WEBVTT

1
00:00:23.310 --> 00:00:31.809
Alejandro Suarez: Welcome, folks. We're going to give just a couple of minutes for everyone to get settled into the zoom room and we'll get started shortly

2
00:00:41.290 --> 00:00:56.039
Alejandro Suarez: as we're getting started, you're welcome to take a look at some of the links on screen. These are some of the initial opportunities and ways to provide feedback for the near pilot launch and what is available currently.

3
00:01:33.880 --> 00:01:35.629
Alejandro Suarez: So I think we're

4
00:01:35.760 --> 00:01:39.530
Alejandro Suarez: getting to still going up a little bit and get started

5
00:01:39.580 --> 00:01:41.000
Alejandro Suarez: one more minute.

6
00:02:04.030 --> 00:02:14.150
Alejandro Suarez: Oh, and uh apologies! If you saw a sharepoint com link that was uh, not the not the right link. We will get the appropriate link sent out uh in a moment.

7
00:02:19.500 --> 00:02:23.899
Alejandro Suarez: You should now see some of the links to the resources that are on our current slide.

8
00:03:08.150 --> 00:03:19.020
Alejandro Suarez: All right. Well, thank you. Everyone for joining us for this pilot Webinar about the National Ai Research Resource pilot effort.

9
00:03:19.030 --> 00:03:34.050
Alejandro Suarez: This Webinar is being recorded uh slides for this Webinar, as well as a recording of our discussion today will be available on the Nsf. Event page from which you registered in the days following this session.

10
00:03:34.060 --> 00:03:50.140
Alejandro Suarez: So without further ado. I'd like to introduce our office director of the Nsf. Office of advanced Cyber infrastructure, Katie and Tepas, who will talk a little bit about the vision for the National Ai research resource and the launch of the pilot effort.

11
00:03:50.830 --> 00:03:52.190
Katie Antypas (NSF): What do I have, Katie?

12
00:03:52.200 --> 00:04:17.150
Katie Antypas (NSF): Great? Thank you so much, Al: and welcome everyone. It is incredibly exciting to to be able to talk to you. It looks like we have more than five hundred people that are are joining, which is is just amazing. It's been fantastic to see the the the interest from the community on on the near pilot. So now let's go to the the next slide here.

13
00:04:17.350 --> 00:04:47.340
Katie Antypas (NSF): So I wanted to start out with the vision for the Mayor. The nearest stands for the National Ai research resource, and and the vision for the Nara is really to provide a ah, a national research infrastructure that connects the research community to the computing, the data, the training models software and user support necessary to empower and drive Ai innovation

14
00:04:47.350 --> 00:04:52.580
Katie Antypas (NSF): in in the Us. Ecosystem here.

15
00:04:52.590 --> 00:05:19.060
Katie Antypas (NSF): So there are a number of key goals that were identified as part of the near one is to spur innovation, for ah, for the country increase the diversity of talent in ai improve overall capacity in ai r and d, and to advance what we're calling trustworthy Ai or responsible Ai, and we'll talk a little bit more about that as we as we move forward.

16
00:05:21.620 --> 00:05:43.429
Katie Antypas (NSF): You know I'm sometimes asked. You know why? Why do we need an error? We We have big big compute, and and and already in this country, Don't we? But I think The The key challenge here is that there are so many members of the community that have the ability to contribute to the Ai innovation ecosystem,

17
00:05:43.440 --> 00:06:10.760
Katie Antypas (NSF): and we know that Ai holds incredible potential to solve some of our key societal problems. But many, many researchers and educators lack the resources that are necessary even to begin participating in in this ecosystem. And so we really need to make sure that researchers that are investing in investigating it in Ai or using Ai to really serve the public good,

18
00:06:10.770 --> 00:06:28.180
Katie Antypas (NSF): that they have access to to resources. And we need to make sure that the next generation of Ai Ai leaders, that they have access to the resources in their own own classrooms, and ah! To be able to to participate.

19
00:06:31.460 --> 00:06:40.290
Katie Antypas (NSF): So I wanted to provide a little bit of history on where the concept of the near came from, and how we got to the stage where we are today.

20
00:06:40.410 --> 00:07:09.810
Katie Antypas (NSF): So in twenty twenty the Ai initiative act created a congressionally mandated mayor task force. That task force had twelve members, for from academia or from government and or from industry. This task force worked together for eighteen months, and they were in charge with examining whether this concept of veneer was was feasible, and you know, if it's, and if so, you know, to provide an implementation plan

21
00:07:09.910 --> 00:07:22.859
Katie Antypas (NSF): for um for the near. And so they worked um and published a report. Ah, this purple report, you see, in the upper right, and it was published almost exactly a year ago in January of twenty, twenty, three

22
00:07:23.680 --> 00:07:38.080
Katie Antypas (NSF): um in spring of last year an interagency working group started um collaborating to determine. You know how to um begin working in this area, how we might have a pilot for the near,

23
00:07:38.090 --> 00:08:06.240
Katie Antypas (NSF): and it was really the Ai executive order that was released at the end of October, and that was kind of a driving force in moving this forward quickly, the Ai. Ah, the trust! Excuse me, Executive order directed, and Nsf. And our collaborating agencies to launch the near pilot within ninety days, and so we launched the near pilot on January twenty fourth, and you'll hear more about that from

24
00:08:06.250 --> 00:08:07.870
Katie Antypas (NSF): from our team here.

25
00:08:10.700 --> 00:08:38.229
Katie Antypas (NSF): So the in the vision for the ner is that the research community that includes researchers, educators, and students. They gain access to a set of distributed resources through a a near portal, and those resources include um ah cloud computing high-performance computing data, sets, models, software and other tools.

26
00:08:38.409 --> 00:08:43.719
Katie Antypas (NSF): And this is an an image that's taken directly from the narrow task force record.

27
00:08:43.799 --> 00:09:03.059
Katie Antypas (NSF): I think it's important to show everyone here at the scale of um. The vision for the near task force, you know, proposes a program. It's a large program, two point six billion dollars over a six-year period.

28
00:09:03.100 --> 00:09:18.690
Katie Antypas (NSF): Now no, this is this is in the report. This is part of the task force. Ah, um! Ah report no funding has yet been appropriated for the near. But just this is to give you a ah a sense of the scope of the narrative.

29
00:09:18.700 --> 00:09:23.210
Katie Antypas (NSF): Now that's not where we are today. Today we are

30
00:09:23.220 --> 00:09:51.820
Katie Antypas (NSF): ah in the pilot stage of the Narrow. And so our goals with this pilot are really to demonstrate the value and impact of of this broader, near concept, and to support and well to do that, we want to support really novel and transformative Ai research, and also reach new communities, communities across our country in every state and at every type of institution, and really gain experience for building the out the full mayor.

31
00:09:51.830 --> 00:10:21.620
Katie Antypas (NSF): So we are directed in the Ai executive order to launch the near pilot with our existing resources, Um, and and the existing resources of some of our our partner agencies as Well, We also have in kind of contributions from a number of non-governmental partners, including industry, private industry as well as um and so this is a a best effort. Um proof of concept approach that really leverages

32
00:10:21.630 --> 00:10:28.999
Katie Antypas (NSF): what agencies and industry and nonprofits are bringing together to demonstrate this concept of thin air.

33
00:10:29.560 --> 00:10:35.290
Katie Antypas (NSF): So Al I think i'm passing it back to you and the team for the next set of the slides.

34
00:10:35.480 --> 00:10:52.330
Alejandro Suarez: Thank you, Katie, and I forgot to introduce myself. My name is Alejandra Suarez. I'm. A program director within the office advanced cyber infrastructure and I co-chair with my colleagues, Sharon, who you'll hear from shortly the pilot program management office for the near pilot.

35
00:10:52.340 --> 00:11:02.480
Alejandro Suarez: So I want to go over what exactly was launched on January the twenty fourth, and what is available for researchers to access as part of the pilot today.

36
00:11:02.490 --> 00:11:26.469
Alejandro Suarez: So on the left you will see a link to the first call for access to computing resources, including gpus that is available to researchers and educators. This call asks for a short three page proposals. It is available now and is open through March the first, so you can receive access to some resources that i'll go in just a few slides.

