 Good morning, everyone. My name is Catherine Boer as chair of the National Academies of Sciences, Engineering and Medicine Committee on enhancing coordination between land grant colleges and universities. I would like to welcome everyone to today's virtual workshop session entitled, enhancing collaboration and deepening impact. Can data science and artificial intelligence otherwise known as AI enable new collaborative platforms between diverse land grant institutions and create more impactful outcomes. To answer this question, we have a great keynote speaker and a wonderful group of discussions who will be introduced to you by Dr. Harold Schmidt, a member of our committee, who was also co founder and general partner of March capital US, and a senior scholar in the Graduate School of Management at UC Davis. Harold will be moderating today's workshop. Unfortunately, due to the vagaries of international air travel, we will not be joined online this today by Dr. Olga Bolden tiller. As with Harold Olga is also a member of our study committee, and she represents Tuskegee University as one of the collaborating institutions in the USDA NSF funded artificial intelligence institutes about which you will hear more later during the panel session. Olga's flight back from Ghana was delayed. So she is somewhere in the air over the Atlantic Ocean at this moment. We are very pleased that Tuskegee's Dr. Gregory Bernard has graciously agreed to participate in Olga's place on today's panel. Fortunately, Olga was able to report some introductory remarks to help set the stage for the workshop, which we're going to play in just a minute. But before we begin, I want to let you know that the meeting is being recorded, and the recording will be posted on the project website about a week after today's meeting. Also, there are several other members of the study committee online today. And in the interest of time, rather than have them introduce themselves, I'm going to ask the study staff to put up a slide with a list of the committee membership. Today's virtual workshop, as well as one tomorrow afternoon on creating and sustaining a culture of collaboration are intended to inform the final report of the committee, which will be released in September. We are also in the process of organizing one final working session on the topic of capacity, and we hope to announce the date and the time for that session soon. The committee's report will make recommendations on how to encourage collaborations across the land grant system that will be successful and impactful. So we are looking forward to a robust discussion today that will help provide the committee with insights as it develops those recommendations. During the question and answer period, I'd like to ask everyone online to please be mindful of the fact that the committee has made no conclusions about anything yet. So please don't leave this workshop today thinking otherwise comments made by members of the committee should not be interpreted as positions of the committee. Further, please recognize that committee members typically ask probing questions in these information gathering sessions that may not be indicative of their personal views. I would also like to note that there is a Q&A box that the public can use to ask questions of today's speakers, and we will aim to get some of those questions answered as time permits today. Please comment. I'm going to turn the meeting over to Harold for his remarks before we hear Olga's pre recorded comments. So Harold, if you would please. Thanks Catherine and just to make sure you can hear me. Yes. Good. Excellent. Well, yeah, thanks very much for that Catherine and looking forward to the session. We've been talking about the sort of profound role that data science and artificial intelligence predictive analytics advanced data analytics can play in terms of the way it's reinvented collaboration across many sectors food and agriculture is no different and will be no different. So the topic is perfect for you know the mission of this committee and we're looking forward to that so I'll stop there and let's I'm looking forward to hearing Olga's comments even though they were done under duress clearly. But thank goodness you was able to talk about connectivity across the world she was still able to do that with you know put these together which is great so I'm looking forward to hearing her comments obviously we've been talking ahead of this session and then we'll come back and I'll introduce Tom and and we'll get we'll get to his comments. Hello. And welcome to the first installment of workshops for enhancing coordination between land grant universities and colleges. My name is Dr Olga Bolden tiller and I service the dean and research director for the College of Agriculture environment and nutrition sciences at Tuskegee University and am a proud member of the blue ribbon panel for this program. Earlier this year a group the blue ribbon panel on enhancing coordination between land grant universities and colleges was brought together to identify key factors for successful outcomes of coordinated and collaborative projects between colleges and universities in the land grant system, including involving historically black colleges and universities and other institution types, which address national challenges and global food security. In the spring of this year, the panel provided preliminary observations about the nature of collaborative activity across the land grant system and the potential to enhance its impacts. These preliminary observations were open for comments from the public, which have served to inform a series of workshops, including this one today, which is entitled enhancing collaboration and deepening impact. Can data science and artificial intelligence enable new collaborative platforms between diverse land grant institutions and create more impactful outcomes. The preliminary observations focus on a variety of factors, including case studies, such as the Collaborative Artificial Intelligence Institutes funded by USDA NIFA and NSF, which in 2020 were initiated to accelerate research, expand America's workforce and transform the future of the system. However, the power of data science and analytics in food and agriculture can only fulfill its potential if all land grants contribute and have access to well curated and properly integrated data assets. Today we have the opportunity to learn how data science, including predictive analytics and artificial intelligence, can transform the food and agriculture sector from members of these institutes and explore how initiatives such as the AI institutes can be optimized in the future to enhance land grant institution collaborations by looking at barriers to collaborations, as well as examples of how some of these barriers were overcome as a part of the establishment of institutes such as the AI farms Institute, artificial intelligence for future agricultural resilience management and sustainability, which is led by the University of Illinois at the University of Illinois in Champaign, in collaboration with a number of institutions including the University of Chicago, Michigan State University, Argonne National Labs, USDA ARS, the Danforth Plant Science Center, and my home institution Tuskeet University, in collaboration with partners such as Earth Science, Microsoft, and IBM. In exploring enhancing coordination between land grant universities and colleges, 17 preliminary observations were identified, which fell into one of four sections or categories. These findings were open for comments in the spring of the year, and these comments serve to inform the workshops, such as the one here today. The four sections included collaboration in the land grant system. This collaboration was identified, but it was noted that it didn't expand to all land grant institution types equally. Section two, the rationale for collaboration. There are questions that need to be addressed that are ever evolving and it will take input from everyone to find the answers to ensure maximum impacts. And thirdly, which accounted for eight of the 17 preliminary findings fell into section three barriers to collaboration and ideas for overcoming them. And finally, section four of the preliminary report, amplifying and communicating the impacts and outcomes of collaboration. It was noted that collaborations will need to change and evolve and can take time to address key questions and issues that will have broad impacts. For our AI farms Institute, additional ones included a lack of information about the distribution of expertise at institutions across the land grant system, or other valuable assets that could make the collaboration be more enhanced. The time available for planning collaborations, leading collaborations that require team building. And this is not always a skill set that is had by individuals who might be leading these different types of collaborations. And then the institutional differences related to administrative procedures and policies, especially as they pertain to different institution types. I have included an asterisk by several of the barriers or challenges that we face as a part of AI farms, which included the foreseen here, as well as item seven and eight with the historical inequities for having to handicap the ability of 1890 institutions to be full partners, as well as insufficient or inadequate time and resources to support new collaborative projects. These were some of the barriers that we indeed did identify as a part of the creation of the AI farms Institute. And I look forward to the discussion, the robust discussion in the second hour, where these barriers were addressed in terms of how they could be overcome so that we could successfully become a part of one of the artificial intelligence institutes. Thank you for your time. We look forward to this discussion and your participation today and the outcomes of this program. Well, that that pretty much summarized why I really, really enjoy working with Olga. That was even even when she's even when she's flying back from, you know, a long business trip she's able to absolutely, you know, summarize and put things out there perfectly. And that was a good, you know, that was a perfect setup for our session today and Tom sorry you're going to have a difficult act to follow after that so I hope you've got a lot of answers from that. So I'll introduce, you know, there's a great setup I'll introduce Tom now and turn it over to him. He's the position he's currently in as Vice Chancellor for information technology and data at the University of California Irvine and University of California health. And he's responsible for ensuring the effective and strategic use of data and technology across all aspects of the enterprise. So, you know, honestly, the sorts of, you know, the sorts of framing or the framing that Olga just did and the sorts of barriers, you know, these are things that Tom works with on a daily basis in his current position and I got to know him. Actually, when he was at the UC system position so University of California Office of the President where he was Vice President and Chief Information Officer in Oakland. And, and I got to know Tom as he was, you know, doing this sort of difficult task and overcoming many of the barriers that Olga just laid out. In terms of cross campus integration of really complicated, you know, digital assets or data assets. And prior to this, Tom has significant industry experience in the healthcare industry and, and has been he's essentially made a career out of building bridges spanning differences between organizations to get to a better place and he's, he's, he's, he's an exemplary leader in terms of collaborative innovation, and especially in the data sciences space so it's a real pleasure to have him here. And Tom, you know, turn it over to you now and looking forward to what you've got to say the goal is to finish your presentation and no more than 30 minutes and then have 15 minutes for some q amp a and I'll start off with a few questions and then turn it over to the, to the committee to ask theirs. So, over to you and looking forward to hearing your thoughts. Fantastic thank you and you know Howard thank you for the introduction to the committee thank you very much. This is, you know, an honor to be able to kind of share some of my experiences insights learnings along the way. So much of what has been said already about reinventing collaboration collaboration needs to change and evolve are things that I truly believe in and I'm going to share a little bit about the evolution of how I got here to what I'm doing today and the things along the way that have shaped that point of view and the people who helped have also influenced that point of view, and then share, because I consider myself always an outsider spending a majority of my career in industry, as I'll show you very quickly, but now enough years inside of research university environments to be able to understand enough to be able to share my perspectives about how