How the Department of Energy Is Tapping AI to Transform Science, Industry and Government - Ep. 236
24:51 Share

Hello, and welcome to the NVIDIA A I podcast on your host. No cabbies. This past october, leaders from the public sector join in video customers and partners at A I summit in washington, D. C.

For three days of connection and discussion around all of the innovative and meaningful work being done with A I one of the panels at the sunset AI for science, energy and security, focused on how artificial intelligence is transforming scientists, discovery, economic growth and national security, and how the U. S. Government can lead the way in developing safe and trust for the A.

I. To address national and global chAllenges. It's a great discussion, and I encourage you to watch IT via video on demand. But even Better than that, we're fortunate enough to have the chance to deve a bit deeper into the role the U. S, department of energy is playing N, A, I development spoiler alert.

They're doing a lot more than you might think with us now as Helena food, director of the office of critical emerging technology C, E, T. At the U. S.

Department of energy director, food previously served this director for technology and national security at the White house national security council as director of international science and technology CoOperation, interested research for the office of science at the department of energy and at the way house office of science and technology policy director, whose credentials extend much further than that. But i'll leave IT there so we can welcome her onto the show. Clean food. Thank you so much for taking the time to join the video. I podcast very bigger.

No, thanks for inviting me. So perhaps .

we can start with the C, E, T. And what your role is as director. Tell us a bit about that.

sure. So the office of critical and emerging technologies, or C E T, because we love a good chronic, is a new office at the department. IT was launched in december of twenty twenty three. Uh, so we're less than a year old coming on that one year anniversary.

And our senior leadership a at the department really wanted to have a central node within the department that was focused on leveraging capabilities and experts across the department added seventeen national labs in specific areas of critical emerging technologies, and our office focuses on four specific technology areas. A I mro, electronics, want to information science and biotechnology. And these four we really see as foundational technologies that enable so many other the emerge technologies.

absolutely. And so how does that fit into the broader scope? We were chatting for a second before we started recording. I was watching the panel you were on and there are opening reMarks from the secretary of energy as well. So again, it's a great panel um and I learned a lot of watching IT and in particular I was struck I think there is a joke made on the panel about calling yet department of everything, I was struck by how much the deal we broadly ader. So maybe you could take a second if you and kind of speak to a little bit to that .

yeah I think that's a that's a joke that our secretary made. But actually many secretaries of, uh, energy have made this joke because as everyone comes into the department, they see the vast scope of what we actually do here at the v. So as you said, many people think about due as the clean energy department and and we are, uh, but we do do so much more.

Uh, we fund basic scientific discovery and research. We steward a scientific infrastructure all across the country. And these are some of the world's most powerful x rays, particle accelerators, colliders, sort of really, really big science.

We advances energy research applications, which is part that I think many people know and and touch. Uh yeah, that comes to dev. Um as well as a where are the sector risk management agency for the electricity sector.

But we also play a key role in the national security space through the national nuclear security administration. And on top of that, we store the seventeen national labs, which are located all across the country. Many of them were buried out of the manhattan project, but today, they serve as that scientific infrastructure backbone of the united states and the capability not just for D.

O. E, but for the broader U. S. government. Um and so I touched a little about that scientific infrastructure. Dev has thirty four scientific user facilities. Now these are places where scientists within the united states, from academia, from industry as well as international scientists, they can access these facilities on appeal review basis. Of course, these are the kinds of particle accelerators, molecular foundry ies, supercomputers that the department are stewards for the country.

So when I think about and and I think a lot of the conversations that, that having on the podcast, when I think about the relationship between A I accelerated computing, these other advanced technologies and emerging technologies and energy, I sort of think of two different things. And i'm curious how how you think about IT. If you think about IT and similar bucket tonight, one is compute, is energy A I.

All this stuff, the bigger the faster, the more powerful than more energy. So there's that, and there's been talk lately about, you know, is, is, is nuclear are going to play all these different things. How does clean energy or no bug? He added.