37
00:11:26.480 --> 00:11:53.029
Alejandro Suarez: There is also an initial catalog of some ah data sets. Ah, other Ah! Pre-trained models, some research research platforms that are having contributed from some of our agency partners. We have ten agency partners that are available on this this pilot Portal page, and we have twenty five nongovernmental partners. That will be. Ah, we'll be adding those resources to this catalogue, coming soon, so

38
00:11:53.160 --> 00:12:22.350
Alejandro Suarez: to reiterate the access to some of the computing and test that systems from Doe and Nsf. Funded projects is available now. So these select data sets from agency partners is available. Now we are happy to announce that in a late March timeframe, pending some funding and system availability, we'll have a second open opportunity to access additional resources, a wider range of capabilities, and we'll get into some detail about what those capabilities will look like for.

39
00:12:24.830 --> 00:12:54.320
Alejandro Suarez: So this is a large partnership. This is not just an Nsf. Activity. We have eleven agencies, totals that are participating across several areas of government, and we have many contributors from the private sector in terms of ah industry, individuals, industry, companies, nonprofits, foundations that are contributing in kind resources, and there are more that we will be joining as the pilot.

40
00:12:54.330 --> 00:13:06.930
Alejandro Suarez: It proceeds. You can find more details about these contributions at the link in the top right of your screen, and it is available in the Nsf. News site that was linked earlier in the chat.

41
00:13:08.620 --> 00:13:36.599
Alejandro Suarez: So what are the types of contributions that are being made here? Well, initially, there are many different kinds of contributions have been announced. These include cloud computing credits, access to associated models, access to computing hardware systems, test beds, licenses to different software. Platforms, open large language models, software libraries, privacy enhancing platforms, Api access education platforms, enhanced training

42
00:13:36.610 --> 00:13:47.420
Alejandro Suarez: user support types, experiences. These are all the types of contributions, many of them coming from the non-government sector that we are going to be integrating into the pilot over time

43
00:13:49.640 --> 00:14:19.609
Alejandro Suarez: in terms of the organization of the pilot we have essentially a a scaled down model of what Katie presented earlier. That was in the full task force report we will have. We have already launched this near pilot Org ah Portal, that us-based researchers, educators, and students can can come in and see what types of resources are available, and we are updating that in conjunction with the authorities that are putting that website together for us, as well as working with the resources and opportunities that are going to be made available

44
00:14:19.620 --> 00:14:41.200
Alejandro Suarez: through that pilot as a reminder. The idea of the near pilot is to provide the research resources to provide the infrastructure, to allow scientists that are researching Ai or using Ai for their research to get it done. This program is not meant to fund and use a research in and of itself

45
00:14:42.810 --> 00:15:12.729
Alejandro Suarez: in terms of all of the different activities that can be enabled by the Mayor that we are focusing on over the course of the pilot. We have categorised these into four. What we are calling tent poles. Those are in near open which is fully open access to Ai resources near secure, which is being led in collaboration with department of energy and Nih to look at privacy and security preserving resources that may need special treatment due to the effect of the models you're generating the data

46
00:15:12.740 --> 00:15:37.370
Alejandro Suarez: you are using near software, which are how the software stack for an air would look like, What are the different types of platforms, tools, and services that are necessary for researchers to best get their research done and their classroom, which will focus on the reaching new communities, reaching new generations through education, training and user support and outreach materials

47
00:15:38.790 --> 00:15:56.320
Alejandro Suarez: in terms of near-pilot users. We are very interested in Ai researchers that may need infrastructure to get their research done. Domain Scientists that are applying Ai into their research domains, as well as students and educators that are looking to educate future generations on

48
00:15:56.330 --> 00:16:15.069
Alejandro Suarez: how to use Ai tools throughout the education throughout their education in terms of the types of individuals that can access the air pilot resources. We're looking for folks that are from us based institutions. But this need not only be academic institutions

49
00:16:15.080 --> 00:16:35.839
Alejandro Suarez: if you're affiliated with a nonprofit and Ffrdc. State, local or tribal agencies, or even startups and small businesses that may have small business innovation. Research grants, or other types of federal grants are welcome to apply. The main focus, however, is on individuals that are affiliated with us based institutions

50
00:16:37.950 --> 00:16:57.889
Alejandro Suarez: in terms of the initial priority areas that we are focusing on in this. In this pilot at the moment we are looking at these four general themes. These include safe, secure, and trustworthy. Ai includes areas of health care areas of environment and infrastructure and Ai education.

51
00:16:57.900 --> 00:17:18.199
Alejandro Suarez: Note that these are not necessarily and sf research priorities. But these reflect the research priorities of having a government-wide scale resource and these are focus themes but that doesn't necessarily mean that These are the only things that you can do research on when having access to the the Mayor pilot.

52
00:17:18.849 --> 00:17:30.409
Alejandro Suarez: So this is an overview of what's available. Now, I'm going to hand it over to my colleague, Sharon Jiva, who will talk about what we are moving towards in future phases of the pilot.

53
00:17:31.720 --> 00:17:46.890
Sharon Geva: Thank you. Al Um. My name is Shrian Brody Geva. I am also a program director in Oac, and along with Al um Coli, the pilot near pilot Program Management office. So as we move

54
00:17:46.900 --> 00:18:11.919
Sharon Geva: beyond this initial launch we have a few areas that we are focusing on specifically, and you'll see a bit more information about them in the next slides, and one of the important things is governance advisory groups and technical working groups. To be able to create this this larger scale of pilot than the one that we initially launched with

55
00:18:11.930 --> 00:18:16.750
Sharon Geva: another important part is integrating additional partner resources.

56
00:18:16.910 --> 00:18:35.769
Sharon Geva: You'll see in a minute. We had a a small number of original um initial launch resources that we integrated, and we will be continuing to integrate new ones, all contributed by our agency colleagues or by um industry

57
00:18:35.870 --> 00:18:47.510
Sharon Geva: of partners. Something that we we're starting. Working on specifically now is the user support and the outreach to new communities.

58
00:18:47.520 --> 00:19:05.560
Sharon Geva: And um, very soon we're going to be starting um community meetings and technical workshops. So that we can assure that this is a community-designed project and that we can get input from various various channels from the community itself.

59
00:19:07.650 --> 00:19:24.750
Sharon Geva: So the current computing resources that are available and were made available on January the twenty fourth in the initial launch are a number of large systems from the Us. The primitive energy, and from Nsf. Itself.

60
00:19:24.760 --> 00:19:40.279
Sharon Geva: These are names that many of you are probably familiar with frontier alone. Star six from tech. It, has a large Cpu Gpu resources delta from Ncsa with gpu-focused

61
00:19:40.570 --> 00:19:59.170
Sharon Geva: uh system neocortex, which is a test bed with wafer scale ai accelerators Summit the Doe Large capacity Cpu Gpu System and the Argon National Lab Ai test bed that has various Ai accelerators on it.

62
00:19:59.180 --> 00:20:04.469
Sharon Geva: Coming in late March we will be opening up a call for additional

63
00:20:04.650 --> 00:20:13.239
Sharon Geva: resources that are computational, more advanced computing systems than we launched with initially,

64
00:20:13.360 --> 00:20:23.389
Sharon Geva: but also access to cloud resources and to train models, community platforms and collaboration opportunities. You can see a list of

65
00:20:23.920 --> 00:20:36.330
Sharon Geva: our partners and contributions. If you can scan this Qr code or go to this link here. Um, And of course everything is pending system and funding availability.

66
00:20:40.270 --> 00:20:53.350
Sharon Geva: So how how do we see the scaling of the narrow pilot at this initial stage, but the pilot launched on January the twenty, fourth, two thousand and twenty four. We had a pilot portal that

67
00:20:53.360 --> 00:21:21.220
Sharon Geva: when live we had a an initial call for the access to those systems that you just saw in the previous slide. And um it it comes with a uh, a thematic focus of um responsible and trustworthy. Ai research. Um, not not a broad um thematic focus. So, as al noted, we we will be also looking at requests for for broader.

68
00:21:21.230 --> 00:21:22.390
Sharon Geva: Oh,

69
00:21:22.470 --> 00:21:32.349
Sharon Geva: allocations! Then these thematic focuses! We have some initial datasets that are available from agencies and partners,

70
00:21:32.360 --> 00:21:50.000
Sharon Geva: and we also launched a request for information survey of researcher and educator use cases that you can also see on the near-pilot org um homepage and there is a link there, and you can go directly to the survey.