things could be better and I'm going to share those examples as we go through. And so I'm going to, I'm going to go right here and say, you know, Harold shared about my background I've spent time a lot of time in the healthcare industry. I've been both a technology executive and a business executive in that capacity. I want to talk about when I came to the University of California eight years ago, first out of system role. And for those of you who are part of university systems. You know that while your system and your one entity on one respects in a lot of ways, you're very autonomous in the way that you operate, and in the way that you collaborate. I thought I would share because it's really important to the strategies that I've been asked to build and how I've gone about trying to build these collaborative models is what I call the people that I've had a chance to interact with and work with and work under over the course of the future, which I call X factor leaders or transformational leaders either people bring big changes to environments that they're asked to lead in. And the characteristics that I think make them special that I try to emulate in how I behave but also to bring the strategies that I see them use in the terms of the impact they're trying to bring into the environment that I'm asked to impact today. And so, you know, just real quickly five characteristics I see very saw very consistently in that and again I've worked on it so my experiences are shaped by leaders from the United States from Europe from China. You know, from South America so I have a broad base of experience in terms of kind of where I've seen leadership, let's say, but first thing that great leaders do is they take the complexity that comes along with our world and they bring it down to simplified narratives and worldviews that are helpful for large amounts of people to understand and line themselves up around. This is something that has to be done and comes into how do you get multiple institutions together. So driving ambitions for the whole enterprise, even though they have an area of responsibility they think about the entire organization that they're a part of. And for the context here I'm going to change enterprise to the word ecosystem because we need to think about developing ecosystems and we're going to talk about what that means and how we enable that in our strategies and really put it to work through data platforms and data science platforms. This isn't evangelize and this is not a unique term this actually comes back from a book that was published in the 90s. The genius of the and they don't let people settle on or it's this or that they drive their, you know the conversation into, we can do this and that and that's very very important in ecosystem thinking. They play well on teams that they don't leave. All everyone I'm sure who's joined this call understands leadership and how to move a topic and an organization forward. But can you do that when you're not in the driver seat. I've seen great leaders understand the value of influence. The value of bridge building the value of molding ideas and conversations and this is something that excellent transformational leaders do. And of course they think about not just effective teams but building the leaders that that need to drive the organization because you never do these things alone. I share these with you because they've been important to the way that I think about the problem that that you've asked me to talk about today and I wanted to share that beforehand. And so if we think then about you know the question that Harold asked me to come and talk about my experiences. I'm going to start with the one that he mentioned when he and I were first introduced. My role as chief information officer at the University of California was really one to say how we strategically think about the use of technology going forward. And in my experiences coming out of healthcare and the transformations of healthcare in terms of the power around data. Was we really needed to think about the data assets that we had. And how do we start leveraging them to generate higher levels of value in 2015 now we're now going back seven years ago I pitched a strategy to the health side of the university. With its six health centers, and said, you know, we did not have, you know, an aggregated patient database of our patients that we've been seeing in our health systems. That seems crazy when I think about it today but if you've been to one you see you've been the one you see there's a lot of autonomous operation, even though we're under one board of regions, you know the autonomy of decision making has always been such that it actually can be as difficult to get. UC Davis and UCLA to work together as could be to get UC Davis to work with Cleveland Clinic. And so we had to go through an exercise of building a collaborative framework for how do we bring this data together that was partially technical, but much more importantly, it was the organizational construct the rules of engagement for how we were going to do this. And what I put on the right side there is you know we have now a mechanism for now maintaining this data asset and you can see the size and scale that we're able to do. And this is only for electronic health records, but it shows you that 17 million represented lives right and you can see that the number of procedures, you know, over 500 million procedures over 2 billion vital science and test results. So these are things now that over time have generated and we generated a what we call this Center for Data Driven Insights and Innovation now led by Dr. Toby who has his appointment at UCSF and operates as our UC Health Chief Data Scientist. This sits under him, but you can see now that we've taken this data asset and we've put it to work across a variety of different value propositions for the university. And that impact the quality of care and our ability to look at populations and study things like health disparities to look into actual the specifics of clinical outcomes and how do different types of patients with different types of diseases, maybe using different types of drugs that that are prescribed to them, how do those outcomes differ and why and why do they happen at different ratios at different facilities. We use it now those same data assets and put it in the hands of our researcher and drive both are basic as well as translational clinical science and research programs right so what we've done is we built a data asset. And we now put it to to work to create that you know the best research that we possibly good for any researcher who wants to leverage this data that drives our clinical enterprise, which is seeing millions of patients every year and trying to provide the best quality of care, because we're part of the public system roughly 30% of the population that we see comes from, you know the Medicaid, you know, Medicaid pool, we really really look at the, you know, how did the disparities of access to health and the quality of care that they receive. How does that different because we have these things at scale and compare them. So it truly is an asset that is driving both the research and practice of health care in our enterprise. Now I've thrown this, I've thrown a blank slide in here as a way to remind me to tell you a story. As Dr. Butte and I were building this. There were, there were some things that we realized and one of the things we realized is that no one's got enough data to really look and study the really complex problems that come along with health care. So in late 2016 early 2017 we embarked on an effort to reach out to a small number of colleague institutions. These are institutions I'm not going to name but you would know them. They are all in the top 10 recognized by US News and World Report, very much known for not just the quality of care that they give, but also for their, their research prowess. We approached them and said, look, what we've done at the University of California with our health systems bring in our data together, we think is expandable to a multi institutional play, and to really build up an ecosystem, you know, academic medical centers who are looking at very, very large populations but also tapping into the best minds that go beyond even what the University of California could amass. But here's what's really interesting about that conversation at the CEO level and Chancellor level of those organizations. We had complete alignment that this was the right thing to do this. There was value for all of us doing this, where we got stuck was in a lot of the compliance regulatory legal privacy. We had hurdles there that both organizationally those organizations couldn't get comfortable with, and also from a technology perspective we didn't have the type of platforms we needed to demonstrate that we could satisfy the requirements that are coming out of some of those functions. And so ultimately, the energy dissipated and we never pulled it off. I tell you that story because when I came to Irvine, which was a specific choice and a role that we created here around the strategic use of data and technology. And it was with things that that were informed by by that last effort number one was, we only did at the University of California with our UC health clinical data warehouse the electronic health record data, which is very robust and is what clinicians say and what they use to drive decisions, but it was only one type of data to really understand the individual, whether that individual be a patient, or a health consumer who wants to stay healthy. And the reality is is the explosion of data in healthcare. All right, now allows us to understand an individual on a phenotypic genotypic, multi oh my molecular behavioral and environmental level, and we can track that now longitudinally over time. And so it's with that as I came to Irvine and they said, look, we need you a unique strategy that's going to differentiate us from, you know, our peers across the country. It was with the concept of the explosion of there were things that were changing that we were going to build our strategy around. The explosion is just one aspect of it right and that these data types are not just available but they come in very, very different formats and that needs to be supported in some way, and the evolution of platforms between what we were talking about in 2016 2017 and the platforms that we've been able to build our strategies around today have evolved and better supported this concept of multimodal data platforms that allow you to bring together in our specific example. And then there's often new data that comes out of electronic health care, a medical image, a genomic data set, a waveform that comes from EKG or something taking an EEG of the brain. So we're able to integrate those things together. Clearly, one of the things that has emerged a lot in the last, you know, depending on say last five years or 10 years is the reemergence of AI. I was trained in some of these things when I was at university a very, very long time ago. I was also trained as a systems thinker. And so now that we have enough data, we can now use some of these advanced analytical techniques that we just couldn't didn't have the data for and we didn't have the computing platforms to take advantage of. Things have now become more mature. And the concept also of distributed analytics that the data does not have to all be brought to one place but that we can do analytics in a more distributed platform, allowing data to be, you know, separate from the standpoint of how it's protected and accessed, but seen as one logical system from an analytical perspective. So these things have developed quite a bit over the last five years it has changed the calculus around building our strategy. Two of the other things that I put at the bottom here are really important because we're talking about data science platforms to drive to drive collaboration and of all collaboration to new models. And it starts with an understanding and building, and this is something that you know it's taken time but my role was really built around influence in the organization is data really is strategy. Right data is the value proposition data is the thing that can bring multi interdisciplinary professionals together with a commonality of something to work from. The other thing is it's healthy organization understand is the strategy I was building for UCI were not limited. And matter of fact not designed just for UCI, but they were built around building ecosystems within which we were a part of facility building and facilitating and take advantage of network effect, which in network effect. Because the more players that are involved the more valuable the network becomes. And so I launched a program under my office, you know, you know, which we call the collaborators at UCI. The first examples in healthcare and I'll share examples out of healthcare but it's not unique to healthcare what we found as we step back from what we built