All this relate to powering the GPS, the data centers, all that stuff. And then the other side of IT is how can all the technology in form, finding new sources of newble energy, making IT more efficient, all of that good stuff, to make, you know, the whole process of creating and using energy Better. For I can even imagine how complex these things are when you think about them. But how do you describe the relationship between energy, and specifically what you are doing, C, E, T, N, D, A. And relationship to all technology stuff?

Great question. And really, the department of energy is sitting at that, this intersection of A I as applied to the energy sector and the energy availability for A I to power the data centers that these A I models are trained. And so we are absolutely laser focus on both of these issues.

And I think one really concrete example um of where the sort of all comes together really is in some of the partnerships that we've been able to develop with industry to drive energy efficiency of computing, right? This is the story behind a the excess scale computing project and the initiative that U. E. Has been very, very involved in over the last decade. And you know, the very beginnings of that was this Price opposition, that to actually get to excess scale computing, you would need to make huge advances and energy efficiency.

And I stop you for a second and just asked to kind of define X, A scale computing everyone or or the project. I should say.

sure. sure. So this is something that was really done in deep partnership between the office of science at the department of energy and the national nuclear security administration, OK. So this is a really a committed partnership that was begun back in two thousand sixteen and running through twenty twenty four.

And this was really about how we could build some of the most powerful and fastest supercomputers in the world, right? And so x scales really referring to how quickly how many floating point Operations per second could actually be performed. Operations alive .

said, define like what do you mean? But really, I just wanted to talk about the project.

but that's a great definition. So thank you. You know back at at that time, do we was already investing in advanced supercomputing, but we recognize that he really pushed the boundaries of science, uh, energy applications and national security. We would need even faster and more powerful supercomputers to do the kinds of exquisite modeling and simulation I would be needed for the nuclear stockpile, but also a model climate um and so there were so many different applications in that space.

But really at the time there was an energy efficiency question, how could we build these kinds of huge, huge supercomputers and how would we manage the power on globe and the cost associated with power in the supercomputers? And so way back during that time, the goals that were set for excess scale were not set by what was technically possible at the time. But we're really set by just a number how much we were willing to pay with the life of the stem um and then just divided IT by five uh and that was the goal.

And really at the time, I think many folks thought we are crazy, but I think that's really the the magic of D V H. I. I don't want to call the magic is not magic.

It's just the ability of the department to partner deeply with industry to accomplish things at scale. I think that is really how I would think about where do we fits in the larger A I ecosystem. I think one other thing that would be really helpful to kind of way. So I talked about A I and energy and energy for A I as one intersection where D O E sits. I think the other one that is not very widely known, but kind of goes to all the different parts of the O E, is that we sit at this intersection of open science and classified .

science at the open science and national security. Classified science kind of know things that there .

is many a national security applications of science, of course, and much of that really is dependent on advances in open science. And I think the fact of deal, we as you know, a decision drivers R N D agency um that is able to bridge these two arenas and to do that at scale, I think scale again is an important d marter. There is really what makes working a deal we are so exciting.

And Frankly, as we bring you back to um our office of critical emersion technologies, before technology areas I mentioned you know A I mr. Electronics, quana, information science and biotechnology, there are all still use technologies. So I think really as we think about you know, both opportunities to advance these technologies and opportunities to manage the risks of those technologies, you really do need that duality of approach, right, where the science is going and what the national security implications are going to be to rest this effectively.

Are there specific A I related initiatives going on within D O E with N C T that you can talk about?

sure. yes. And we're very excited about the opportunity space of A S I aren't already do have A A very exciting uh proposed initiative called fast frontiers in A I for science, security and technology um which we see as a real opportunity space for a step change in U S.

Government capability in A I. We also see the advancements that industry is making. We need to be able to harness those advancements for our mission space.

And I think the recently issued uh national security memorandum on A I just last week, um really direct departments and agencies in the national security space to do just this. But we also think that there's even more to do, right? So fastest are organized in four general buckets, one around data. One around a dancing computer, one on developing new kinds of models and the fourth around the application space. So how do we actually use these models to solve things in the real world?

