 As live participants start streaming here, I want to say welcome. This is Meng, the John A. Everson Dean of College of Engineering at Purdue, and we have a very special monthly virtual fireside with alum and also an event with an outstanding group of colleagues, Neil Armstrong Distinguished Investing Fellows to Purdue Engineering. I will turn the mic over to Armin, who was the mind behind the creation of this program to introduce the panelists. Thank you, Armin. Thank you so much, Meng. Back in 2019, we celebrated 50 years since Purdue Engineering alumnus Neil Armstrong's one small step inspired the entire world about the limitlessness of human and technological achievement. About a year, Purdue Engineering was really proud to launch this program to bring highly accomplished and world-recognized scholars and practitioners to Purdue Engineering to really collaborate and catalyze interactions that would increase the visibility and impact of our students, faculty, schools, and college. We're really delighted that today, five out of seven of our Neil Armstrong Distinguished Investing Professors are able to join us for this panel, and I'd like to do a brief introduction of them. As I read your names, just please, who worked at Lockheed Martin from 1985 to 2011 in many different technical leadership roles, and is world-renowned for inventing the vertical lift concept in the F-35JSF Joint Strike Fighter. He's a member of the NAE, a fellow of the AIAA. He's the recipient of several awards, including the IAAA's Neubuld VSTOL and Aircraft Design Awards and the Daniel Guggenheim Medal, Lockheed Martin's Kelly Johnson's Inventors Awards, and the Aerostar and NOVA Award, SAE's Aerospace Vehicle Design and Development Award, and the American Helicopter Society's Pauli Hoiter Award. He was recognized as engineer of the year by Design News Magazine in 2004. Paul is working with the School of Aeronautics and Astronautics in many topics, including aerodynamics and new concepts in aircraft design. Welcome Paul. Thank you. Dara and Takabi is the Bacardi and Stockholm Water Foundation Professor in the Department of Civil and Environmental Engineering and Professor in the Department of Earth Atmospheric and Planetary Sciences at MIT. He's an expert in hydrological remote sensing, and is currently the science team leader for NASA's Soil Moisture Active Passive Mission. He's a member of the NAE and a fellow of the IEEE, the American Geophysical Union, and the American Meteorological Society. And Dara is working in the School of Aeronautics and Astronautics as well, really on new remote sensing techniques for soil and water, in particular the use of signals of opportunity in the P-band below 400 megahertz. Welcome, Dara. Thank you. Also joining us today is Enrique Iglesias, who is the Theodore Firmulent Professor in Chemical Engineering at UC Berkeley. His research addresses the design and synthesis of structural mechanistic characterization of inorganic solids used as catalysts for chemical reactions that are important to the production, conversion and use of energy carriers in sustainable petrochemical synthesis and in the protection of the environment. He's a member of NAE, AAAS, and the National Academy of Inventors. He has received several awards for his contributions to chemistry and chemical engineering. Most recently, he received the 2019 Michael Buttert Award for advancement in catalysis and the AICHE's William Walker Award for Excellence in Contributions to Chemical Engineering Literature. Enrique is working with faculty and students in CSTAR, which is our Center for Innovative and Strategic Transformation of Algae Resources Center on Research, on Industry Innovation, Entrepreneurship and Educational Activities. Welcome, Enrique. Thank you. Also joining us today is Randall Poston. Dr. Poston is senior principal at Pivot Engineers, a structural engineering consulting firm in Austin, Texas. He's an internationally recognized expert in structural engineering practice. And he's an inductee in the National Academy of Engineering as well. Randy has established himself as one of the preeminent structural consultants in the United States and has authored and delivered hundreds of papers and presentations related to the structural engineering industry. He's championed the repair of existing structures for upwards of 30 years and dedicated his career to advancing the state of structural engineering knowledge. He was the past chair of the American Concrete Institute's Committee 318 Structural Building Code, where he oversaw a monumental effort to completely reorganize the concrete code, the first in undertaking of its time, of its kind in the history of ACI. The engineering news record, ENR, named him a top 25 newspaper in 2014 for his Code Reorganization Leadership, and ACI bestowed upon him the Henry L. Kennedy Award for his work. Randy is working with faculty and students in the Life School of Civil Engineering in the area of structural engineering. Welcome. Welcome, Randy. Thank you very much. It's a pleasure to be here. Mordecai Moti-Sagev is the Robert Shilman Distinguished Professor of Physics and Electrical Engineering at the Technion in Israel, where he received his BSc and PhD as well. After post-doc at Caltech, he joined Princeton, where he rose through the ranks very quickly from assistant to full professor before being recruited back as distinguished professor at Technion in 2009. Moti's interests are mainly photonics, solitons, lasers, and quantum optics. These are topics with issues collaborating with faculty and electrical and computer engineering here at Purdue, but also in physics. He has won numerous international awards, amongst them the 2007 Quantum Electronics Prize of the European Physics Society, the 2009 Max Born Award of the OSA, and the 2014 Arthur Shalo Prize of the APS in 2011. He was elected to the Israel