for our our health strategy is, is let's say transferable to other domains, the first principles that allow you to basically define the steps of building platforms and ecosystems. And so the word collaboratory is not a unique word. It's, I found a definition going all the way back to 2007 around it's really about facilitating human interactions around a common research area, and providing access to people who are known and unknown. What we've done is we repurposed that word and say, look, in the world that we live in today, data is the engine that drives insight, innovation, collaboration and impact. And the unique thing of data as an asset is that single asset can serve multiple purposes at the same time. Very different from your automobile which are either driving and using or it's sitting in your in your driveway, right. And so we know that you know we're using data as the engine. Data alone is not valuable right it takes people with the right contextual knowledge to really put data to work and develop insights from it. We think of these data assets is not just a data platform but we call it a data operating system and I'll show you an example of what we mean it's like the data comes out of the environment in which it's generated. It's brought into environment where it could be combined with other types of data given to broad sets of subject matter experts insights develop and then a mechanism to bring those insights back to the operating system. In this case, back to the clinical care environment where we actually deliver better care, and I'll share you a specific example of that that came out of the pandemic. And then our collaborators are not confined to UCI matter of fact they're better if they are not confined UCI because we want to really drive the ecosystem and ecosystem they're dribbling around enablement connectedness inclusiveness and boundary listen the way that we define them. And so this is just a picture of how I describe to people what a collaborator is doing, it's bringing together lots of different types of partners and actors around data, where the goals are drive better outcomes drive more impact drive, you know, deeper research. And it becomes things that attract the right type of partners to us, whether those are partners are giving us complimentary capabilities are partners who sometimes bring to us the dollars that drive the development of the platform or activity within the ecosystem. I won't go into detail but you know one of the things as we did this in the health care and we stood up the Collaboratory for Health and Wellness, we came to a strategic decision that we wanted to build an Institute on top of the Collaboratory concept and so we represented as the Access of the wheel, and then we are putting these pieces and parts from across our university institution who are bringing in. What's interesting is that every time I show this slide and if I show it to you three months from now there's a new player that's coming around this wheel because we keep tapping into different subject matter experts, because they bring an expertise and a perspective to the data that may not have been represented before. And so the value grows as we add more and more players into the Collaboratory ecosystem concept. I thought I mentioned in terms of what we're doing is you know there's there's aspects of this and sometimes it can get hard to distinguish about whether we're talking about the platform that the data exists within, and the platform which is what we're facilitating by the platform. Certainly there's a lot about you know data here how do you structure the data, how do you do metadata management, how do you harmonize the data, whether it's coming data coming from different institutions for harmonization or harmonizing data on a dimension of it's coming from different sources and how do we bring together EHR data and medical imaging data or data that's coming out of a federal registry. There's way in which we continue to invest and improve the quality and value of the data asset. There's ways that we control the who accesses the data and around what rules. There are ways in which that we're facilitating these interactions where we're able to actually take it all the way back to demonstrating attribution to where did the data come from that was used in in building the model. As well as you know who were the key players, if we get the publication go back and who really contributed into this these are things that this platform can do, we've designed the platform to do and we're leveraging and testing out more and more and how we can do that. The whole idea here is that we want to facilitate team science through the platform and generate new insights and start by telling you a story that happened during the pandemic. It's a UCI story, but we're, you know, it parlayed itself into a multi institutional story. We had this concept stood up for UCI when the pandemic and the surges hit, and we put this model to work where we took data out of the health systems about the COVID patients that were coming into our facility were landing into inpatient stays. We brought them back to the to let's call it the data science platform and we built a predictive model around really understanding which patients were the most sick and that we're going to be the most likely to land into an inpatient stay or into the ICU. But we didn't just build a predictive model we took the model back into the patient care environment and then under the, you know, under the guidance of our chair of medicine, it became a tool that was used within the clinician workflow. So we put it in from a predictive model we ran it alongside so that clinicians could see it as kind of an A B test, when we thought that that model could actually help us make better faster decisions around patients. You know, we now implemented into the workflow and became part of our standard, you know, care workflow practice. That turned out to be a really good decision because we were benchmark after the winter of 2021 surge UCI had the second best survivability rate of COVID inpatient stays of anywhere in the country benchmark by a third party organization. We brought that back to not just the fact that we had data curated data built a model but that we actually had a process for implementing that model back into the patient care environment, and then also continuing to feed the data as we were making decisions back into the data set nice platform and continuing to revalid that validate and update the model so that the model would stay current with the patients that we were seeing and treating this caught the attention of the White House Task Force for COVID and they asked us to apply our the model to monoclonal antibodies. And because they had come out with emergency use, and we were one of the facilities that was using it and what we found through our models was that the criteria by which patients were allowed to be prescribed it was far too narrow. Our model showed informing policy that we could actually extend the model to more patients, and that became a critical influence for them to revise a protocol and ultimately more patients across the United States were able to get access to that. The team group came back and then funded us to work on a multiple institutional basis. We've got some published work so I can say is with the Mayo Clinic. It was with inter mountain in the mountain, mountain West region and with Houston Methodist and UCI, we actually use patient data we used imaging data we genomic data around patients and we actually informed understanding patients, along with the variance that they got. So we now proved out that we could move this to a multi institutional model. And we also proved out that using the platforms that you know that we have developed, we could deliver data faster and higher quality than some of the other institutions that we work with. You know this operating system for us is also around putting the scaffolding around it, you know to be able to support the research interaction so we found value in putting a collaboratory front door on the on the front end. It's not just about having a platform it's about helping, helping whoever wants to work with the data, get to the right data with the right tools and the right type of computing platform to do the work that they want to do. To facilitate, you know, not just the research but the interactions and the parties involved in the research and we've set up mechanisms for doing that. Another slide because I want to tell another story. You know, we were asked to pitch it an NIH grant that they that they put out last year, wanting to look at the intersection of health disparities and artificial intelligence and how these two things could. Be addressed through AI much the same types of questions we're asking here in land grants and agricultural systems. We pitched this concept to them. We pitched a concept that look it's not just about data platforms, it's about an ecosystem where you're able to onboard organizations who actually have the data on the underserved this nation which we actually have a really good partner I happen to be on their board. That you bring the data in, you put the data to work through your data science platforms you generate insights and tools but really importantly if you want to have real impact, is you have to be able to push those full back out to people who are actually caring for the patients and do in such in a way that they can actually use these tools to deliver better care, and the feedback mechanism to to to be able to improve those models over time. The concept for not just the data infrastructure we pitched a building a data ecosystem not just a data infrastructure. You know, and the way I describe it is, I, it's a little of this, right, which is, people are really busy people have built their careers around certain success and certain paradigms, you know, the concept of incremental improvements, you know, kind of can make sense if I can find the time and the energy to think about something different but what we were pitching is something completely different for that person on the left. We were talking about a platform, and we were pitching not just solving, taking rocks, you know, from the hillside, you know, to building the wall, we were talking about a platform that's going to take water from the river to the village and to take crops from the field to the village and take people from this village to that field. We were pitching something very very different we just don't think they were able to wrap their head around what we were trying to offer. Right. We actually have gone out and found a set of like minded people and are taking this concept through a public private partnership that we call the global health ecosystem. What I can tell you is, we have an FFR DC involved, we have members of the federal government involved. We have academic institutions involved. We have technology companies involved. We have scientists companies and pharma companies involved already around this concept of not just building a data infrastructure or data science platform, but really building an ecosystem that allows you to bring parties together on board them quickly and solve problems by bringing the right data together. And making it in the most usable form that you need for the problem statement that is being generated, or let's say it has been asked, and how to do that at velocity and at scale. And, you know, we've even now had discussions with regulators from the healthcare environment wanting to come in who are really wanting to say, look, the real world evidence aspect has to become part of the way that we think about bringing things, whether they be drugs or medical devices to market. So we want to understand this global health ecosystem concept for how can we use the data actually help us get to decisions better decisions faster from a regulatory perspective, which is quite unique and means that people are trying to think outside the box. So, you know, just to kind of wrap up, you know, the collaborative concept for us at UCI, you know, is being extended to other domains. Certainly the one I've talked about here and the one that's most mature today is around our health and wellness initiative and how we're building an ecosystem that parties from across the country can play. We're doing the same thing around student success initiatives, we're now have the same type of platform and the same type of multi institutional collaboration that's coming together. And the one that I'm really trying to take the playbook out and play against is how do we think about the same thing from a climate solutions sustainability perspective. Certainly a lot of diverse types of data that needs to come together, lots of institutions who are really asking themselves, how do we work closer together and bring things to impact more quickly. And that's also something that I'm working with our domain experts on for that. Just some final thoughts. You know, I think just in general with this group, you