere All At Once”. Of course, there is an explosion of academic interest in the topic, with rapid growth in the number of students wanting to work in the area, departments wanting to hire in the area, and conferences and journals getting huge numbers of submissions in the area. But we’re also seeing increases in interest from other countries wanting to collaborate with US scientists on AI-related topics, other parts of the US Government looking for partnerships and guidance on the topic, and public calls for policy makers to do more to help society navigate this moment in time. As IIS is the intellectual and spiritual home of AI within NSF, I’m involved in a lot of these discussions.

 

One item you might have seen is last week’s White House Fact Sheet on responsible AI innovation: https://www.whitehouse.gov/briefing-room/statements-releases/2023/05/04/fact-sheet-biden-harris-administration-announces-new-actions-to-promote-responsible-ai-innovation-that-protects-americans-rights-and-safety/ . NSF (including me but definitely not only me!) contributed a number of items to the folks putting together this announcement. Our goal was to highlight the amazing work being done in the scientific community to help make sure AI technology is a positive force in society. The list we sent was substantially boiled down, but we expect to have other opportunities to contribute to this important conversation in the months to come. If you have suggestions for items to amplify that show ways research is focused on protecting “rights and safety” in the age of AI, I’d love to hear them!

 

“AI Everywhere All At Once” is an exciting time for all of us who are associated with IIS both inside and outside NSF. Of course, IIS is more than just AI, but this attention is giving us some great visibility and even some clout that extends beyond a narrow view of what AI is. I hope you are finding it exciting and even inspiring and not overly stressful. (I’m sure you are getting questions, if not from journalists then from friends and family. It’s nice that people are interested in what we do!) This too shall pass, but, for now, there’s a lot to contribute.

 

-Michael