Academy of Sciences in 2015 to the National Academy of Science in the US, and in 2021 to the American Academy of Arts and Sciences, AAAS. In 2014, Moti won the Israel Prize, which is the highest honor in Israel, and in 2019 he won the E-Met Prize. Above all his achievements, Moti takes pride in the success of his graduate students and post-docs, among them are 23, who are professors in the US, Germany, Taiwan, Croatia, Italy, India, China, and Israel. Many hold senior R&D positions in industry. Welcome, Moti. Thank you very much. Pleasure to be here. And moderating our Farsight chat today is my boss, Monk Chiang, the John A. Edwards and Dean, the College of Engineering, and Dorosco George, Distinguished Professor of Electrical and Computer Engineering. Over to you, Ma. Thank you so much, Arvind. Just one minor correction on no one's boss. I assure you, no one thinks of me as their boss. Now, it's, however, a great pleasure to be here in the same virtual room as our five new Armstrong Distinguished Visiting Professors here, at Purdue Engineering. And as you can tell that this is truly an incredibly talented group. And thank you for sparing some time every year visiting us at Purdue and working with our faculty and students here. And you represent quite a cross section of different engineering disciplines and also from different parts of the country and the world. So to those who are watching in real time, you can also ask questions in the chat box, although I've already generated quite a few to start out with. And those who are watching later asynchronously, you know, feel free to reach out to these incredible researchers with your additional questions. Well, the theme here today is looking at the virtual physical interface and how that's going to address the education research societal challenges. Well, maybe I'll just spend one minute giving my bias to view on what I mean by this virtual physical interface here. So if you think about what we code and what we touch, they increasingly interact with each other. One form of interaction is I want to run physical experiment at the wind tunnel and I want to extract a mathematical representation from it. Another could be that I got remote sensing in the building outside the building and the interactions is physical. But the way that I design the interaction relies on software, relies on codes. Another one could be look at autonomy as a standard example today that you've got softwares driving the car in some sense. So you've got an autonomous connected, almost intelligent thing driven by machine learning, driven by software engineering. But they are driving physical objects. They are moving sometimes physical, dangerous, heavy or light physical objects. So that decisioning and control interaction in real time is another type of interface. Now, the first question is a generic one and I will give that to all the panelists is what do you think whether it's in your own area or neighboring disciplines is the most critical or interesting kind of interactions between the bytes and atoms or between the virtual and the physical worlds? And with five panelists, you know, maybe I'll start out just sort of first name alphabetically if you will, you know, start out with Dara and then Enrique and then Marty and Paul and Randall, please. Yeah. Thank you for this panel and really a pleasure to be here. I think in my field, actually, the use of the virtual world was early on one of the problems in computer science, you know, numerical weather forecast was something that was one of the early benchmark sort of virtual world applications. And this is going back to 1950s in Princeton, actually. But it hasn't, throughout all these years, it hasn't made experimentation and physical representations of processes obsolete at all. It has made them all the more necessary. So it's not like one is replacing the other, but just expanding the capability and creating new problems to be solved in a sense. So, you know, I think numerical weather forecast is a good example of how early on the virtual was adopted and it's part of the necessary arm in that, you know, aviation without weather forecast would be a very dangerous thing. Yes. Thank you, Dara and Enrique, please. So if I divide my discipline into three components, one of them is the chemistry component, one of them is the engineering component and the other one is the classroom. And if I may speak briefly about each one of them, I think that would make it fairly complete. So when it comes to the chemistry part of chemical engineering and what we do in rearranging molecules, clearly the advances have come from being able to describe the motion of electrons and their positions using equations that are too difficult to solve and as a result of it have become increasingly less approximate and increasingly more visual. The ability to think in multi-dimensional space since every vibrational mode of a molecule is actually a potential reaction coordinate is something that humans are not able to visualize, but machines are able to provide for us at least a final answer. And perhaps in the less chemical end of things and more physical, I think the protein folding problem is a classical example of how it is that machines with some learning capability are able to now solve problems that were not solvable before. In the engineering side of what I do, I think the machines have for a long time allowed us to interact with processes in ways that allows to sense things that we cannot see, detect problems that we would not otherwise detect and act on those problems with remedies that lie supported by a model that tells you what to do when something happens. A lot of that has been moved to the virtual side as a result of the fact that safety concerns associated with operation of chemical processes are best handled from a distance. And I think today for the