know, with this committee is talking about it's really important as is, you know, we have to think about what does it mean to be a research university in the 21st century. And that's something that our chancellor provost as me to bring to the table and to push the conversation about how we need to evolve and change. And say it's not just about what we've done in the past about creating and disseminating knowledge. It's not just about the models we use today for, you know, grant funding and you know these universities have agreed to work together but to really think about platforms that can curate aggregate curate and build valuable data asset that can be addressed that could be put against specific problem statements that are generated. Regardless of one of those problem statements come out of the federal government, some other funding agency or private industry that you know we spend our time putting our knowledge into practice like we did in the coven example. And that you know that we take our practices and we export them and facilitate larger ecosystems but everybody's got data. Right. And typically as universities we've gone from data to information and knowledge and we've stopped there. We have to be in the insight and wisdom game, you know, from my perspective and that's what I've been bringing to our strategic conversations. The reason I talk about this concept of, you know, coming out of industry, one of the things you learn is that there's different types of innovation. Right. There's product innovation, there's process innovation, and then there's business model innovation and any business professor innovation will tell you that the most impactful, you know, is the business model innovation. It's what drives the investment from, you know, from venture capital private equity, it's what drives valuations. And so what we're really talking about here and what a challenging group to make think about is, don't just think about data science platform, but think about the ecosystems that you can generate and facilitate by using these platforms and putting data at the center of the conversation. Decisions get made off of data by subject matter experts who know how to do something who understand the context around the data and can do something novel and different with them. And so really what we're talking about here and what the evolution should be to be focused on is not incremental, but really a business model information to really reestablish what it means to be a research university or a set of land grant universities. This slide is really new. It comes from industry. It's what if you are in corporate America and you're talking about your innovation strategies. This is the kind of slide that your head of strategy or your head of your chief innovation officer is bringing to the executive boardroom. It talks about innovation on the front end in the back end. What's interesting and why I put that in there is with or without us industry is doing this already thinking this way already. And we run the risk of being less relevant in their world if we don't match them or potentially lead them in the ecosystem What's interesting if you look at this slide, I could give you a version of this slide from 2016 that had everything at the top, but not this at the bottom. This is what's been brand new. And if you think about, you know, kind of what I talked about, it's about changing away from thinking about innovation is being new ideas that we talk about for enterprise versus creating an ecosystem that has ideas source from everywhere, that we facilitate bringing together what brings them together. I would argue to you that it's the data that is the common clay that the artisans are working with together. And some of the things I've learned about building these ecosystems is, you have to have somebody responsible for cultivating it for establishing the rules around it, whether it's the rules of engagement. Creating the sense of community, ensuring that the right data is there and that you have the right metadata management strategy someone has to own that how the rules are for on ramping new entrance into the ecosystem is really important. Because remember in ecosystems, the number of participants is the is the key. The number of participants and the number of connections are the key measures of success. And so this is some of the things that I've learned, and just want to say at this point I'm love to have a dialogue with with the committee members and thank you for the opportunity to participate today. Tom that was fantastic and so we've got about 10 minutes here and and also we can for the next session wrap. You know you're you're sort of part of the ecosystem so to speak here so anyway so we can also you know have questions run into that so quick one quick comment is a lot of what you spoke about really really tied together very well with our previous workshop which was on team science and a great presentation by by Jenny cross on that you know and your X Factor leader played there but also a gentleman named Andy Hargadon at UC Davis and Jenny both talk about you know the network piece so I learned the phrase. It's not the product that's the innovation is the network that's the innovation and I think that's you know baked into your comments and baked into our, you know land grant collaborations and I think you added to it with this data is the strategy so I just wanted to really make sure we don't lose that point as the committee goes you know forward with its thinking is that data is you know data is more than an asset, it actually can be and should be the strategy to certain levels so really key points I've got two questions that. I want to make sure get asked that you know drive committee, you know thinking and deliberations and then if we have time, you know we can we can take in a few more but one is. One is that every time you, you know you some sort of health or health care type of terminology so you know precision health or, you know, patient care, you know, or whatever. For me, I hear precision agriculture precision nutrition plant health instead of patient health, instead of global health care ecosystem, global food, you know an agriculture ecosystem or you know land grants ecosystem so I guess the point is, you know, I hear the data science and you know it's piece as an agnostic platform but there's more, there's more to it than than just that so I'd really like to hear your thoughts on a. You know just to drive the point of everything you're saying has application to food nag I think that's correct we'd be good to hear from you on that. And then number two is from a, you know, think you know unique strategy thinking perspective for food nag given that you have spent your time in the University of California and food nag has been part of the agreement. If you could comment on any unique strategy aspects we should be thinking about there as opposed to what you did with the health and health care piece. Yeah, so the first question absolutely and that's why the collaboratories at UCI, you know is is a program and not just specific to health right because as we step back what we saw. I'm not trained as a clinical professional right I'm really a you know a business technologist and so as I step back and I listen to people's vocabulary from different domains what resonates to me is the commonality. You know, over what people are talking about just contextually different. And so, I think it absolutely applies to the food and expense right I mean, there's that data being generated the sensors that are being deployed or more and more right so the data volume is exploding the diversity around the data that you want to combine together, the different subject matter experts that you're asking to bring together and create team science is diversifying all the time. You have, you know, federal entities that are interested in funding, but also have a regulatory role. So I think, I think if the model fits very well with food nag right and again this is you know what I see is that it's a play but that's why I call it to play book right the playbook just needs to be applied to a different context, the plays are going to be a little different but the principles behind it are very similar. Can you repeat the second question. Yeah, so the, so the, the, the second question is, is, you know, yes so great on the playbook similar, you know agnostic yes but you know given your experience with the UC and having food nag on your plate is there something, you know, particular from a strategy orientation perspective that you would say we need to pay particular attention to from the land grant side in that regard that's different than than than the strategy you deployed in health and health care. I don't know if they think of anything that's different right I think I think the concept that can get very vexing when you kind of get into the details is the what we call meta meta data management's right it's really around the definitions of the data and how much alignment that we that can you get around that. And I think that that's something that you know a multi institutional coalition should come together and put some standards around somewhere along the line earlier is better. The other thing is is the capabilities to link data sets that are really really different because those that doesn't happen. It takes work. It takes work and it takes expertise and there should be a capability built that is likely to likely for most universities, unaffordable, but collectively shared, you know, is both good from the standpoint of sharing the cost of building that capability, as well as the diffusion of that capability existing in different environments where then many institutions could be taking advantage of kind of a common built capability. So those are the two to come to my quick review. All right, then just one other question and then there's a couple in the q amp a and, and, and there's a hand up from one of the participants but the other question is so sort of building on the genius of and that you mentioned. With the land grants we have the, you know, the 1862 1890s 1994s and so, you know, the really we have this opportunity to enhance collaboration between all the land grants so in this context of the genius of and what I mean are there are things that come to mind for you where it's like actually, you know, although there's a lot of barriers and and Olga did a great job of laying out those barriers. The opportunity you see in terms of using data science to enable the 62s 90s and 94s to collaborate is there something special there that you would be excited about actually just out of curiosity. Well, I think I think there's a couple of things that come to mind. So, so one is, you know, the concept around, you know, the ecosystem and the network effect in, you know, like we very specifically put inclusiveness as one of the characteristics around our concept, which is, you know, we want to have all players that equal kind of value participation because those institutions have just different perspectives, different history lived experiences and different perspectives, but they're very valuable to studying the problems that that we have in our society. The other thing is the opportunity around training grants and training grants and how those things can be strategically used, not just to build capabilities but building capabilities into our underrepresented groups across the country. We both have those in the form of training grants in my office actually has as built. I'm just calling it a an experiential learning program so so that are we have underrepresented students who compliment what they learned in the classroom with going out and solving real world problems. And those aren't just in healthcare but they're using data science and they're out there solving problems in healthcare in health and wellness. They're understanding climate issues. And so I think it's the opportunity of finding those underrepresented populations and bringing opportunity to them to build skills, but also to have experiences that will help them, you know, as they leave university and go and build their careers and lives. And so, one time for one more question and Vikram since you're going to be talking soon, I'm going to, I'm going to enable our committee chair Catherine to ask the question. So, so Catherine you get executive privilege. What's your question. Well thank you very much Harold wonderful, wonderful presentation. Tom so thank you so much. I am so intrigued by the fact that that that you reached out to a variety of different institutions in an area that frankly is fairly easy to imagine the value of developing a system of data sharing, which is health care and well being and the larger the data set obviously the more robust your outcomes, as you understand what's going on with people, and yet you ran into the set of barriers, particularly, I'm going to call it at the risk management legal area if I understood you correctly. It really right at that space. And so