71
00:21:50.010 --> 00:22:00.509
Sharon Geva: We expect to support about twenty five to fifty projects in this first initial launch that will continue until the next call,

72
00:22:01.250 --> 00:22:21.369
Sharon Geva: which will be the enhanced pilot in the spring of two thousand and twenty-four, and which will add access to additional computing resources. We'll have future calls on targeted topics that may not be the same as the as the theme that we had now enhanced user functionality, and we expect to

73
00:22:21.380 --> 00:22:27.080
Sharon Geva: support a much lighter number of projects about one hundred to three hundred.

74
00:22:29.690 --> 00:22:47.819
Sharon Geva: One of the the important things that we are we are um starting to to build out now is support for narrow pilot users and the outreach to new communities. These are things that that are important. They're called out in the executive order, and they're They're critical to be able to

75
00:22:47.830 --> 00:22:55.209
Sharon Geva: build out the support and the community. In the Nation's near-pilot users

76
00:22:55.220 --> 00:23:08.820
Sharon Geva: we have to be able to reach out to new communities. We we are working through various channels to do that consortia, state and regional networks, and so on. We are also starting to one hundred and one

77
00:23:08.830 --> 00:23:28.309
Sharon Geva: work on um a classroom resource. You saw that in in one of the slides at the beginning in our classroom is is one of these areas that that we are targeting for this pilot, and we especially would like to target rural and underserved institutions.

78
00:23:28.320 --> 00:23:47.450
Sharon Geva: And then continuing our collaboration with with our various agency partners. Um, we we are trying to um make sure that we can consult with other agency projects, and that we can build some crosswalks and some um. Some ways of combining

79
00:23:47.460 --> 00:23:51.380
Sharon Geva: these various contributed resources

80
00:23:54.290 --> 00:23:55.860
Sharon Geva: and um

81
00:23:56.220 --> 00:24:01.110
Sharon Geva: and I'll. I'll pass this on to my colleague, Lisa.

82
00:24:02.830 --> 00:24:09.490
Lisa Mirel: Great. Thank you, Sharon. My name is Lisa Merle, and I am the statistical adviser at the

83
00:24:09.500 --> 00:24:14.650
National Center for Science and Engineering statistics which is within Msf.

84
00:24:14.710 --> 00:24:29.130
Lisa Mirel: And i'm really grateful for Alejandro and the neare team for inviting me to be part of this Webinar, and really hear about all of this exciting work, and really kind of talk about ways of sort of talking about agency projects, and how we might be able to align some of the work that we're doing

85
00:24:29.140 --> 00:24:38.579
Lisa Mirel: so on there. You know the pilot is taking off at Ncscs we are actually conducting um what's called the National Secure Data Service Demonstration Project,

86
00:24:38.590 --> 00:24:43.990
Lisa Mirel: And I think that there really is a lot of work between the two projects that can complement each other.

87
00:24:44.000 --> 00:24:52.080
Lisa Mirel: And within our agency we are starting to discuss ways that we can connect these initiatives and really help synergize our experiences.

88
00:24:52.090 --> 00:25:02.459
Lisa Mirel: So just a very, you know. Brief overview um The demonstration work that we are doing for the National Secure Demonstration project. Um! It was established through the Trips and Science Act,

89
00:25:02.470 --> 00:25:20.810
Lisa Mirel: and we've been working on these projects for about two years, and we have many more projects that are continuing to move forward, and there is some connection in terms of. We have projects looking at privacy, preserving technologies, and we actually have a new request for solution that is out right now looking to build a secure compute environment.

90
00:25:20.820 --> 00:25:33.320
Lisa Mirel: So just the last bill that we just wanted to mention that we are having an upcoming webinar where we're going to talk about just different ideas for new projects, and we do have a specific breakout session that really is going to focus on Ai.

91
00:25:33.330 --> 00:25:39.960
Lisa Mirel: So that was kind of all I wanted to mention, but just really looking forward to future collaborations. And I will now turn it back to Sharon.

92
00:25:42.110 --> 00:26:01.669
Sharon Geva: Thanks, Lisa. Um! I just wanted to to quickly show that we we are Um, only a couple of people here on this Webinar, but but we have a whole team. Um, who is working on this near pilot from Oc. From a cross-s and A,

93
00:26:01.680 --> 00:26:12.930
Sharon Geva: and in addition to those that you can see here. There are also many others, from from from our partner agencies, and from the community itself, that are all working on this.

94
00:26:14.070 --> 00:26:18.120
Sharon Geva: So at this point, I think i'll turn it back to Al.

95
00:26:18.280 --> 00:26:19.989
Alejandro Suarez: Thank you, Sharon.

96
00:26:20.000 --> 00:26:49.289
Alejandro Suarez: So this is essentially everything we want it to cover right now in the in terms of what is available to researchers as part of the pilot effort. And what's coming very soon. So stay tuned to the near pilot org space for the future call happening in late march. As we transition to answering more of your questions, we'd like to linger on this slide a little bit, because this is something that we really like to get your feedback on. We noted it earlier in the deck. We're just repeating it. Here we have a

97
00:26:49.300 --> 00:26:53.589
Alejandro Suarez: Nsf: dear colleague letter that calls for

98
00:26:53.600 --> 00:27:22.689
Alejandro Suarez: Yeah. The members of the community researchers, educators, even students, may fill this out a quick survey where you can describe what you know, discipline you're working in. What are the use cases that you can see for using an air? And what are the challenges that you're You're facing from accessing Ai tools and resources for your current research. This is going to be incredibly helpful for us to be informed about how the near pilot

99
00:27:22.700 --> 00:27:52.089
Alejandro Suarez: we'll continue to progress. We also have an air pilot announcements, mailing list. So if you send an email to the email address at the bottom there that near pile of announcements subscribe request, we will be able to keep you informed about updates such as the new Ah, call that the light will likely happen in in late March, and other announcements about the Mayor pilot. So at this point we see many questions coming in many questions that have been answered.

100
00:27:52.100 --> 00:28:01.790
Alejandro Suarez: We are happy to start going through some of your questions, and answering. Some of them live, as many of them have been answered through through text.

101
00:28:04.420 --> 00:28:06.520
Alejandro Suarez: And, Katie, please go ahead.

102
00:28:06.530 --> 00:28:20.229
Katie Antypas (NSF): Sure I um. I saw one question, and I want to make sure we really emphasize that this um the nearer isn't is an infrastructure. It's an infrastructure for the research, community,

103
00:28:20.300 --> 00:28:36.380
Katie Antypas (NSF): and I make an example. This is an infrastructure, you know. Not a telescope, but it's it's like a telescope in that it is made available to the community. The Mayor pilot is not funding researchers own

104
00:28:36.390 --> 00:28:47.230
Katie Antypas (NSF): time or or students. It's providing access to this to this resource. And so there is a question. The

105
00:28:47.240 --> 00:29:10.650
Katie Antypas (NSF): the process for requesting resources. It's actually not um an Nsf process it's run through, run through one of our awardees, and it's much lighter weight. So I think it's a a couple page apple uh application process. Um! And so there was a question about um a bio sketch, and you'll see in this application.

106
00:29:10.660 --> 00:29:23.750
Katie Antypas (NSF): Ah, that is that that is online, that that particular part is not needed. They will check to see if you have other funding when you apply for those resources.

107
00:29:25.870 --> 00:29:27.370
Alejandro Suarez: Thank you, Katie

108
00:29:28.800 --> 00:29:29.920
Alejandro Suarez: Sharon,

109
00:29:30.450 --> 00:29:49.849
Sharon Geva: I can answer. I saw a few questions about access, and overlap with the near pilot. So it is true that a number of the resources, even in the initial launch, are resources that you can request allocations on

110
00:29:49.860 --> 00:29:52.079
Sharon Geva: through access as well.