Yes, there's been um actually was recording a podcast early this week with somebody and use the same lines forgive me listeners, but there is kind of this sense that twenty, twenty two, twenty twenty three in in the generate A I space in particular or sort of the years that l and the idea of models kind of explore the mainstream and other started, you know, to take notice and really think about we need to invest, we need to do things.

And then twenty twenty four now has been the year where folks in the application space are really starting to home in on, okay, how do we take all this power and translated into something that is, you know practical, useful and easy for an end user to can pick up and go with? Is there a similar sort of timeline or you kind of in the same sort of head space with relative to building the applications and that focus now in your government work? Or or is IT a different? Are you a different so I .

think that's a really interesting question because when IT comes to agency deployment of industry models, many of these are large language models, right? And so we see tremendous efficiency, potential efficiencies within the department on utilizing these. And we have some things that we're beta testing now and excited about them. But what we are talking about and fast is really around how we harness the scientific data that do.

We generates just some of the most unclassified and classified scientific experimental data in the world, you can and how we really advance computing in this space in a way, uh, that h that we did were able to do through the access scale computing project, but really think further beyond at the horion. Yeah and then the models themselves, we really think that much of industry obviously is focused on large lan, which models and there's been a lot of exciting developments. And even those models are becoming much Better at scientific reasoning.

Really, we're seeing tremendous advances in that space, and that really excites us because that means the time to discover, whether is for in discovery science or in the energy applications or in national security, that time is going to be cut so significant. But we also think that there is going to need to be new kinds of models, additional kinds of models that are not language based, but are, you know, graph neural networks or physics informed models, that are the kinds of models that need to be able to learn right from the kinds of data that we hold. And that can be then shown to translate into the physical world, which is why we have lavatories.

Um so I think this is this is where we get so excited because we think the opportunities for partnership with industry are really enormous and a place where there hasn't yet been the same kind of attention and focus more broadly because Frankly, these are things that are in the realm of science, very like government investment in science. That's why there is that sort of public private partnership opportunity. If I could just add one thing though, please, because there's been a lot of attention when we talk about for the A I for science, there's been a lot of attention, for example, on an f 是 and you know I think what is important to note there is your alcohol statistical models。 They're trained on data from experimentally determined protein structures that were only made possible by the use of dues unique march, get light sources, OK and the protein structure data that was used, and that is free and open to the public, are stored on D O, E funded routine database, which gets funding from other U.

S. Government departments and agencies. But I think that's an example of where there is that public private dimension to even industry advancements.

My yesterday is Helena food. Helena is the director of the office of critical and emerging technologies at the U. S. Department of energy. And as we've been talking about the department of energy, the work that they're doing, the work the director foods team is doing as well, touches so many different parts of private life, public life, different parts of the the government working, you know, with science and industry and everything are so many different places. Let's kind of land on, uh, a note of opportunity. And i'm not going to ask you to predict the future, but what are some of the spaces that you secular your teams are really excited about when IT comes to the applications of lunch language models? A I, any of the technology you we've been talking about, our other things we haven't gotten to yet.

Thanks for that question. Uh, I I think I might answer in two time scales. The first is the near term. Uh, so we have a number of activities already under way that are really seeking to harness these powerful tools. One example here is really around smart grads.

And using A I for example, are great deployment office funded cotler general electric to deploy utility data smart meters, which have gp s inside of them to customer. And you know, we think that there's an opportunity also for local models that help to aggregate loads like electric vehicles, he pumps uh, and that provide utilities insight while also protecting customer privacy. Our office policy and uh, pacific northwest national lab are also looking this voltaic initiative, which is really around how we can harness large language models for permit and how we find efficiencies in that process.

Permitting is in granting permits to .

citizens to do permit process like, uh, environmental permitting reviews. Um and in fact, one of the things that we were able to do just earlier this year was to take the entire corpus. I've leap a data over the last decade or so and make that A I ready and available and is now open and available to the public for researchers to work on our industry to develop neutrals around.