most part, most of the operators of those chemical processes are trained with the moral equivalent of a flight simulator in order to be able to detect problems in silico so that when they happen in practice, they know what to do. When it comes to the classroom, and this is a conversation that perhaps we will return to, but when it comes to the classroom, I think the recent emergencies that we have faced have made it clear that we can deliver information to more people more efficiently as a result of our ability to communicate over distances and over different times by recording, by talking to large audiences, so just the ones that we have gathered here today. What in my view has not happened in my classroom is that I think the recipients of the information have not really adapted to how to absorb knowledge that comes in a remote form. And if there's a subject that is dear to my heart and if the conversation lead us there, I think that there is an untraining of humans that needs to take place before we can accept knowledge that comes in a different mode and without the social component of human interactions. Thank you, Erika. We'll come back to some of those points. And Mardi, please. Well, I want to go back in history of science, 40 years, early 80s. Richard Feynman, already Nobel laureate, this is like about 10 years or so before he passed away, realized that there are several models, actually many models of physical processes and materials that we assume that we think we know the equations, but we can't solve them. And the whole classical, all the computing power on Earth together, taken together, cannot solve them because these problems diverge and it may take them thousands or millions of years or so to solve. Some of these problems actually also we know from day to day life, like, for example, this traveling salesman and so forth, optimization problems. So Richard Feynman suggested at that time to do a quantum simulator, which is like a special purpose quantum computer. For some years, people really didn't really consider that seriously, but now everybody's talking about quantum computers. So I'd like to talk about those, really the ability to compute processes that we think we know the equations and we can't solve. There are so many effects and phenomena in the physical world around us and also in engineering applications that we would like to be able to compute. So obviously we are facing a major challenge. There are many companies and countries that are involved in this business. The bottom line is that if 10 years ago I thought that maybe we are near a quantum computer, now we think that we are still looking at a real quantum computer about 10 years away. But let's imagine that we have one, which we don't. And I am happy to provoke IBM and Google and many others and we don't have a real useful quantum computer that can do something that is really important, not only breaking codes for national security or something, but for example, computing some molecules so we can design better drugs for personalized medicine, for example. They can't do it today. And so I'm happy to challenge any of one of these that will tell us that we'll have it in a year or two. I can tell you that it's not so coming in the near future. But this is where I think the major challenge is going and I think we are going there more and more. I think we will have one, we will have, but it's not as near as people think. And I think it gets our imagination quite excited to be able to understand the world around us and also use it. Designing better drugs is just one example out of very many. I think Dara talked about weather forecast. This is another problem of this kind that a quantum computer would be able to simulate better. Will it solve it? I don't think so. At least because there are several arguments why probably not because it's a chaotic system at least sometimes. But there's a lot to do and major challenges for the next generation. Thank you, Marty. And Paul, please. Yeah. For us in the airspace field, you know, 50 years ago there was no software on airplanes. The F 16 was unstable and pitch and had fly by a wire hit ahead at like 250,000 lines of software code. But today the Dreamliner and the Joint Strike Fighter have over 5 million lines. In fact, they're both programs ran into problems in developing the software. So it's become very important. And we joke in the industry that the airframe is a dust mix and software. So for us, the interactions between software and simulation involve airplanes that are pretty well defined. I mean, a jet airplane looks like a Boeing 707 transport airplane looks like a C 130. But if you're doing something completely new, like the stealth F 117, you can't really rely on software. And our philosophy there is you do enough with the software to make sure you haven't got a dumb idea. And then you go to rapid prototyping and you test it. And in fact, that's part of the scientific method, isn't it? I mean, how something your hypothesis moves to a theory is you predict something that no one's seen before, then you go out and see if you find it. And if you find enough times it becomes a theory. So there's important interaction between theory and experiment and I could go into it. Perhaps I will live in later questions. A lot of examples on our field where we thought we had the answer and we neglected an effect or a term and we were completely wrong about it. No, well, thank you. We'll come back perhaps to some of those very telling examples, Paul and Randall, please. Yeah, I several things come to come to mind and hearing the various panelists talk and certainly my, my world as I'm really a practicing engineer structural engineer or civil engineer dealing with with infrastructure so it's you know very physical sorts of things and not. We do rely heavily on, you know, computerization and software. I look