111
00:29:52.090 --> 00:30:10.330
Sharon Geva: But a few, a few important things to note near pilot is not taking any of the capacity of these systems that that is allocated to access. So this is in addition to the capacity that is allocated

112
00:30:10.420 --> 00:30:25.810
Sharon Geva: to access via access um from these systems. And the question was, What's the difference? So there, there's a couple of differences that you might want to take into account if you consider applying through through near

113
00:30:25.820 --> 00:30:40.519
Sharon Geva: um. One of them is the the near call, at least the initial call is is for a six-month period of of usage. And the other very important thing is the focus of the projects

114
00:30:40.530 --> 00:31:00.090
Sharon Geva: that that will be awarded time on these systems and the focus here has to be completely aligned with with the um nearer, and goals, and with the thematic focus of the specific call. So there there. There is no

115
00:31:00.100 --> 00:31:26.230
Sharon Geva: no reason to choose one or the other you you can apply to to more than one. But there in mind that that there are specific eligibility requirements for for applying for an allocation on there, even if it's on a system that you are used to requesting on access, and along with the eligibility there. There's also the time constraint on the on the length,

116
00:31:26.240 --> 00:31:32.769
Sharon Geva: allocation, and and the thematic focus of the project that you are proposing.

117
00:31:35.570 --> 00:31:54.659
Alejandro Suarez: Thank you, Sharon. Ah, I will note I've seen some questions that talk about are there? Ah! Subgroups on fair principles. Aspects. How is Ip being handled when resources are being used, I will

118
00:31:54.670 --> 00:32:22.949
Alejandro Suarez: point out with regards to the initial allocations call. There is a requirement that anyone who uses the near pilot computing resources in this. Ah! This round has to openly publish their findings of of their research. So any if you're doing something that results are going to be proprietarily handled due to the institution you're in. That. Would that would not be allowable through the near pilot,

119
00:32:22.960 --> 00:32:36.050
Alejandro Suarez: however, Otherwise, just like as Sharon was talking about our access program, which is our general program for that Nsf. Runs for access to computing resources for scientific research. More broadly,

120
00:32:36.100 --> 00:32:50.239
Alejandro Suarez: we would give the same advice as Ah, that program, which is things like handling Ip, that is, up to your institution. How you do that through your your standard research projects, such as if you are receiving a Federal grant.

121
00:32:51.230 --> 00:32:52.640
Alejandro Suarez: Katie, please go ahead.

122
00:32:54.090 --> 00:33:09.030
Katie Antypas (NSF): It looks like there's a couple questions related to private data, secure data, and so I want to speak a little bit more about the near near secure thrust, and they are secure.

123
00:33:09.040 --> 00:33:19.270
Katie Antypas (NSF): Area is going to be led by Nih and Doe, and so we we do want to make the

124
00:33:19.430 --> 00:33:30.870
Katie Antypas (NSF): opportunities available through near secure. We realize that there are certain Ai research that requires an additional level of

125
00:33:30.880 --> 00:33:42.180
Katie Antypas (NSF): ah privacy and security preserving functionality. And so I would say, stay, stay tuned for more information on on their, so on their secure.

126
00:33:43.180 --> 00:33:44.440
Alejandro Suarez: Thank you, Kerry.

127
00:33:49.100 --> 00:33:50.380
Alejandro Suarez: See?

128
00:33:50.560 --> 00:33:52.990
Alejandro Suarez: Ah, Bill, see a question. You'd like to.

129
00:33:53.000 --> 00:34:01.439
Bill Miller/NSF: Yeah, thanks. There was a question about supplements to Awards research awards, and I just thought I would answer

130
00:34:01.670 --> 00:34:09.110
Bill Miller/NSF: easier to answer it by voice. So everybody hears again it's. It goes back to the question about the near pilot

131
00:34:09.170 --> 00:34:15.730
Bill Miller/NSF: provisioning resources for the for the researchers. We're not supporting the research itself,

132
00:34:15.920 --> 00:34:27.350
Bill Miller/NSF: and these resources are not being provided at at a cost. So if you get an allocation on computing and near-pilot computing resources, it's not going to be charged to your grant or anything.

133
00:34:27.580 --> 00:34:47.439
Bill Miller/NSF: Ah, so if you are interested, as this questioner is interested in supplements to your research grants, we encourage you to reach out directly to your research. Ah, program officer, but you know for your award, and that you can discuss research supplements, but it's not connected to the narrative

134
00:34:47.449 --> 00:34:51.700
Bill Miller/NSF: pilot infrastructure uh effort that we're doing here.

135
00:34:53.510 --> 00:34:54.720
Alejandro Suarez: Thank you, Bill.

136
00:35:01.310 --> 00:35:02.439
Alejandro Suarez: I'm sure

137
00:35:03.980 --> 00:35:12.130
Sharon Geva: there was a question about, if you can. If there's a mechanism to do exploration without applying

138
00:35:12.490 --> 00:35:28.709
Sharon Geva: for for access to these to these resources. So if these are computational resources, then there is no mechanism it has to other than going through a request for access through the application.

139
00:35:28.720 --> 00:35:45.489
Sharon Geva: There are, however, many data sets that you can see on the website listed on the website that you can. You can access without having to get an allocation formally approved.

140
00:35:45.880 --> 00:36:00.030
Sharon Geva: And as we go as we go to the next phase you will see many different types of of resources. Some of them will be available only through allocation. Some of them will be available only,

141
00:36:00.040 --> 00:36:05.930
Sharon Geva: or will be available without having to go through the allocation list, but they're going to be

142
00:36:06.420 --> 00:36:16.829
Sharon Geva: clearly delineated. Which of them are require allocation requests, and which aren't, and you can see that today on the website, the call for allocations

143
00:36:16.840 --> 00:36:31.629
Bill Miller/NSF: is is for resources that that require going through the process and the the open data sets and so on. You do not have to fill out forms or or go through any process for them.

144
00:36:38.400 --> 00:36:57.859
Bill Miller/NSF: Another question, and I see Ah, Barone is also typing a response which I will let him do. But I just want It's It's a question that has broad relevance from an anonymous attendee who is asking about for students or newer researchers. Those not familiar with access or the allocation process. The bar for applying is is high,

145
00:36:57.870 --> 00:37:06.200
Bill Miller/NSF: and it would be great if there were a much more lightweight allocation process for the initial step of allocation. So onboarding to the near can be easier. So

146
00:37:06.510 --> 00:37:29.769
Bill Miller/NSF: this is very important to us to reach. Ah, it's important to access as well the access program, but for the near pilot it's also very important to reach broadly, and we are, We encourage Ah! New users and and students. Ah! Meritorious activities that they would like to do to to apply. Right now we have the one first

147
00:37:29.780 --> 00:37:45.950
Bill Miller/NSF: an early open call, and it's well described what resources are available there. It's a relatively simple form to fill out with a relatively short, I think, three-page application, and so it's not overly onerous.

148
00:37:45.970 --> 00:38:01.270
Bill Miller/NSF: We are looking at in the future, and the next open call there will be a more broad array of resources and and different ways of engaging, but we encourage you to apply, because we want

149
00:38:01.280 --> 00:38:14.190
Bill Miller/NSF: mit Ctl, and to try to make sure that your experience is good. So don't be thrown off or or hesitant to do so. Do your best to describe what it is that you're trying to do. One hundred and fifty.

150
00:38:14.200 --> 00:38:20.140
Bill Miller/NSF: Answer the questions, and we will. It will be addressed accordingly.

151
00:38:21.970 --> 00:38:26.399
Alejandro Suarez: Thank you, Bill. I see another question that is asked.

152
00:38:26.410 --> 00:38:49.709
Alejandro Suarez: There are already research infrastructures in many other areas like health, automotive security include elements of Ai. How do we ensure coordination and consistency? Across all these efforts, and with fragmentation, one of the ah strong points for the near pilot effort, as well as what was called for in the full. There is that we have a steering committee of all of our agency partners that are

153
00:38:49.740 --> 00:39:12.570
Alejandro Suarez: bringing up what are the different activities within Ai, the infrastructures within Ai, that we are all funding and understanding how to best work together to ensure that there is coordination between all of those different resources that are going to be available. In addition, we are continuing our conversations with the non government contributors to ensure that the

154
00:39:12.580 --> 00:39:39.429
Alejandro Suarez: resources that have available have the right match to the research use cases of the members of the research community that are going to be using them, and it connects to emissions of the different agencies that are going to be involved in the near-pilot effort. And So All of this we're looking at what is available, and looking to best synergize the different kinds of contributions that are coming from all our all our contributors. In this

155
00:39:44.170 --> 00:39:49.779
Alejandro Suarez: I'll just keep writing through reading through some of these

156
00:39:49.790 --> 00:40:18.209
Alejandro Suarez: ah questions someone had asked, Is there a quarter limitation for the the resources available? Is academic personal use acceptable right now, when it comes to the initial call. Um! They are going to be evaluated for the the scientifical and technical appropriateness and the feasibility of using that specific resource. So you have to have a research use case, and you have to have a reason to

157
00:40:18.220 --> 00:40:36.549
Alejandro Suarez: the type of resource that you are requesting. So with regards to whether it's a personal project, or if it's your academic research project, it has to have a research use case that you are stating in your application to request access to these resources.