my question is, do you think and this is just an opinion question that that the passage of time since the time when you reached out just try to start that and the pandemic and the success that you have now been able to demonstrate with the network that you've built among the UCs, do you think you might get a different kind of reception to to that notion of creating that kind of network, several years ago and and if so, what do you think that the biggest changes might be. The answer is yes right I think you know I think even without the pandemic we would have had some movement as you know as more examples are out there and more opportunities to challenge you've said no but look at these organizations that are saying yes and look at the benefit that you might because it's a balancing test of the opportunity that comes through data with the risks of violating privacy, regulation, etc. So, we have a very, I actually am responsible for the governance data governance conversations at our institution regardless of what type of data. So I think it did and I think the pandemic also because it forced us to do things that we just couldn't see we couldn't come to the normal no to there were just too many lives at risk and so it's accelerated our acceptance of things that that I think is help healthcare will help healthcare. Now that being said, I'll go back and say that University of California we've always kind of put ourselves out there saying we want to convene people. And so we had back in 2017 after we built a clinical data warehouse and people start challenging what we were using it for. We convened a set of experts across our institution invited outside experts, and we actually had a health data task force commissioned by our president. And there's a report that comes out of that but the more important thing is the informed conversations that happened in the committee and then cross pollinate into the local conversations which it's not the people that close minded it's just they haven't necessarily thought and heard from another perspective to get to a common shared understanding of yes, right and or what's the right path to yes. So the governance part, right, which again I think the ecosystem can play a role in bringing the right set of subject matter experts together to challenge convention for the sake of progress is an important thing that that needs to live somewhere. Where is that going to live as the land grant universities work together. Terrific thank you. Thanks Tom. So we're going to there's actually a bunch of questions in the q amp a and etc so I'll go through and for the ones that are directed to me I'll answer them while our next speaker is Well, there are some that are directed to you. So, so yeah, enjoy that. So we'll for the next hour or 55 minutes or so we'll switch over to the to the section on the on the USDA NIFA plus National Science Foundation funded initiatives and data sciences and advanced analytics and this was Tom really gave us a great view I would say on the macro perspective at a, you know, and sort of a global perspective macros perspective deep organizational perspective etc. And now, you know, because of these USDA NIFA and NSF funded centers and data sciences and advanced data analytics focused on food and ag. We have this coincidentally really perfect. You know, early stage case studies if you will to take a look at collaboration between land grants and not just 1862s but there's also one of the one of the groups are presented here and Olga and Tuskegee folks are part of it. There's a 62 and a 90 collaboration which is really cool so we have an opportunity to dig into a couple of, you know, pilots if you will because it is early days I think the first. I think the first grants were given out in 2020 the first cycle if I remember correctly and then there's been another tranche and I think there'll be more but Steven Thompson will tell us more about that but anyway we have this great opportunity to match the zoom out which I think Tom will look into but in a very deep way and now we're going to zoom in with some of these institutes that are focused specifically in the food and ag space and actually have as a key criteria the collaboration between land grant campuses and land grant campuses so with that what we're going to do the format here will be approximately five minute presentations by a five of our panel members which were privileged to have joined us here. And then after that we'll, we'll have a panel discussion again I'll ask a few questions that are top of mind from a committee perspective and then, of course open it up to the committee and, and if there's time will also get to q amp a from the audience but the main order here will be, will start with Steven, then we'll go to Gabe, then Mason, then Vikram, and then Gregory, and so with that Steven, I'd like for you to, if you're on if you can go ahead and, and pop up. I just want to introduce you you're the national program leader USDA National Institute of Food and Agriculture and supporting research education extension activities there and you've got a background and in engineering processes and precision agriculture and you know, clearly, very very well suited to be playing a key role in these data science initiatives and advanced analytics initiatives, and so we're looking forward to hearing your thoughts, and looking forward to hearing your thoughts, especially on how these, and you know these this initiative can drive enhanced collaboration across land grants so thank you very much for for joining us really appreciate it. I don't have a PowerPoint for this five minutes, which you'll be glad to know that that's refreshing good. Yeah, but I'm basically going to, you know, introduce the two that are coming. Also, give you a couple of updates on the data science and AI, or shall I say AI related initiatives that we have collaborated with NSF as a later on. So, um, as you probably know, USDA NIFA is the primary grant funding arm of the USDA. In terms of the AI institutes we have funded four of those. We have money for one more. As far as I know that's it. But the two you're going to hear from from the first year, you've already heard from Olga. The first year of funding. The question's about various, I'd rather take those in panel as they come with the with the rest of them. I have two updates, however, we have a new program called the open data framework that funds one grant to the tune of about a million and it's basically set to create a secure data repository and cooperative where producers universities and nonprofits can store and share data to foster ag innovation and this is run by Ann Stapleton, who by the way, works with me on the AI Institute, she handles the front end and dealing with the panels and various things with Jim Donnellan and Rebecca Hoa, and I run. I manage the back end primarily the post award management. Both of us have a big job in our respective roles that it works out really well. Then one more update which is related to AI is you probably know that NSF is sunsetting the National Robotics Initiative. In favor of the foundational research and robotics program. NIFA has decided to continue our interface with the FRR program. So we're on the front end of getting that cleared and getting all the paperwork and collaborations. So, this was brief and I told you it would be so I'm going to turn it back to you Harold and I look forward to seeing the presentations that are coming up. Thanks. Great. Thank you very much. Steven so I've there's definitely a few questions that I know I want to ask you but I'll respect what you said and wait to do it during the during the panel session. And appreciate those comments in that overview. So, Gabe, you'll be up next and I'll let you introduce the crew that that that you and Mason are representing AFIS alongside the AI farms piece. So, Gabe is the chief innovation officer for University of California Agricultural Natural Resources UC ANR Wendy Powers course is part of this committee and she's part of that crew. She provides leadership to UC ANR is information tech unit to support programmatic educational administrative and marketing oriented projects and Gabe is, you know, he's an extraordinary collaborator. You know, great connectivity throughout the the UC system so you know things and cross campus ways within the system but he also reaches out literally across the world in terms of multi university interactions has been building relationships with with European universities, especially bucketing in in the, you know, the Dutch sort of cohort, but games, like Tom is an exemplary, you know, representative in terms of being a collaborative innovator so Gabe over to you for some comments regarding AFIS and then then we'll move on to Mason, after that. Okay, so I do have slides and I'm going to try to get through them really quickly in five minutes, but I do believe pictures can can, you know, speak 1000 words here. So yeah, Gabe got see chief innovation officer I won't say more than Harold's generous introduction Harold thank you very much for that. Let me just sort of move on to, can you can you guys see the screen, make sure that you can see that we can. Yep, we can. The University of California is one of the not the largest land grant institutions in the United States. It has 10 campuses, five medical centers and three national laboratories and our statewide cooperative extension system run by UC and are, you know, and as you can collaborate across the UC itself is a challenge. And so we have, of course been collaborating extensively throughout, you know, the last 100 years, but internally across the vast networks can be a real key challenge with over 500,000 faculty staff and students. So, just wanted to sort of set the stage of, we've got a big UC system land grant system that we collaborate with and work to collaborate across, you know, really the nation of the world. This diagram, I thought could be helpful to, are you seeing land grant knowledge transfer, you see that. Okay, this diagram really attempts to visualize how we knowledge transfer occurs essentially collaborating with between industry in the public. Really starting with colleges and universities conducting basic and applied research that leads to academic public publishing and teaching and training of students in Ag and natural resources that may lead to pilots or demonstrations often delivered through cooperative extension. That is through direct consultation by cooperative extension advisors and agents field days and events and very specific publications and tools that are really designed to be more traditional academic applied than traditional academic publications. And certainly an important piece, newer pieces the tech transfer and licensing the transfer that technology so it can grow scale and mature from lab to commercial product through either a student or faculty entrepreneur or partnering with a company. In 2019, I think is the correct year. You see Davis you see Berkeley Cornell University of Illinois Urbana Champagne, USDA RS and our own UC and our came together all the end grants to establish an AI Institute, which Steve mentioned and has been previously mentioned through the NSF and USDA that really is aimed at leveraging AI to transform the food system and advance the state of AI itself. And so that's really the mission is leveraging transformative AI for ethical production of safe sustainable nutritious food with less resources. So, the focus of this is a really across the whole entire food supply from the breeding of new crops to farming to food processing to advancing health and human nutrition, and overall ensuring equitable, equitable participation and ethics of these new AI based technologies and the industries that are transformed by them and deploying the underlying and technologies that will ultimately lead to the transformation here. It's a very large team a faces across institutions of faculty staff and students and now an extended network of industry and undergraduate students as well. There's a number of different projects here in these areas that are really aimed at produce producing both what we would call use inspired research, but also focusing on how do we translate these, these technologies and activities into the markets like open data collaboratives and create new industry university partnerships that are driving new data sets to underlie all these new technologies that we're working to transform these industries. There are workforce development activities like career exploration fellowships and AI boot camps and creative ways to get young people and frankly new people a very diverse backgrounds and skill sets interested in food and agriculture through technology. So just some key takeaways as I try to stay within my five minutes. AI and AgriFood has required and will continue to require radical new collaboration across disciplines. And we have found in many cases that seasoned scientists and AgriFood don't know the lingo of artificial intelligence and vice versa. So we're