back on on my career and, similar to what Paul was saying, we barely had, you know, computer analysis that we could do and use for structural analysis and the like and now it's, you know, basically you build a model and visualize it and visualize the the the forces and the stresses there in the model in order to do your, you know, invest do your investigation or do your design so I'm of the belief to that, you know, it's it's a tool to help advance our knowledge and our in our in our building you, you know, this the important part to me is whatever simulation or analysis you're doing has to be an accurate representation of the physical so you need somehow to tie the simulations back to reality and whether that's from physical models from prototypes to experiments and in the laboratory. I think that's a that's an absolute must at least in in our field. Yes. Well, thank you so much everyone. You know, I want to follow up on this theme on the classroom and education in engineering. And I'll ask two related questions. We see that a lot of engineering students, not as much at Purdue, perhaps, but I have heard at peer institutions, public and private other top schools, where engineering students are leaving the physical side of the degree choices and crowding into a lot in computer science or data science or AI degrees. And, you know, is there a competitive job market you see out there or anything that can be done to give these students a balanced set of choices. I'm not questioning their rational decision making ability. I'm just simply saying, as educators, you know, to give them truly a well informed decision process, and giving give them the range of choices, you know, I give one example in the semiconductors. You know, we see that hundreds of billions of dollars are going to be spent this decade in the US and some other places to build a new fax. Some of the fax will require BS or MS, if not PhD level engineers to work in them. But just getting students to be interested in taking semiconductor courses while rolling out a whole set of dedicated degrees from Purdue next month. Nations first, I think. But it's one thing for us to offer it. It's another for them to be interested, even when there seems to be a pretty good labor market, maybe I don't know how well they pay compared to jobs on programming for AI applications. So any thoughts on that? Well, I think there's a huge labor market out there. One of our big problems on the joint strike fighter was firing enough software engineers who wanted to come from the gaming companies, the video companies, etc, to work in aerospace. They didn't realize that there was jobs for software engineers in aerospace. So there's a huge demand for them. There are lots of jobs they can do with the artificial intelligence in the automobile industry. I think the students are aware of that and that's why they're going into software engineering. But our problem is letting them know that there's software engineering to be done in the other fields. Civil engineering, mechanical engineering, aerospace engineering as well. Thank you, Paul. Anyone else? I've heard some fellow deans of top schools sort of wondering whether they are seeing this trend towards a ginormous percentage of engineering students into computer science. Whereas the degrees such as electrical or chemical or civil mechanical error and so on. Our error may be slightly different just because there's a lot of very hot news out there around the arrow and astro. But a lot of the views they're looking at the number of BS degree enrollment is shrinking. Again, I want to qualify a Purdue engineering that's really not the phenomena we're looking at shrinkage. No, quite the opposite. But in many other places, they see that you can add up all the other engineering degree BS enrollment and they're still smaller than the enrollment in CS. Maybe that's just the way it's going to be. I'm not saying it's not healthy, but just wondering about your thoughts. If you find that there are any schools out there that need more chemical engineers, UC Berkeley is happy to provide some because our enrollments are continuously increased for the last five years. Chemical engineers have a well deserved reputation for claiming that they can do anything, including software engineering if they're called upon doing it. But we do take a certain position that we tend to be two users rather than two makers. And in that sense, I think that at least from the way that we train our students, we clearly have to bring back the logic of programming that now is done by numerical packages. We need to teach them again how to write a logic diagram because it will help them not only in writing software, but also in reasoning through complex problems in a logical way. So my feeling is that in spite of the data science that today attracts many of the younger members of the engineering community, that at the end is going to be the ability to use statistics probability to be able to know the logic of programming. That is ultimately going to lead them to be able to apply the algorithms as they're written to do machine learning artificial intelligence without ever having to become the ones that actually develop the software and the hardware to be able to do it. I view it as a return to my roots to some extent, because that is a way that chemical engineering was taught 40 years ago. Okay, when we had to program in C or Fortran, for example, we didn't have a package like MATLAB to be able to do it. And we took a course in statistics and probability that served us well outside anything that had to do with data. So I don't think that at least in my discipline, we have to worry about an exodus into the more glamorous areas today. Okay, thank you, Enrique. And Dara, you've got your hand up. Yeah, the current market trends and in data science and AI among students may continue, it