158
00:40:45.460 --> 00:40:47.319
Alejandro Suarez: Seeing other

159
00:40:47.370 --> 00:40:49.259
Alejandro Suarez: activities here,

160
00:40:52.200 --> 00:41:06.659
Alejandro Suarez: someone had asked, Are there any groups involved in environmental projects? Not just academic research, but real world application? Someone's addressing clean water issues in Florida,

161
00:41:06.670 --> 00:41:34.230
Alejandro Suarez: Marlin. I don't know if I don't want to put you on the spot here, but there are some ah additional projects that I think are Ah! That that have good synergy with this in terms of the near pilot. There'll be more coming soon, I think, in that March call, especially when it comes to environmental impact. That was one of the priority thrust areas that we talked about in the thematic slide earlier.

162
00:41:34.240 --> 00:41:38.790
Alejandro Suarez: I don't know if there's anything else you wish to add on the environmental aspect to Marlon.

163
00:41:42.070 --> 00:41:52.389
Marlon Pierce: Yeah, thanks. It was also for supporting climate research. But the question specifically, if we're talking about the same one was,

164
00:41:52.400 --> 00:41:54.509
Marlon Pierce: I think, wondering about

165
00:41:54.620 --> 00:41:57.640
Marlon Pierce: if you're a non-academic

166
00:41:57.670 --> 00:42:08.519
Marlon Pierce: if you're not academic. But you're involved in these sorts of research projects or environmental projects. Are you eligible for the getting access to the air at this time?

167
00:42:10.420 --> 00:42:27.770
Alejandro Suarez: Thanks, Marlin. I think I would go back to the types of users that we are targeting in this call, And I think it's not just academic research organizations. If you have a use case that has to do with.

168
00:42:27.780 --> 00:42:56.740
Alejandro Suarez: There. There is some research involved, but you may not be in an academic institution if you're at a a nonprofit or a State agency that is looking to do a study on something such as the clean water issue that was discussed in the in the Q. And A. Then that could be something that you could request access to resources through the initial allocation. You just have to discuss that in the form of a research question that is being addressed through the help of access to this infrastructure.

169
00:43:03.820 --> 00:43:05.399
Alejandro Suarez: All right,

170
00:43:09.070 --> 00:43:10.660
Alejandro Suarez: let's see

171
00:43:10.730 --> 00:43:14.949
Alejandro Suarez: other. I see some folks are asking the

172
00:43:15.640 --> 00:43:20.269
Alejandro Suarez: there specific requirements for reporting at the end of the pilot. So the

173
00:43:20.280 --> 00:43:49.679
Alejandro Suarez: I think there's a couple of different ways to interpret that question. There are reporting requirements for initial users of the pilot, and we will that that is available in the allocations call. You can see some more detail about what is asked of the researchers that use those resources. We will continue giving updates about what capabilities are made available through the pilot, as we as we continue, and we'll be getting more feedback from the community as well as the number of users that

174
00:43:49.690 --> 00:43:51.060
we bring on.

175
00:43:51.650 --> 00:43:53.189
Katie. I see you have.

176
00:43:53.200 --> 00:44:12.149
Katie Antypas (NSF): Yeah, a couple of things. So there's a number of questions on um. I feel like near classroom and from educators, and I just want to follow up on what Bill said that. Um, this we call it open call, or this open opportunity for access is really targeting the near the near open thrust,

177
00:44:12.160 --> 00:44:22.479
Katie Antypas (NSF): please, I I would say, please stay tuned for near classroom, so we certainly do not want the boundaries to fall like outside of your

178
00:44:22.490 --> 00:44:52.230
Katie Antypas (NSF): you. You know a a semester that's not the an intention at all. We understand the near classroom effort. Um probably needs to have a lower barrier to access, and also probably needs to come in a different way. And so we are. You know, We just launched this pilot about three weeks ago near classroom. We're still working out some of the details there, but I appreciate a lot of these comments about simplicity for for educators, um, you know, and and so just us.

179
00:44:52.240 --> 00:45:03.010
Katie Antypas (NSF): They stay tuned for more information on their classroom, and how to get those resources and apply for those resources for your community.

180
00:45:03.020 --> 00:45:18.310
Katie Antypas (NSF): I saw one other question sort of around Max, and what we plan to to publish. Certainly, once we get all of the use cases in. We will

181
00:45:18.320 --> 00:45:33.719
Katie Antypas (NSF): certainly have proud presentations and public material on those survey results that was going to be really important to get that community feedback on that we also will be um sharing.

182
00:45:34.220 --> 00:45:46.490
Katie Antypas (NSF): You know information about the the number of projects that have been included. Certainly kind of the demographics of those of those. You know the the science areas as well.

183
00:45:48.190 --> 00:46:05.119
Katie Antypas (NSF): But someone else asks, Has anything like this been done before? I'm going through kind of a different list here. I just want to note that many of you may have been familiar with the Covid, nineteen Hpc. Consortium consortia that brought together

184
00:46:05.130 --> 00:46:10.210
industry, academia, and government resources to address

185
00:46:10.220 --> 00:46:38.969
Katie Antypas (NSF): computational science. Research is so to associated with the pandemic, and in many ways I would say, Um, we are informed by by that process I think we're We're building off of it. The The near pilot has is integrating more than just compute, and so it's. It's more complex. It's a larger effort. It's more for doing it more formally. But that is a a recent effort that we've worked with

186
00:46:38.980 --> 00:46:41.520
Katie Antypas (NSF): both public and private sector partners.

187
00:46:42.760 --> 00:46:45.189
Katie Antypas (NSF): Thank you, Katy.

188
00:46:45.670 --> 00:46:48.220
Alejandro Suarez: Ah! So that

189
00:46:48.230 --> 00:47:07.910
Alejandro Suarez: there's a few questions that have been asked about timing and time span what is available when for how long? Um! So is the access afforded as part of the pilot indefinitely? Or is it rotated for a period of time? The initial call, which is open through March first is for six months

190
00:47:07.920 --> 00:47:37.370
Alejandro Suarez: allocation. So it's for you to do. Yeah, for piloting efforts. We're looking at short timeframe projects where you will be able to have some results within six months of your allocation in terms of budget. I think Katie and some others may have commented on this before this Right now, is, There's not something that we have any budget in terms of dollars available for researchers. If there are active awards you have that you may reach out to your

191
00:47:37.440 --> 00:47:55.450
Alejandro Suarez: program officer about. You know questions for supplementing that award. If you have additional scope, you, you would like to add to that. But what the near pilot is making available is access to this infrastructure, access to the computing research, research, infrastructure, and eventually

192
00:47:55.460 --> 00:48:13.209
Alejandro Suarez: access to these cloud resources, these models, et cetera. And then there was a discussion about. Are there mechanisms to contribute back, or how how are folks contributing back to the near pilot? And for that, from an individual

193
00:48:13.220 --> 00:48:20.469
Alejandro Suarez: researcher perspective we have. I think I've talked about the reporting requirements that we have. We want to

194
00:48:20.480 --> 00:48:49.990
Alejandro Suarez: ensure that the near-pilot user community is informed about what is, what is available. What are the highlights from the research that is coming out? What are the demonstration projects that have highlighted the unique infrastructure that was made available to them, or combination of infrastructure that was made available that allowed them to make new scientific discoveries. This is all something that we're keeping in mind as we grow the pilot, and as we onboard new users

195
00:48:54.240 --> 00:48:57.899
Alejandro Suarez: and I see many of our questions are:

196
00:48:58.090 --> 00:49:01.530
Alejandro Suarez: i'm. In the closed questions.

197
00:49:03.890 --> 00:49:17.449
Alejandro Suarez: We can continue going down. We discuss the educators and how to approach their allocations. Call. Do stay tuned for the near classroom piece, Katie. I see you your hands up

198
00:49:18.000 --> 00:49:39.309
Katie Antypas (NSF): right. I see there's a couple questions around data sets and data set contributions, although the data sets that we have available now are provided by government agencies. And so we're working through a process for accepting data sets from non-governmental partners and so stay tuned for that.