driving a lot of those radical new collaborations with institutes like this. So AI for AgriFood is very early land grants can lead in this area through data, the data set development and a range of activities in partnership with each other and with industry. And really establish leadership that will drive the future of these new technologies by working together. And lastly, really to leverage those land grant relationships to the food system. The land grant has very, very deep ties to every part of the food system in their region. We need to really tap into those and connect them across regions, so that we can both really establish that sort of horizontal flow between all these diverse institutions but also that vertical flow between industry and our governments and our regions of which we already have deep connections in the AgriFood sector. So let's do that and turn it back to you, Harold. Thanks, Gabe. That was really, really helpful. And so now Mason, or else we'll, we'll make some comments. So Mason, actually, he's an assistant professor at UC Davis that he also he can't and the genius of and to build on Tom's, you know, commentary. He came out of Apple. And so he's a, you know, exemplary sort of industry academic interface person and, and, and being here at UC Davis I'm able to actually watch him sort of integrate across colleges and schools as well so not only industry academia but within academia so in any case Mason is a key player in this AI Institute so Mason I'll turn it over to you for some comments. Great. Thanks a lot, Harold. And thank you all for having me. I'll just jump right in since limited on time. My big goal here is to give an idea for some opportunities in terms of where we can find this interface between the institutes and land grants and academia in terms of tech transfer and what we're doing and thinking how we can learn from what other industries have done. So here I've got a graph this is a from the AI index report in 2021 showing all of the AI job postings as a percent of the total job postings across numerous sectors in the economy, which you can actually see it's funny because I think what we used to show a figure from McKinsey from about 2015 or 16 which would show the least digitized sector of the economy being agriculture and maybe there's still a lot of truth to that, but we can also see here that there's a ton of hiring going on in agriculture for AI as of 2021. So if we look more specifically at this trend you can also see that this is recent right this is a big uptake in the last two three years in agriculture where we have a jump from 1% to over 2% of the total hiring is happening just in a matter of a year so there's a lot of interest in industry and a lot of effort going into this. You can see on the academic side as well that we have, when we do a quick search and machine learning and agriculture AI and agriculture are really exponential type of increase in the number of publications happening through 2021 up till 2025 where I mean of course this is a line out but we can see there's a lot of growth in the last few years that corresponds with what's happening in industry as well. So we've got a lot of interest in this idea of AI and agriculture and food. And so the question is, how we can amplify and accelerate the broader impacts of agricultural AI. And so we know that as these AI institutes, one of our big roles is foundational R&D development but infrastructure development right so how do we sort of develop the tools and infrastructure that can accelerate and amplify our impacts across society, food processors, farmers in an environment so this is something we're really keen on understanding as a role. So to understand this better perhaps we can look at other industries and where investments have been happening recently to think where we might find models for this type of acceleration. So we have a lot of money this is showing the global private investment in AI by focus area in 2019 and 2020 and we can see that some of the top areas are maybe some of those application areas that we might think of our medicine. So drugs cancer molecular drug discovery, which we've heard about earlier, autonomous vehicles and robotics are also a big one that have happened and then a lot of this down here is actually a bunch of this is either infrastructure or finance and kind of retail. So what's happened across these industries that have been doing this a little longer than we have. Well, some of the infrastructure that's enabled this sort of amplification of impacts and acceleration impacts are data set centralization and standardization. Right, we need more data sets and we need them to be in one place standardized so that everyone doesn't need to remake the data set themselves. So this is better model performance and benchmarking and competitions has been a huge part of developing a community and an interface between industry and academia and building around a common goal, whether that has to do with examples of image net for image classification, drug protein folding discovery there's been a lot of innovation around that and other types of medical imaging as well. So one more of these, excuse me is model code and weight sharing this is once you train to model we need to be able to share that code we used and we need to more broadly as a community be sharing weights, meaning in other words, the, the trained brain of the model needs to be shareable. Right. So this I think is the development of AI coupled simulators this has been a big thing that's happened across whether we're talking about robotics more generally and warehouse applications or autonomous driving applications simulators that might exist for simulating protein folding and molecular types of models, or even, for example, in the bottom human physical simulations. So all of these kinds of coupled AI coupled simulators are needed in agriculture and food as well. So many agricultural data sets out there so I just reference one paper here's 34 just from one paper, many, many out there. There's a lot of unique problems to building models for agriculture we don't see another domain so we need to focus on how do we build these unique models. There's a lot of potential for synthetic and simulated data as I mentioned earlier. I think this is a role that we can play as the AI Institute is building out both the AI ag AI data benchmarks and models and developing the infrastructure around these coupled simulators that we can share and develop as a community and that's a perfect interface as it's been shown in other areas for going between industry society and academia so I think that's that's a big role we've got and a lot of us are pushing that forward. I'm at five minutes I'd say one one project we're doing here is called AgML this is starting to do this but by no means are we the only ones working on this but I think they're concerted centralized effort is key. So thank you I'll stop there and we'll let Harold jump back in. Thanks very much Mason that was that was great actually and I appreciate that you did the segue to AI farms there for Vikram to talk about and the visual which was perfect and Vikram I was just you know and looking at your background that reminded me I wanted to say that and you're at University of Illinois and you're you know you're the director of this AI Institute that you'll describe in a moment but you and Ilyas Tagopoulos who wasn't able to be here with us you know you and Ilyas represent something very very important which is you guys are leaders in computer sciences and AI with the pedigrees and experiences and track records to back that up and and you're you know thankfully for whatever reasons it is really appreciative your and you've chosen to be an academia and you've chosen to be a land grant institutions and the talent you know the talent situation is real and data sciences and computer science and AI and so you know we're having a lot of great conversations about looking forward and and how this can be the next collaboration amongst land grants one of the questions is how do we make sure that people like you and Ilyas actually you know really are motivated for whatever reason to participate with us so I just want to say grateful that you are and really looking forward to hearing hearing your comments about the Institute you leave. Thank you thank you very much. I couldn't agree more I do think that. Sorry, I lost my screen. I do think that education is going to be a hugely important part of making AI and data science successful as a way to create these kinds of collaborations and to develop the technologies that we need and in fact I'll say a few more words about that. I really appreciate the opportunity to be here. And as Harrell said I'm the PI of the AI farms, AI Institute, we have seven institutions that are part of this and Olga talked briefly about this earlier in our talks I won't spend too much more time except to say that three of these are land grant universities so Illinois is one, but also Michigan State University and Tuskegee. So we have a number of collaborative project between these institutions. I am not going to spend too much time talking about AI farms broadly. So, because I really do want to focus on the collaborations that we are trying to put in place and give these examples of how I think institutions can in fact work together in practice and actually give examples from two projects one is AI farms and the other is crops, which is an NSF science technology center which I talk about in a moment. In AI farms of. We have several examples of collaboration between Illinois and Tuskegee, which is an 1890 land grant institution, and these collaborations have revolved around a number of different areas. The first two sets are one mechanism for enabling sharing but in fact AI, as I'm sure you all know is much broader than that and so some of the areas we've collaborated on include computer vision, robotics and a common robotics platform can really enable collaboration in important ways. Machine learning for hyperspectral data which is very much based on data sets, but also farm bots which is an automated gantry for autonomous planting activities. We also have a very closely joint program for summer undergraduates that we work on very closely together and I'll say a little bit more about education in a moment. In a second project which is called crops the Center for Research and Programmable Plant Systems. This is an NSF science and technology center where we are partnering with Cornell, and there again we are collaborating on a number of areas that directly relate to AI and data science where the broad goal in crops is to use plants, breed plants as sensors of their own environment to give early warning for example of nutrient stress, drought stress, disease and so on. We're developing this concept we call the Internet of Living Things, IOLT which is a variant of IOT and using robotics not only for phenotyping but also for mobile sensing platforms to be able to collect data from plants in fields at large scale in order to collect this kind of data and that directly drives machine learning techniques on data driven science to be able to sense behavior in the fields but also understand plant biology. These are just some specific examples of areas where we are developing collaborations, but I wanted to say a little bit more about a longer term effort to build a large scale shared data set, because I think it illustrates both the huge opportunities but also some of the challenges in creating these data sets. As a previous speaker noted Mason said, developing data, common shared data sets like ImageNet for example and many others has been foundational to enabling collaboration and enabling progress in many other disciplines. There are a number of significant challenges in doing this for the culture because of the sheer diversity of plant behavior in different environments. And so we have developed by we, I'm sorry I'm using some of the royalty and I had no role in this and I claim no credit. This is an effort that's been going on for many years across four large data centers. And they have developed a large data set that includes genome sequences for over 800 variants of sorghum along with a wide range of different phenotypes for these sorghum lines in many different locations in the different environmental conditions across many years. And this data set is now driving new research both in a bioenergy research center called Kaby and AI forms, including new AI research on high throughput phenotyping on genome sequencing on genome wide association studies and how AI can be used to to tackle these kinds of problems. And I think what this example illustrates is that to be able to develop a data set that can actually be correlated across different environmental populations and and where you can actually do scientifically grounded well grounded studies requires a lot of coordination, a lot of investment over many, many years. At the same time, I don't want to overstate the problem either I think