We are working with other agencies in the federal permitting council to see how they can potentially use this tool for their permitting processes. So that's exciting and that at the federal level. But the full take initiative is actually also focused on expanding to state at local area because there is even more variants across state and local ordinances, for example, for E B charging or for uh other any other kind of uh sighting yeah the mention of the .

grid and the smart red meters with you know the G P S and then able to computer on the fly I mean that he hits home for me you know i'm kind of the stereotypical northern california have got E V. The whole thing.

But IT really makes me think about all of the opportunities in the physical world that if the infrastructure in place and you've got this data work with, which obviously is huge, I can even imagine the the throws of data that are around waiting to be process to to use. But you know, you can deploy things at the edge like that, right to a customer. In my case, you know, to where I live in a resident person can have that smart meter or going to have, you know them, the ability to shift the loads locally. And I can only imagine the possibilities that opens up for great resilience. And and you know all of those things.

it's exactly right. And I think moving beyond the short term into that medium and long term time scale, we are so excited about the opportunity of A I applied to materials discovery, which I think has applications across science, energy.

national security, right?

We often talk about this example from uh p and microsoft to look at the universe of potential materials and battery that is. But I think that's just only the tip of the story, right? Obviously, battery materials, carbon captures, orbans hydrogen catalysts.

There is so much opportunity to discover new materials that will have both impacts on like abundancy and affordability of energy but also for strategic technologies like critical minerals and hyper sonics, right? There's many, many applications um across this factor similarly in the physical sciences, right there's obviously uh a lot of work already underway on physical form models for climate. Um we need to be able to bridge climate and weather across multiple time skills and that will have huge implications for disaster response and prepared ness and you know just climate modeling more generally.

yes. Yeah, as you said, that just kind of the tip of the eyes and the idea of you know seventy percent more efficient car batteries built from renewable energy is sort of mind blowing to me anyway but that really is just kind of the the beginning of the story, right? And the early tings here in general, before we wrap up any closing thoughts, people listening out there, perhaps there are you know, organization leaders, uh, you know, I T leaders, people in their whether to private or public sector who are really starting to work on ah they understand the opportunity.

They've use some A I tools. They're thinking in that same way, right, that kind of short and kind of mid term time scales advice you would give to somebody who is you know whether the organization is smaller or governmental scale, who's trying to kind of lead that initiative to take advantage of of A I and and a thought for you sustainably minded way. Any words of advice from your experience?

Yeah I am also the acting chief, A I officer or the department of energy as I think about how the department of energy is going to utilize A I and work very closely with our achieve information office on this. We think really and this goes for reno within india, we also companies as they think about 定 ai， we really need to innovate up that letter of trust.

So we really need to think about I mean, when I think about this for D A V, I think about what are the immediate use cases, how do we make sure that they're not right and safety impacting, what are the processes we have in place to manage those risks. But I think the overall advice here really would be, no, this is a transformative tool. IT is a tool in the tool.

Bell is very powerful. We should all figure out how to use this to the best possibility to amplify and augment the work that we're doing. Um so we are thinking about this day and day out of due. We think that partnerships are going to be essential to our success. Um when IT comes to applying A I for abroad, science, energy and national security mission.

anta asked for a listeners who would like to learn more about any of the many things that that you covered. What the office of critical technology is, dog is doing good work where some places they can go online .

to get started yeah well, the office of pretty al emerging technologies has a website, and that is a place where we really look to point to all of the different and amazing activities happening across the department and the national laboratory. So that's a really good resource to learn more about. Are frontiers and A I for science, security and technology initiative and also test spas that are available idea we training opportunities at the national laboratories foundation models that we already developed uh in partnership with others and so much .

more and time oh, again, thank you so much for taking the time directer food to come on the podcast and kind of extend the conversation. You started the A I summit and hopefully we can do this again down the line. In the meantime, i'm really excited to follow the work that you and your teams and the rest of the folks in the government are doing to make life Better for all of us. So thank you.

Thank you. Thanks so much.