199
00:49:40.900 --> 00:49:42.279
Alejandro Suarez: Thank you, Katy.

200
00:49:42.490 --> 00:49:43.649
Alejandro Suarez: Ah,

201
00:49:43.940 --> 00:49:46.810
Alejandro Suarez: with regards to the educator call.

202
00:49:46.820 --> 00:50:15.509
Alejandro Suarez: Katie talked about. Their classroom is going to be coming online soon, and there will be more opportunities available through that. If you have a research project that involves working with, you know, serving something to a community, and you feel that that could be ah described in a proposal for access to the existing resources. You're welcome to apply to the near-pilot dot org Call that's currently currently up. If you have any questions about that, you can reach out to the

203
00:50:15.520 --> 00:50:22.160
Alejandro Suarez: the Help folks on the allocations team who may give you a bit additional guidance There,

204
00:50:36.750 --> 00:50:38.010
Alejandro Suarez: Let's see

205
00:50:38.330 --> 00:50:40.479
Alejandro Suarez: other questions.

206
00:50:42.270 --> 00:50:50.530
Alejandro Suarez: There's a question about whether the spring call will be a six-month span as well. That is yet to be determined. How long the

207
00:50:50.540 --> 00:51:07.839
Alejandro Suarez: the allocations awarded through that call will be stay tuned to the near pilot org space and the announcements lists. Once we're ready to officially announce the spring call, we will have more details on what those allocations will look like.

208
00:51:19.810 --> 00:51:27.519
Katie Antypas (NSF): Yes, Katie or I can take one. What is the criteria of success for the pilot? I think this is a great question.

209
00:51:27.530 --> 00:51:56.099
Katie Antypas (NSF): Um um recall what we're really trying to do with this pilot is demonstrate the the potential impact and that demonstrate the the potential for the concept of a full there. And so we want to do this in a couple ways. We want to reach really broad and new communities. So we are going to be, you know. Um tracking the new users, the new researchers that are connected getting connected to research, to to resources.

210
00:51:56.110 --> 00:52:15.549
Katie Antypas (NSF): And then we also want to have some amazing um and novel science that is running and using the pilot resources that you know, we can demonstrate that the the pilot is supporting, You know, Ai research in the public in the public's interest.

211
00:52:18.120 --> 00:52:19.520
Alejandro Suarez: Thank you, Katie.

212
00:52:21.450 --> 00:52:33.480
Alejandro Suarez: I think we discussed the criteria for contributing data sets. Katie noted the connection. The The process that we're working with our non-government entities, we will be able to share more on very soon.

213
00:52:33.870 --> 00:52:48.309
Alejandro Suarez: We've, discussed the how educators may approach this. Ah! The near pilot allocations call it, was a discussion about access by the Ckd. You may be answering this in in chat

214
00:52:48.320 --> 00:52:58.810
Alejandro Suarez: so essentially what we're differentiating Here is there's the access to the resource which you can apply for essentially computer time

215
00:52:58.820 --> 00:53:21.679
Alejandro Suarez: versus the funding of. You know some of your salary support for your graduate students. If there's some equipment that you, your your research team is buying, that's what we're not able to support at this time. So that's why there's no there's no dollar budget that we're awarding for these allocations. If you have a research project that requires you

216
00:53:21.690 --> 00:53:27.619
training a model on a set of gpus for a certain length of time to be able to

217
00:53:28.450 --> 00:53:36.339
Alejandro Suarez: connect to one of your your research projects to make a discovery. That is what you can apply for through the allocations team, and that's what would be awarded that computing time.

218
00:53:40.270 --> 00:53:46.919
Alejandro Suarez: There's discussing about the dollar amount of in-kind contributions

219
00:53:46.930 --> 00:54:15.409
Alejandro Suarez: that would be provided by commercial cloud you can find more information in the There was the Nsf. News link that we shared in the chat a little bit earlier. We can repost that. That talks about all of the different contributions that have been announced so far, and Katie noted that we're in discussions on uh, exactly what will come when, in terms of those allocations, and what form they will take as they get uh shared with the research community itself.

220
00:54:15.840 --> 00:54:16.939
Alejandro Suarez: Built.

221
00:54:17.570 --> 00:54:28.599
Bill Miller/NSF: Yeah, answering a question from Kurt Theorem. Then about looking to other government agencies, websites for their aspirations, or as Nsf. Aggregating and integrating across the Federal landscape.

222
00:54:28.710 --> 00:54:41.150
Bill Miller/NSF: Nsf. Is leading the near pilot on behalf of the other agencies, and we've done it for the and Nsf. Website. We We put it in the link in there earlier. Ah, and the and the portal.

223
00:54:41.160 --> 00:54:54.349
Bill Miller/NSF: All the text that refers to other agencies were worked out with the agencies that said um. Some agencies we, I think several agencies may have ah written on their websites about the near pilot, and have,

224
00:54:54.360 --> 00:55:08.739
Bill Miller/NSF: you know, in alignment with their own missions. So there's nothing to ah not discourage you, or we do encourage you to take a look at other agencies particularly, I think, for example,

225
00:55:08.750 --> 00:55:24.340
Bill Miller/NSF: the we made the their summit. Ah, supercomputer a portion of that ah ah! Available as part of the first open call. That's open right now for computing resources, and I know that they made an announcement about that.

226
00:55:24.350 --> 00:55:34.180
Bill Miller/NSF: I don't remember if other agencies have done something similar, but they they they are perfectly welcome to do that. It's a collaborative and collegial effort.

227
00:55:35.840 --> 00:55:37.019
Alejandro Suarez: Thank you, Bell.

228
00:55:40.040 --> 00:55:41.800
Alejandro Suarez: I think

229
00:55:42.010 --> 00:55:46.760
Alejandro Suarez: there's some more discussions about the six-month period, I would

230
00:55:46.940 --> 00:55:58.140
Alejandro Suarez: say, You can. You can reach out to the allocations team through near Pilot Org sl allocations for more information about the six months period. We can

231
00:55:58.150 --> 00:56:12.650
Alejandro Suarez: chat separately if you Ah, if if folks have questions we can, we can also put some more contact information in the chat. Um, you know, if you have questions specific to your own, your own research case

232
00:56:19.570 --> 00:56:25.789
Alejandro Suarez: as a question about whether compute facilities will be isolated. Oh, i'll go ahead, Katie. I see you have your hand up.

233
00:56:25.800 --> 00:56:27.640
Katie Antypas (NSF): Um. So yeah, I should.

234
00:56:27.770 --> 00:56:40.820
Katie Antypas (NSF): So a question about um. The contributed resources and the scale of those resources. I put a link into the chat, and if you kind of scroll down um, we have

235
00:56:41.090 --> 00:57:09.409
Katie Antypas (NSF): ah provided a a little summary for each of the twenty five partners in terms of their contribution. Some of our partners provided a dollar amount, others others did not. And so you you can. You can go and and and look there. Some of the contributions are right in the tens of of millions of dollars. Other contributions are, say, for

236
00:57:09.780 --> 00:57:21.890
Katie Antypas (NSF): ah access to software that is, has no license for for near-pilot users. Others are a contributions in terms of collaboration. So you can kind of peruse that.

237
00:57:21.900 --> 00:57:23.770
Katie Antypas (NSF): And hopefully, that's helpful.

238
00:57:25.290 --> 00:57:26.629
Alejandro Suarez: Thank you, Kitty.

239
00:57:28.000 --> 00:57:43.179
Alejandro Suarez: I think there was a discussion about whether the computing resources could be accessible individually, or their plans to string them together to form larger compute forms. That's a great question.

240
00:57:43.220 --> 00:57:46.650
Alejandro Suarez: I see, Veroon, you have a note on there, or maybe you're

241
00:57:46.710 --> 00:57:49.090
Alejandro Suarez: tossing it aside to me.