there are smaller data sets that can be done in a single project over a few years that are valuable as well. They're deliberately planned and carefully coordinated and I do think that this kind of coordination and investment is what's critical in order to be able to develop use scientifically useful data sets. In the interest of time I'm just going to cover one more slide because I wanted to say a few words about education I think that both as Mason said, as I said, and others have said before. The most training and and the talent gap in AI is absolutely crucial today. And at the same time we it's sort of as an opportunity to because this clearly extraordinary demand for AI among students in fact over 60% of the applicants to our graduate program and computer science at Illinois, to do AI research for their PhDs 60% out of over 1000 applications is quite an extraordinary number. And that demand I think enables sort of gives us an opportunity to create more collaborative efforts across institutions. For example, through online courses so so we're developing at Illinois and online degree program, which is a master of engineering in digital lag that covers topics like machine learning robotics IOT and others. And this is available both in terms of certificates which are specific topics as well as a full degree, but it's available to students all around the country can really enable collaboration across different schools. We are doing an AI summer school for our training students which has gotten over 30 students from, I forget the exact number but more than 10 institutions that had hackathon at the end of it. We are developing a program called I can to enable pathways into computer science and non CS majors. And again, this is being done largely online that wasn't the original intent but covert is sort of driven us in that direction. But in fact now that enables people from all over the country and even outside to be able to develop the computer science skills they need to enter the science. And there are many other opportunities as well where education can really become a mechanism to collaborate across universities, including land grant universities. And I think that's something that we really should try and take advantage of. I'll stop there I'll be happy to take more questions after the end of this sequence. Awesome. Thank you for that. And I think you know capitalizing on the extraordinary demand for AI I'd also put in that there is, there is this growing extraordinary I guess interest by, you know, the next generation let's call it of students who really really you know they get fundamentally that food and ag is, you know, a mega sector that can impact the grand challenges we face and so hopefully that's something that can help us get more and better talent into the space as well. So, our last panelist will be Gregory Bernard and Gregory, a thank you for joining us be I'm not prepared to introduce you because I was going to introduce Olga, and so thank you very very much for agreeing to join obviously your colleague Olga is in the air right now and she had great comments earlier but I will leave it to you if you wouldn't mind introducing yourself briefly and and also obviously you're part of the AI farms. Interaction that Vikram is leading so looking forward to hearing hearing your thoughts and especially looking forward to hearing how the interaction is going, you know, between Tuskegee and the others in context to the land grant collaborations between the 62s and the 90s so anyway over to you Gregory. No problem and thank you so much for inviting me. Yes, I did get this notice a last night, but our Dean she's so busy, and she works so so very hard. I'm happy to step in and thank you for a director showing for inviting me and also helping me help and prepare. And so very quickly I'll talk about the importance of our collaboration with the Center of Digital Agriculture through AI farms. And of course, to preface that we know, if you look at evolutionary genetics, we know that diversity is the key to longevity. And so, in living systems, as well as living organizations, diversity is very, very, very, very key. And so one of the things director show and asked me what were some of the ingredients that made our interaction with the Center of Digital Agriculture and University of Illinois. So impactful. And I told him what one of the more important components was inclusion and empowerment. And Dr. Davi, and of course, Dean bone tiller, brought us in at the ground floor, the AI farms project and so we were included and then we were tasked to actually develop a micro site AI micro site at Tuskegee based on the AI to center at the center of Illinois the center of the digital digital ag. And so we were asked to develop our own research objectives that coincides with the overall AI farms mission. And so that was very important to actually empower us to develop the solutions in assistance. And so, Dr. Ainsworth, Dr. Harri, we've worked with them and we have continued to continue to work with them since the start of the AI farms project and we collaborate on several other grants. One grant now that I was just notified that we are submitting now that I need to actually submit to our OSP today, so that they can review it and submit it before the deadline, and I'll talk briefly about that in a second. And so one of the projects that we've been working here at Tuskegee under developing the AI micro site analogous to the one at the University of Illinois the center of digital ag, or the inclusion of autonomous farming tools, such as the farm button so last summer we had our students and our student interns actually assemble the farm by which was developed by Roy Orrinson and others. And we've had students, Jasmine Boone and Christian Peterson traveled to University of Illinois to get experiential training in AI analysis of ethograms. And so this has been very impactful, particularly as we use this information to develop ethograms. And they're a site specific for our Tuskegee animal production. The grant that I'm speaking of now that that we're working to turn in today is going to focus on using autonomous tools and crop health diagnostics for minority and small scale farmers and so Dr. Dr. Davi, Director Davi is actually served as a co-PI as well as Dr. Ainsworth and Dr. Chahari and Dr. Upa Lapati. And so our goal here is to take the information and the knowledge that we're receiving from our collaboration with AI farms and help to further expose and provide these technologies to our small scale farmers and particularly our minority small scale farmers. We know that 90% of all farming comes from small scale farmers, but they are resource limited and they do not have a lot of access to the technologies that are involved with modern precision farming. And so we are mandated to actually provide extension and provide these technologies. And in this grant that we are developing now will provide a free crop health diagnostic for minority and small scale farmers, both in Alabama and Illinois. So that will facilitate a exchange of ideas and technologies between both universities and of course both states as we try to directly target the small scale farmers minority and small scale farmers for inclusion to these technologies. And so from the from the work that we do, of course, we do a lot of outreach and extension as myself as a professor have a teaching research and extension component. And so the technologies and the things that we we develop and we with the Center of Digital Ag, we actually use that to expand to other teachers and so we've had visiting collaborators that have visited our facilities and to understand and identify some of the autonomous tools that we're working on so they can implement them into their school systems. We work a lot with the local small scale farmers and trying to, as I said before, introduce these technologies to them and have them properly included, as well as our stakeholders student participants and other stakeholders that we will meet doing with our farmers conferences, the PwC professional agriculture workers conference, and many professional meetings we can actively engage with our stakeholders and target audiences. That being of course the small scale farmers, farmers in general and then USDA. A lot of work in the local school systems. And so we some are you programs and in fact I'm in to right now, and which we try to get this information this knowledge transferred and make it translatable so that we empower the next generation can see that we actually have a lot of student collaboration so they are farms. We meet a lot as Dr. Davi knows, and I think that that was very important as we talked about inclusion as a farm initiative was in its infancy. Dr. Davi was very keen on having this meet weekly and in order to gauge our thoughts and help us to build our collaboration and in an efficient way from the ground up. And that was very empowering for us, and also empowering for our students as you see here, even through virtual meetings we were able to process and relate what developments that we are doing here on our end to support the forms overall objectives. And so finally, as I said before it's about empowering the next generation if you look at this young ladies t shirt, which I love this shirt. To get princess, I want to be a scientist. And so we, we empower our students and TRQ is one of the students in our to you start student agriculture research team which is funded by USDA funding. They actually these these students actually work on grant projects, particularly our USDA projects and help to complete grant objectives. And so that helps us to harness their power and to provide a lot of highly skilled agricultural workers to enter into the workforce. And I think I will end with that. I hope I've addressed and answered any, any questions. Thank you for some great fodder for questions. So I want to thank you again for for doing this and actually you and Tom, both. So Tom who started and you just finished up you just really connected really well to this previous workshop I mentioned that that we had on team science. You know, we had a lot of friends who spoke and what you mentioned about, you know, Vikram and how you guys were meeting weekly. So you, you were building, you know, building this small team, the right way from the beginning, you know, and it produces. So anyway, really appreciate that. I just wanted to call that out. So from a panel perspective so we've got, we've got 15 minutes will do wrap up at 55 minutes past the hour so got about 15 minutes here. Any questions that I'm going to start with and then, and then the, the, you know, the committee I'll have rights to ask questions. So one, one question Steven I just wanted to start specifically with you. And that is, you know, since you're in this over in sort of oversight position with respect to USDA NIFA. Do you have any thoughts that you wanted to highlight from Tom's talk given his sort of macro overview is there anything that really caught your attention that is really on you guys's mind at the USDA level, you know, from a macro level. So Steven, you know, any, any thoughts from your end, you're on mute. Okay, kind of varies a lot. Yeah, um, I was, I want to talk a little about what Greg mentioned first. And I have a, I have a question because Dennis said at the doggy and I that NIFA have been tossing around an idea to come up with a program that targets small and mid-sized farms. The AI space. Right now it's very broad program. Most of the applications are centered on the large farms. And so a question I had for Greg. Let's see if I missed that did you play for a grant in this area. Did I catch that correctly and where did you send it. So the Sustainable Agricultural Systems Grant, which is due tomorrow. Our submitting application today and so I just got word that our Office of Sponsored Programs is ready to review it for the final submission so Dr. Davi is actually serving as a co-PI along with Dr. Dr. Ainsworth and Dr. Upalapati from the Center of Digital Ag at the University of Illinois. So this is also another follow up of the Farms of the Future Grant that was awarded to Dr. Chahari also addresses small scale farmers and the use of cover cropping using these tools. And so we want to get these technologies to the small scale farmer who need them who need the exposure and to provide these free diagnostic services for them so we can increase their sustainability and agriculture production, which we all know is very important we just talked about how 90% of all farmers from small scale farmers and then of course our minority farmers of course going to need even more targeted inclusion. And so we are thankful to work with the Center of Digital Ag and Dr. Dr. Davi so that we can actually develop this AI microsite here at Tuskegee so we can empower and expose local small scale farmers here as well as the University of Illinois so the grant we're working on now we're developing now for both target these minority and small scale farmers