may saturate, it may be that the new equilibrium is going to be that the majority of students are in that field. But it also depends on their ambition is is the creative process that what is it that you're applying? What is the innovation that you're trying to apply data science to? It's the idea of the app not coding the app, which is the issue. And that goes, where do creative ideas come from? Is it, can it be acquired purely by virtual education or does it require a culture and a incubation space like a university campus? And also, that means that the sort of physical sciences like chemical engineering, civil engineering, biology and all that have to persist with some level of activity because that's where the problems come from. It's not just coding, it's coding what and also it goes into the universities using the fact that there's all disciplines, social sciences, history and how do those inspire creative ideas? How do those get engaged? It's not just social science requirements, it's where ideas come from. Yes, I want to follow up on that and I know there are also audience questions in real time streaming in, but Marty has a point. So I think that in terms of undergraduate education, AI is going to be very big, certainly the next decade, maybe also after that, maybe we'll stay like that for tens of years. But for PhD, I think the ideas need to come from the physical world. There is a limit to how much you can come up with just by doing, working with software or even working with mathematics. Everything else comes from the physical world and I think that in terms of the level of the PhDs, we will still have them more or less uniformly distributed, maybe slightly larger in AI. The undergraduate population will still go to AI simply because they pay more. If you look at what happens now, for example, at the salaries that undergraduate students or those that just graduated or about to graduate, they get offers from companies like Facebook and Google and all kinds of others, smaller companies that are at the level of a senior engineer in the industry and all other kinds of disciplines. Including in very competitive ones like photonics and quantum and so forth, still the Facebook kids get more. So that I think will continue to drive the undergraduate world. But for PhD, I think for research, there's still going to be always, I think, more or less distributed as it is today. Now I want to point here to another thing which is also very interesting. It turns out that today a lot of the cutting-edge research, not development, but true research is done in the industry in places like Facebook. It's hard for us to swallow, hard for us to admit because when you think back, the research at Bell Laboratories, for example, the mythical place, this is like the ideas factory, if you know that book that tells us about how the transistors invented and the laser and satellite communication and so forth. At that time, this was like people viewed it as the world's lab. Today, these are private companies actually. And really the research, the cutting-edge research is done there, not at MIT, maybe also at MIT and places like that. But first of all, it places like Facebook as much as we wouldn't. It's hard for us to swallow. Yes, Facebook, Google, MIT and Purdue. Now, I will come back later to this point to have time. But let me go back to this point already raised by some of the panelists repeatedly. And that is, if I flip the question here and say, what about making either basic programming skills or the mindset of software engineering as a way to create, as a way to do problem solving as a required general education? So now these are two distinct things. The ability to code and the way of thinking as a coder are not identical objects, but maybe either or both. As a gen ed, as we call it here in the U.S. system, just like you have to take a minimum number of courses in history or foreign language, every student, including sociology majors. Now, a sort of related sub-question would be in engineering, we can make it sort of a required. I'll give you two quick examples. One is that we started something called every border engineer codes is EBC. What it does is a non-transcripted, non-credit course on Saturdays, Saturday mornings to offer to students. And it won't even show up on your transcript and you may pass and get a certificate. Well, we grew from 50 students taught by students to this year, 2000 students in one year took the biggest, the class of Purdue engineering is not a class. Not officially recognized or transcripted class. We're also rolling out to coming year as an AI minor, which is offered to all the engineering majors. So you can be a major in civil or chemical engineering and take a minor in AI. So, you know, we're trying to see if there's the best of both worlds here. But back to my question to the pan all panelists is, what do you think about making programming like a basic understanding of history and language general education requirement for all college students? Say I Purdue. So since I brought it up, let me let me make the first comment on it. And it and to me that the logic of programming, the reasoning that goes behind it, the the tenants and the tools of statistics and probability. That is part of an education that is part of what you keep at some point in time when you're not coding anymore and you're not doing experimental design of experiments. And I distinguish that from what I would call not despairingly vocational training, which is training people how to put those to the solution of a particular problem, the coding of a particular exercise or the analysis of a particular data set. And I do put it in the same general education category, both the ability to the logic of coding and statistics and probably to be part of a general education in much the same way as the humanities are. And then the