242
00:57:49.100 --> 00:58:17.890
Alejandro Suarez: We we have. We are looking into what are the best ways to make these resources available, and what are the use cases from the research community, including workflows that could allow research to be the time to science to happen as efficiently as possible. This could include existing data sets that are available on certain resources that may be connected to compute resources in a tighter manner,

243
00:58:17.900 --> 00:58:23.770
Alejandro Suarez: to allow for a calculation to be done quicker. If there are ways to

244
00:58:23.780 --> 00:58:53.210
Alejandro Suarez: move a data set from ah! From one system to another, because there are different aspects of the compute that are more beneficial for different aspects of the workflow than best leveraging high-performance networks to connect the resources and transfer data that way are all things that we're exploring in terms of the near-pilot activities? So How how can these machines best work together, and that is, I think, part of a more general concept within Oac, about all of our advanced people,

245
00:58:53.220 --> 00:59:00.779
Alejandro Suarez: resources that we make available to the research community nationally, and how best they can support researchers in their endeavors.

246
00:59:04.900 --> 00:59:33.750
Alejandro Suarez: And I see. So we have very few. Just one question left here. If near pilot success and funding from there is allocated, what will it be used for? Well, there are a lot of use cases that are discussed in the near task Force report. You're welcome to look at what the vision of the large scare scale mail there might look like, at least as described in that report, I will say that there is currently no,

247
00:59:33.820 --> 00:59:47.429
Alejandro Suarez: no, no legislation has passed that that has created an official there, so there's only speculation that we can get into in terms of specifics there. But i'll hand it over to Katie, who may have more thoughts on this.

248
00:59:47.440 --> 00:59:56.469
Katie Antypas (NSF): Sure, Yeah, if you look at the near task Force report. I think that's your slide. Um uh four or five um al um!

249
00:59:57.070 --> 01:00:03.589
Katie Antypas (NSF): The majority of the resources envisioned Go to.

250
01:00:03.850 --> 01:00:12.169
Katie Antypas (NSF): They they call it, you know, resource providers. So thinking of yeah access it was that next one,

251
01:00:12.770 --> 01:00:13.890
Katie Antypas (NSF): okay,

252
01:00:13.900 --> 01:00:20.149
Katie Antypas (NSF): access or larger compute and data resources, I think it's the next one out

253
01:00:21.530 --> 01:00:24.300
Katie Antypas (NSF): there we go. So. Um

254
01:00:24.310 --> 01:00:51.849
Katie Antypas (NSF): I mean. The the vision is that the majority of the the funding would go to. We say, resource providers. Those are providers of compute and data data software assets. There's also sixty to seventy million dollars for what's called an operating entity that would be a non-governmental ah operating entity. That is running the day-to-day operations of

255
01:00:51.860 --> 01:00:58.999
Katie Antypas (NSF): uh of of an air and then a small amount for external evaluation. As well

256
01:01:04.130 --> 01:01:17.910
Katie Antypas (NSF): please keep the questions coming we have. We have a little more time, and happy to to answer them for you. These are great questions that are coming coming in. We already have a sidebar on where we need more more discussion. So thank you all.

257
01:01:19.070 --> 01:01:20.890
Katie Antypas (NSF): Oh, Bridge,

258
01:01:20.900 --> 01:01:48.629
Alejandro Suarez: I think Bill may be typing an answer to this one, but just some clarity on when it comes to the researchers that are eligible for requesting resources on the on the near pilot. We are looking at researchers from us-based institutions. That is different from saying, You know we are only requiring people to be us citizens or things like that. We require everyone who will be using the

259
01:01:48.640 --> 01:02:02.179
Alejandro Suarez: a resource to be affiliated with a us-based institution for usage of the of the system That is the main requirement under the research, security and research, integrity, requirements of the near pilot.

260
01:02:09.290 --> 01:02:11.159
Alejandro Suarez: Any other questions

261
01:02:12.140 --> 01:02:13.869
Bill Miller/NSF: we managed to

262
01:02:14.520 --> 01:02:17.559
Bill Miller/NSF: handle them all. But there's now a new one. Okay?

263
01:02:22.180 --> 01:02:25.560
Alejandro Suarez: So this one is asking, ah

264
01:02:25.810 --> 01:02:40.689
Alejandro Suarez: use of hugging face data sets or synthetically generated data sets be considered, Viables will access to data sets from the pilot, and for synthetic augmentation afforded data sets. I think there's a lot of a lot of ongoing work on exactly what the

265
01:02:40.700 --> 01:02:51.040
Alejandro Suarez: partner contributions will look like if you're looking to partnership partner contributions. If you're using existing data sets that are available out in the open, then

266
01:02:51.050 --> 01:03:19.710
Alejandro Suarez: there there are ways to connect those to some of our computing allocations when it comes to things that have not been that are not publicly ah available, We're going to have to figure that out. The different use cases. The research use cases you have if you're applying for access to some of our existing allocations. But again, with regards to the non-government contributions, what form. They will look like you can find a bit of a sneak peak on the

267
01:03:19.720 --> 01:03:27.400
Alejandro Suarez: the Nsf. News site that we linked earlier. But more details will come in the coming phases starting in March,

268
01:03:45.540 --> 01:03:52.760
Alejandro Suarez: Does the Nara intend to standardize usage of these resources in the longer term requesting access, infrastructure, use, etc.

269
01:03:52.770 --> 01:04:07.300
Alejandro Suarez: I think that's a great question we are also looking at in terms of the pilot. We're piloting allocation mechanisms. What are the best ways to connect researchers with access to these kinds of resources. We realize that

270
01:04:07.310 --> 01:04:28.749
Alejandro Suarez: an allocatable resource, like a supercomputer, may not be the same as giving out a license to use a certain software stack or giving out an Api key to get access to a certain inference model or access to a large language model. These are all areas that we're going to be exploring over the course of the pilot. Um strong if you have something dead.

271
01:04:28.760 --> 01:04:31.600
Sharon Geva: Um, so? And asked um.

272
01:04:31.750 --> 01:04:53.460
Sharon Geva: Someone asked previously also about. How? How do you know what? What resource you should be requesting access to? And I just wanted to point out that in this this current initial call there. You can select from a list of of the systems that are available. But you can also say, I don't know which one I want. Choose for me,

273
01:04:53.470 --> 01:05:10.669
Sharon Geva: and during the review process, and the and the matching process, they these um that these requests go through it. It will be possible for for people who understand the differences between these machines to to basically choose for you.

274
01:05:11.900 --> 01:05:14.919
Sharon Geva: So if you're if you're not sure which one,

275
01:05:15.020 --> 01:05:25.949
Sharon Geva: you should be selecting that shouldn't stop you from requesting um access to to one of these systems that are in the in the call for allocations.

276
01:05:28.880 --> 01:05:30.350
Alejandro Suarez: Thank you. Sharon.

277
01:05:33.910 --> 01:05:35.229
Alejandro Suarez: Ah, Kitty,

278
01:05:35.440 --> 01:05:50.970
Katie Antypas (NSF): Yeah, um, I think I I want to say a little bit more around um their classroom. So the vision for for near classroom is really to support educators who you know, are,

279
01:05:50.980 --> 01:06:11.099
Katie Antypas (NSF): or either just starting their careers or retraining um, and are are new to the field of of Ai, or using Ai. And so What we have heard from from many communities is that, you know, simply

280
01:06:11.110 --> 01:06:38.239
Katie Antypas (NSF): working on doing a a classroom assignment or a classroom project, and and from a machine learning class can be really challenging for many communities, because even those classrooms Don't, have access to the kind of the computing or the data resources even to do those kind of assignments, and so our intention with their classroom is to provide that infrastructure to

281
01:06:38.250 --> 01:06:56.260
Katie Antypas (NSF): to ah education communities. It's not in the near pilot's scope to provide the curricula. And so we're going to be partnering and actually meeting next week with um some other directorates to to see how we may

282
01:06:56.270 --> 01:07:00.489
Katie Antypas (NSF): make these near classroom resources available.

283
01:07:00.500 --> 01:07:07.330
Katie Antypas (NSF): But the answer is, No, it's not intended to provide continuing education credits. Which was the specific question.

284
01:07:08.740 --> 01:07:10.129
Alejandro Suarez: Thank you, Katy.

285
01:07:10.180 --> 01:07:34.799
Alejandro Suarez: I see Another question, and Bill may be adding to this one as well. There's a question about whether an operating entity will be identified during the pilot, or if it will be during a full funding appropriation, what type of organization would it be? Um, right now? I think it's? It's too early to get into the details. On that we will say that we're we're currently working with existing awards existing funds.