in Alabama and also in the University of Illinois so they're going to enter exchange of technologies as the PIS University of Illinois will travel here and we will travel along with our students to the University of Illinois and meet with the farmers in Illinois and they will meet with the farmers here in Alabama so it becomes a very good cross disciplinary and cross institutional exchange of ideas with the overall goal of supplementing and empowering, not only the next generation of agricultural workers, but our present and next generation of small scale farmers. Okay, yeah, this that's the big money program I, you know, Dennis and I are going to be looking for stakeholder input on what a smaller program like look like. And so we may be calling many people for that but we're trying to we're debating on whether to create a whole new program which might be problematic if we don't have additional funding or modify one of the ones we have. Second thing I want to point out. In general, you know and Ann Stapleton and I have talked about this about the five year limitation on funding for these efforts. And in general, let me pull up something. You know, and we don't really, we're not really sure from government perspective where there is a push for long term cyber infrastructure development for agriculture, you know we have these five, five year limited grants. And many of the institutes can leverage grants that they have, I know AI farms is able to combine and leverage from several different areas we have funded or NSF is funded. However, they're on a similar timescale, which is, you know, at best we may be able to get something for seven years or something if we time these things correctly. So, you know, and it's and made the point this is coming from her really that NSF has put together various approaches but they are limited by the five year limit also and I just, it would, and we follow the farm bill you know we fall follow what Congress wants and being being the agency, we cannot lobby for this. You know, I think that's a real need and, and so I'm going to turn this around a little bit this, this, you know this particular meeting. I'm, I'm looking at it from a value point of view, not to scare anybody that's not my purpose, but to see where we may be able to do better in that respect, you know what are going to, what are the long term sustainable aspects that we continue after five years is over, and Vikram has a comment, I think, I'd be glad to listen to that. Okay, I just want to second what Steve just said and actually turn it to the blue ribbon panel because to me it seems that you folks on this panel may have the standing to make this case that. With NSF there is a different science and technology centers in NSF, even though they're initially funded for five years they're routinely renewed for another five years. So it's very common that they sort of start out with an assumption that it's really going to be a 10 year effort, and in agriculture because so many activities are really have to be studied over the long term to understand the impact for example on science and environment on on many different kinds of data sets that really require longitudinal study. I think longer term investments are critical and so to make a case that agricultural center research centers as well also have a longer term time horizon seems like seems like a really valuable point to be able to make if you can. So just to build on that real quick is, so at one level there's the, you know, at one level there's the, the sort of okay data as a strategy point that Tom laid out, you know, and then sort of within that strategy and laddering up what you guys represent is a very, you know, you're the analytics capability this this program the institutes and so one, one question I wanted to throw out to you guys as a as a panel and, you know, please multiple people answer is, do you think it's possible that these AI institutes, so if it was agreed that land grants were going to, you know, taken a strategic approach data as strategy with these AI institutes, you know, would they be these nodes or centers of excellence for data analytics that could then, you know, sort of enable this overall strategy, I don't think it's possible that every institution, every university or every college is great at advanced data analytics I don't think that works with computer science you know that strategy so it'd be interesting to hear you guys, you know, maybe part of the long term perpetuation of this in the right way is that these institutes you guys have formed are actually these centers of excellence. Maybe that's right maybe that's wrong but I'd love to hear the panel respond to that. I think there's something about that I'm sure the others have thoughts too but I agree I think that these AI institutes are really bringing together each one of them is bringing together multiple in partner institutions. I think every one of us has minority institutions as well as part of this is on the 19 institutions like Tuskegee and and Gregory really has been a leader in in the forms of this role. The first four AI institutes are led by land grant universities so we don't know how I was state Washington State University Davis and all land grants and right so we really are pulling together AI expertise but it's also bringing it across institutions because every one of us is seven or eight or even 10 partners in some cases so I think that they are centers of AI expertise but also shared expertise across multiple institutions. I would agree and I would say that, you know, these are the current centers of excellence for building for the digitization of ag and food tech and building, you know that data as a strategy. If we don't do it here where will else will it happen. I mean right now, the largest, either the largest data sets in agriculture are owned by private companies that you would, you know, can't people can't get access to for for innovating. Or they simply don't exist. And so I think that this is where the intersection of this, these institutes exist to, to be those experts to pull industry other universities, other partners together large and small to to build these new tools to build these data sets there's so many crops. There's so many products there's so many aspects of the food system that all need digitization that if we don't do it in these centers. We're really going to going to lag behind the development of other nations around the world and develop these capabilities like China. So, you know, let's build on that expertise through these centers and it will take time because compared with healthcare which has had digital tools and systems to manage patients for a number of years. Those have not existed and don't exist in the same way in the ag or food sector and need to be created. So, so, so this is where I think these institutes can help and advance the strategy. Thank you for that game. It would be just taking pain attention to time I want to make sure Catherine you get a chance to pitch in your question for the panel. Thank you. And so this question is largely aimed at Vikram but I'd love to hear from anyone else and it's about data. As you spoke about the existing large data sharing that you're already involved in doing. And I'm wondering if you could say a few words about the time commitment and the expertise involved in the governance of the management of these projects and the management of these data because I think my opinion is that that's an area that we rarely give enough attention to as we're setting up these collaborative projects. I think that's a really good point. I will say that the data sets that I was referring to came out of scientific collaborations by small groups of researchers who had this as a primary research goal for themselves. So it didn't need a high level management oversight to make it happen. But in fact, what we've, I think, realized is that we've got lucky that through a combination of a sequence of multiple such collaborations. This data set has developed across all these different centers. And so if you really wanted to do this in a more intentional proactive way from the beginning, rather than just get way to get lucky. I completely agree. I think then we need that kind of management oversight and that, and that planning and the foresight to put together the right people and the right resources to collect that. And you don't need tens of millions of dollars to do it. You need some significant amount of funding but you need some planning and you need some coordination. Other from the panel other thoughts to add on to what Vikram said on that one because I mean that is totally fundamental what Catherine just asked. Yeah, I might add just one thing. I fully agree with what Vikram saying in this would just want to add on that there's an important piece of this is developing and maintaining and making accessible data. It's like you said it's not a small task but it's bigger than we usually think. And it's often, you know when you're thinking of software development team when you're trying to update maintain run quality control these sorts of things that takes many people. Right. It's not and we often maybe it's hard for us to build in those types of and they're usually well paid in industry. So there's a lot of competition for software engineers, data engineers in industry. And as a result it's hard for us to a afford, you know, 180 $200,000 a year salary for a software engineer and much less a team of them. So I think like having that capacity and that available that could be something you know shared like that as it seems like would be a good shared resource across institutes, especially on multi institutional types of I mean when I say multi I mean multiple AI institutes even working on this kind of shared data platform or whatever it might be that that does seem like we could use some additional attention and resources in that area. Thank you. And I know we're nearing the end here and I think the frustration that I feel right now must be shared by everybody, which is, we literally could easily go on for hours on this I mean with no problem whatsoever. So anyway, I just want to, you know, I want to close with a few thoughts here and then Catherine I'll turn it over you to adjourn the meeting. You know a few thoughts one is the connectivity between this session and what Jenny cross talked about in the team science thing is just like, you know, very, very obvious and I think this, you know this concept Tom that you brought out that data is the strategy data has to be the organization therefore to because organization follows strategy. And, and so we need to think pretty deeply about that. And, and I think Vikram, you know what you were referring to and Catherine you were just asking about is really important which is it's it is this investing in the systemic overview so one of the things that came through here and I knew it would, but I want to make sure we don't get distracted by is, are there lots of papers that are being published with respect to AI and and you know data science and food and agriculture yes there's tons of papers. And you know the curve is going exponential and Mason I'm glad you showed that that does not mean that oh everything's great, you know in terms of food and eggs got this, you know, this sort of taken care of and the land grants are all set with their, you know data science platforms and this is all going to go just fine I think what you know what gave you mentioned is this is at a strategic level and enterprise in our nation as an enterprise. And this is a competitive situation, and just because we're publishing a lot of papers and there's a lot of projects doesn't actually mean that we're setting ourselves up to really, you know really embrace this in the future going forward there needs to be a systemic investment, and we, you know there and data science is perfectly positioned to genuinely enhance collaboration between all the land grants and so I think this session did a great job of highlighting that and really really appreciate everyone taking the time to participate. So with that, I'm going to turn it over to you Catherine to take us to the close here. Thank you, Harold, and I'll be very brief. I am very grateful to each and every person who shared with us their expertise and their experience in this particular space I think the examples were just right on and very helpful to the committee in our deliberations and in our thinking. And as, as we draw to conclusion with our work so again I express my, my gratitude to each of you and to each of our participants thank you for your questions. And thank you for joining us today in this particular workshop we hope you'll also join us tomorrow in our upcoming workshop on on a different topic but related to the overall theme. And my gratitude to the National Academies of Science, Engineering and Medicine for setting up this important set of questions for us. And with that, I will say goodbye to all of you and thank you again for participating.