training on how to apply to specific problems. That is something that is local that the students are going to need, but they're not going to take away the particular problem and the particular code but what we taught them about making connections as a result of those general courses. Yeah, I want to add to that that I like to make a distinction between programming or coding and software engineering coding is what an undergraduate does he's given an assignment and he's to write a program. But software engineering says well what is the customer want what's the need what are the requirements, how did I turn that into a program. And that's what's really important I mean it's like teaching someone to run a computer language to learn a computer language that language is going to be obsolete eventually. But computer coding or software engineering is going to persist and that's what they need to really learn. The other comment I'd like to make is students seem to believe the codes more than the experiment they pick up a number from a code and they don't have any reason to no practical experience to say this is wrong it can't be right. They seem to say it came out of a computer program so it must be right just like if you find it on the internet it must be right. Yes. Thank you for highlighting the differentiation between a vocational training of a skill of operating in one particular language versus the mindset, including the ability to not blindly trust code. But I assume that you are implying that if this the mindset we're talking about maybe part of the mindset training is through a concrete programming language in some steps should become a general education requirement for anyone to graduate from a college. Which language though. Right that may again change over time right but that will be purely used as a tool as a vehicle to deliver the educational outcome of training the mindset. So if that's the case, the angle now being get them to be expert in one particular language. I assume some of the panelists are implying that they would be supportive of making that a general education requirement, even for history majors. But while I see Marty and Randall's hands up please. Yeah. Yeah. Marty you can go first if you like. So let me first answer the question in my opinion yes in my opinion this is like mathematics in elementary school. We will have to teach everyone to have some basic coding skills to write a computer program and to do a flow chart and understand what it means just like we required everybody will know some basic operation mathematics and elementary school. And in my opinion at some point it will go down to elementary school for all the kids and they will learn it maybe a third or fourth grade to know something about this. This will also maybe make many people that today you try to teach them how to code. You know if you try to teach them how to code at early age. It may be easier. But I also want to relate this question of yours monk to the previous one and tell you a little story. You know, this is like going back. I think 60 years or so maybe 70 years to Ramanujan. Ramanujan was a mathematician from India that saw that was able to make conjectures major conjectures that sometimes you took years to prove in mathematics. Okay. So one can ask legitimately is will a computer one day be able to make conjectures of that sort. The answer was given two years ago. The big of mine actually my former student you don't come in and wrote a beautiful paper in Nature magazine on the Ramanujan machine. He wrote a code. He meant several undergraduate students or undergraduates wrote a code that on an ordinary computer could run on a laptop I think that can actually come up with conjectures that he takes true mathematicians professional mathematicians sometimes years to prove. They opened the community already and it's all shared online. And if you read that and you will see you'll be amazed some of them were already proved so in other words, a I came up with conjectures just like Ramanujan did. In my opinion, that addresses the previous question but it's a major one. Here we have artificial intelligence coming up with something that we thought can come only from the human mind. Very interesting. Started there. Thank you, Marty. And I think Rando had hands up. Yes, please. Yeah, I just wanted to. Yeah, I certainly think there's some need and basis for and I would even agree that any student at Purdue University ought to have some basic knowledge of programming these days and as we've been discussing the. It's not it's not what's being programmed. It's the idea of connect connecting the the whatever the the problem is or the question is to the to the end result in the process of going through through that in your in your mind and developing developing software. I mean, I look over my career and you know the days of programming Fortran and then had to adapt many times over to different to different languages and the different types of programs and you know if you have that basic programming mindset or knowledge about how how software is developed. I think you you're much more adaptable and I've had to adapt and you know the last 37 years and to this this day. Certainly the things that we have to do have changed we do a lot in remote sensing and then structural health monitoring and a lot of that takes. You know programming of a different of a different set in order to be able to make sure that your systems and systems are are working so I do believe that it's almost fundamental to an education these days. Thank you all by the way I'm trying to text response back on the chat box as well now let me shift gear to ask another question which is one that administrators ever that means of university need to answer quite a bit and that is why if the world. Is going virtual most of the time the same where the margin the value