286
01:07:34.810 --> 01:07:36.770
Alejandro Suarez: Ah! So this is

287
01:07:36.780 --> 01:08:06.359
Alejandro Suarez: whether it's supplementing some projects to allow additional functionality to hook into near pilots. Some of these resources, for example, some of the the near Pilot Org website, that you've seen the allocations mechanism that you've seen how all of that will be coordinated. We're currently working on the best mechanism to enable that coordination and provide more community involvement in that coordination process as the pilot progresses, and more to come on. What an operating energy may look like.

288
01:08:16.430 --> 01:08:27.420
Alejandro Suarez: I see one last question here: economic sciences. We We have the the overall research themes that are that we pointed out

289
01:08:27.430 --> 01:08:42.079
Alejandro Suarez: on this slide here. However, that's not We're not only allowing that these are just the themes that are priorities at the moment, especially the top one safe, secure, trustworthy Ai. If there's some

290
01:08:42.090 --> 01:08:55.370
Alejandro Suarez: research in your area of economic sciences that can connect to usage of Ai resources, and you'd like to request access to the allocations mechanism. You're welcome to apply for that.

291
01:09:06.450 --> 01:09:10.769
Alejandro Suarez: Still some questions coming in um

292
01:09:10.800 --> 01:09:14.190
Alejandro Suarez: for both research and teaching use cases.

293
01:09:14.200 --> 01:09:42.409
Alejandro Suarez: There's questions about how how many times can you apply for access to to to resources? I don't think we have a specific limit right now, but it is. It does close on the first. So I think at this point you may only have time to to get in one application in, but you stay tuned for the next call. That will be happening soon after, if that may be another opportunity to request access to some of these resources,

294
01:09:44.450 --> 01:09:51.319
Bill Miller/NSF: Bill, I think you have a question you'd like just to the to the question about the

295
01:09:52.240 --> 01:10:01.320
Bill Miller/NSF: continuity, I think contingency plans in place for a partner drop. So I think it's important to emphasize that this is it. This is a pilot effort.

296
01:10:01.490 --> 01:10:04.759
Bill Miller/NSF: Uh, and we have uh,

297
01:10:04.920 --> 01:10:22.019
Bill Miller/NSF: many partners who are with us for different differing contributions and differing times. We are, as I'll showed in his slide deck. You know that the aim here is to demonstrate

298
01:10:22.420 --> 01:10:39.939
Bill Miller/NSF: the concept of the there, and at the same time really try our best to deploy resources as the near pilot that are accelerating research on Ai and and research that that employs Ai, and

299
01:10:40.080 --> 01:10:52.650
Bill Miller/NSF: we are Ah! It's a very flexible and very fast-moving effort. So We encourage you to engage in the ways that you you think are the most appropriate,

300
01:10:52.660 --> 01:11:09.800
Bill Miller/NSF: either applying for allocations to our resources, responding to the survey, and generally telling us the survey is a great way of telling us what your needs are, what your use cases are, and we will use the results from that

301
01:11:09.810 --> 01:11:13.050
Bill Miller/NSF: to shape uh what we're doing going forward.

302
01:11:14.260 --> 01:11:20.489
Alejandro Suarez: Thank you, Bill. I think that also ties into how Sharon, If you want to answer next question,

303
01:11:21.250 --> 01:11:22.990
Alejandro Suarez: you may be i'm muted.

304
01:11:25.320 --> 01:11:36.360
Sharon Geva: I was just going to answer this last one here? How will pilot success or denial affect follow-on applications? There is no tie in between the

305
01:11:36.540 --> 01:11:55.870
Sharon Geva: requests or applications for for allocations, so so being denied is not going to um be taken into account when you have a follow on, or or an additional one. But you you will want to pay attention to why

306
01:11:55.880 --> 01:12:08.080
Sharon Geva: you did not receive an allocation and and submit either a different idea or or follow the comments that you receive.

307
01:12:10.800 --> 01:12:12.019
Alejandro Suarez: Thank you, Sharon,

308
01:12:14.130 --> 01:12:19.569
Alejandro Suarez: any other questions. We have a few more minutes with you all. There are

309
01:12:20.240 --> 01:12:21.660
Alejandro Suarez: other questions.

310
01:12:25.610 --> 01:12:34.480
Alejandro Suarez: So any informal communication community like a slack channel and Tpc. Great question. I think we are looking to

311
01:12:34.540 --> 01:12:52.869
Alejandro Suarez: foster a community of near-pilot users, and as the users get on boarded we will share mechanisms for the community to interact with each other for folks that are interested in becoming involved, to have ways of communicating with us for a user

312
01:12:52.940 --> 01:12:59.290
Alejandro Suarez: meetings to to happen, different community events to happen. Those are all coming soon

313
01:12:59.830 --> 01:13:01.309
Alejandro Suarez: and stop.

314
01:13:01.490 --> 01:13:16.229
Alejandro Suarez: We have just as a as a reminder. If you want to stay up to date about what our mechanisms to be involved in here here about our updates. You can subscribe for updates on that email address. That was that was shared in the chat.

315
01:13:16.400 --> 01:13:20.309
Alejandro Suarez: Ah, then, let me make sure I share it to the right group here.

316
01:13:20.320 --> 01:13:23.460
Alejandro Suarez: Ah, there we are. You can have another

317
01:13:23.470 --> 01:13:24.440
Alejandro Suarez: happy with that.

318
01:13:37.200 --> 01:13:42.129
Alejandro Suarez: So the other folks started talking about classes, the bale of mid-march. We we

319
01:13:42.140 --> 01:14:07.429
Alejandro Suarez: we we can't promise that anything will be announced by mid-march, so we we wouldn't say to Ah! To to to have that based on. You know your your class availability, as Katie and and others have pointed out, There's there's more information coming soon on their classroom. Our our colleague, Jenny Lee, is working on that very hard with many of our other colleagues, and there will be more to come on that shortly.

320
01:14:26.760 --> 01:14:32.040
Alejandro Suarez: Oh, that's a good question. How are you defining Ai for

321
01:14:32.050 --> 01:15:00.560
Alejandro Suarez: the neare program? So there's many definitions of Ai. There are, I believe, the near task Force Report goes into a definition of Ai, The White House Task Force for the White House Executive Order on Ai, I think, has a snippet defining Ai. So right now we're We're following the the definitions that are in use by by the White House, by the Federal Government, and I would steer you to there for the the official book that

322
01:15:00.570 --> 01:15:04.980
Alejandro Suarez: that we're using. I unfortunately cannot recite it by memory.

323
01:15:11.680 --> 01:15:21.939
Alejandro Suarez: Good question someone had asked, Is there an intent to overlap the requirements from quantum computing and Cyber Security Preparedness Act with the near pilot. I think the

324
01:15:21.950 --> 01:15:34.009
Alejandro Suarez: the value of quantum computing and cyber security in the context of Ai, and the resources needed to do research on those activities could definitely be within scope.

325
01:15:34.180 --> 01:15:51.550
Alejandro Suarez: I don't know if I can answer any exact overlap in terms of the requirements of that specific act. But cybersecurity, quantum computing research. If there are ai components to those research areas, they can definitely be within scope of usage of near pilot resources.

326
01:16:01.230 --> 01:16:02.959
Alejandro Suarez: The other questions

327
01:16:08.650 --> 01:16:10.049
Alejandro Suarez: all right.

328
01:16:11.660 --> 01:16:41.459
Alejandro Suarez: I think that may do it for today. Thank you so much for joining us on the call. We know there was a lot of excitement about the launch of the near pilot, and many of you are very interested in seeing what has been made available, and what will be coming soon again, as a reminder recording of this presentation will be posted on the Nsf. Event page from which you registered and stay tuned for new announcements, including

329
01:16:41.470 --> 01:16:48.399
Alejandro Suarez: the the March pilot. Call Katie. I don't know if you have any final parting thoughts for us as we finish up.

330
01:16:48.410 --> 01:16:57.689
Katie Antypas (NSF): This is this is um amazing to see the the interest in the pilot. We are really, you know, racing as fast as we can here at Nsf:

331
01:16:57.700 --> 01:17:14.029
Katie Antypas (NSF): Um. And so please subscribe to to updates, and we will get information and updates out to you as really as soon as we can, and as as they're being developed. So thank you all for joining, and we we look forward to the new pilots supporting your research.

332
01:17:15.620 --> 01:17:17.079
Alejandro Suarez: Thank you. Everyone

333
01:17:17.160 --> 01:17:18.889
Alejandro Suarez: have a pleasant rest of your day.