might be often derived should a university student invest heavily in physical facilities for research. Now just in case they are Purdue faculty wondering what's my answer to that question my answer is yes. So you know I don't want to scare anyway anyone on the on the call right now from Purdue but let me ask this question to our visiting distinguished faculty here. What do you think I think we've already touched on this question we said that problems come from physics and you test your hypotheses against physics or so for example Einstein's theory of relativity was. Proven demonstrated when there was an eclipse and the light from a distant star was bent as it passed. You can't do that. I mean you can simulate it but you still have to run out do an experiment to prove that what you've got is true. If enough predictions of a theory or a hypothesis that you don't haven't previously observed or then subsequently observed he said OK that's a law now. So problems come from physics from mechanical systems and proofs come from mechanical systems you can't do science without them. So I agree with you. Yes well certainly there's the ability to falsify and you know how would you falsify without the physical facility that's that's part of the answer. Sure. What do other panelists think. So it is interesting that it in my own field of research the simulation part has become most useful by falsifying hypothesis. And that is because many of them are truly absurd enough that even a low level of simulation would allow you to say that it is not correct. So in that sense I would turn around a little bit OK and say that we can do the kinds of calculations that will rule out many many possibilities that experiments could not rule out. And then we can focus the experiments on those that remain on the issue of whether universities still will require physical space for research. Yes a resolute. Yes in my field without the experiments you're not going to be able to deal with the physical reality. I do have a concern and that is that you know if we say that experimental research takes up space and we say that virtual research does not for purely budgetary reasons. There may be a preference on the part of certain administrators to actually hire people that do calculations and simulations rather than people that require sophisticated equipment. And that is a tendency that we need to resist because in my field that in most others there is a definite need to maintain that state of the art expertise in experimental research. Yes. Well I assure you that here at Purdue we continue to invest heavily. In fact just in hypersonics area long over one hundred ten million dollars of the University Central internal funding in the past one year devoted to setting up a new experimental facilities. And that's just tip of the iceberg and it goes much broader than one area. So Purdue certainly is not leaving the physical side and we try to bring the virtual side into the physical in education from virtual labs to programming not only as a skill but as a mindset. And we try to bring the best of virtual and the physical and research strength and impact together. But we're not leaving that behind. We're not going just with the flow of the fashion or purely because of financial reasons. It is true. Some of the faculty experimentalists to start a package can be quite hefty. It's like pulling teeth. You know it's it pangs me to do it. But you've got to do the right thing and invest. Marty please several things first of all not everything can be simulated at least today on classical computers. There are many many things that cannot be simulated. So for this we have this what we call quantum simulators where what you do actually you create a physical system of cold atoms or cold molecules. Of something like this and then you set up the experiment the potential and you set up the initial conditions and you let it run multiple times and take the data and the ensemble average of the expectation value is the result. What a computer that would one day be powerful enough to give you case on that part we have no choice we have to do quantum simulators. But even for classical simulators I have an opinion actually I have a whole talk about this that explains why they are so important. In principle one can say look there are some classical things that I could simulate why do I need experiments. And the answer is that when you do when you do a simulation what you have is an equation some dynamic equation some boundary conditions some initial conditions and everything is in your starting point and then you let it run. Let's say the computer can solve it but when you do an experiment what you're trying to do actually you're trying to take a system a physical system isolated from the rest of the world and just try to address a particular issue a particular model. Now in in learning how to isolate that system from the rest of the world you learn a lot of things and you learn a lot of you come up with a lot of ideas about your own system and about things additional effects that were not initially included. And in addition to that you can never really see isolate your system or any system completely from the external world and they are always surprises. And I gave actually a talk about this a colloquium at Purdue in the summer and how exactly from this kind of view that we must do experiments came a whole new idea that lately in the last few months went already to the industry and is already there are several companies working on that and maybe one day it will succeed. So that started off as a simulator and ended up with technology. So I strongly believe that you are doing the absolute right thing. You have to invest in experimental systems. There's no replacement. Yes. Well you know I think we are seeing to the same choir here but indeed you know here at Purdue we believe that we have to pay attention to the power of what we code but we're not going to leave behind what we touch. In particular we believe that what we code will be more valuable if at all possible. In some cases not as you just highlighted Marty but even when it is completely doable it will become better more valuable. Absolutely more differentiated. If you put it together with the what we touch part. And that's why you know we continue to invest in Birken and a technology center for the tiny ones. The Bowen Lab which is five story building. You can slide into this building to do earthquake study. That's why we have tremendous investment in aero astro including hypersonics. That's why we are building resilient habitat on the moon with a experimental test bed here. And that's why we have the worst largest indoor outdoor Zhongdong by any university at our airport. So the list goes on now. I want to quickly remaining few minutes and I know that there are great interactions on the chat box. If we don't get there my apologies but I do have the urge to ask a question. Maybe there's like a quick answer. Thoughts from each of you here in the Midwest and now it's not just Purdue also Michigan. There's Illinois and here in the Midwest. Manufacturing is there a hope. And this is not just research education. It's also university as an economic engine as an innovation engine is there hope that technology may rewrite a manufacturing industries equations so that we can bring industry 4.0 back here to the Midwest. Additive manufacturing and new technology that could be applied here in the Midwest. I mean. Yes. Additive is indeed a you know additive slash digital slash other type of process innovations. In addition to material innovations you know may indeed rewrite the economic equation whereby innovation is more valuable than labor. Right. In which case that we might be able to reshore some of those manufacturing jobs back to this part of the country in particular or to any part of this country. I wonder if panelists of any thoughts suggestion that what particular industry sector or what particular manufacturing innovation may help us to rejuvenize a new kind of manufacturing in the US in particular in the Midwest. Not being from the Midwest and not being in the kinds of industries that perhaps are considered to be high technology. I just like to point out that that things are cyclic in nature and that it is difficult to make predictions right especially about the future. And that for example in my area the chemical industry was disappearing from the US. If you look now at the Gulf Coast you will not recognize that prediction anywhere and that came from the unexpected technology of fracturing and the shell gas evolution. That today you know clean energy technology you know affordable takes fuel cells batteries and so on. For the most part we're subcontracted through globalization and for the sake of immediate profits and we're recognizing the tenuous nature of many of those links through the geopolitical issues that we're dealing with right now and those will return and they will retain with better technology requiring more specialized expertise in order to do it. It's not going to destroy the work is going to create work and it's going to bring it back and I think we should prepare for the fact that many of this disappearance and reappearance in general in the US and elsewhere is cyclic and it will come back and it will come back unpredictably and we need to have this insurance policy to make sure that we accept it back into our economy whether it is in the Midwest or elsewhere. Thank you. Well, yes, Marty please. I may I come from a small country that is called the startup nation. I think the secret is right there. In other words, if you think about about work labor and then you will always there will always be countries where they will do what you want to do for cheap labor. There will always be countries like that and there will in the US as a country that pays in our main power costs a lot of money. You can't fight it. You can't really start putting all kinds of customs on taxes and so forth. The only way to fight it is through innovation. That's the secret. In other words, the only way to do that is to come up with new ideas in all kinds of things, not only software in all kinds of things. And what you need to do is actually what you are doing already. You have to bring in the industry, especially the high tech industry, not only the software, but all kinds of them to be near the university. This is really the hope because then the interaction between them will create knowledge, will create new ideas. Eventually manufacturing part of it will be done somewhere else, maybe in Malaysia or Indonesia. You can't fight it. You have to create the knowledge, the innovation. Very well said. It is that innovation power that changes the constraint set rather than struggling within the constraint set. I do optimization for wireless networks and if you are dealing with a distributed optimization problem, ideally you should ask yourself, why can't I just rewrite the constraint set to get to a very different problem in the first place? Well, you know, I just want to quickly respond back to the Mananujan paper there. That's very interesting. Thanks for sharing the link. And this reminds me of a conversation with Eric Schmidt years ago when I was at another institution hosting a panel with him. And I said, this is years ago. I said Google translator. I said, you know, Eric, and he was still the executive chairman of Alphabet at that time. You know, this Google translator doesn't quite work. You know, I put in some words in there. If it's a user manual, you know, how do we turn on the TV? It probably works. But otherwise I said I put in a poem from a East Asian language and try to get it to English and it's not poetic. Now Eric had a very witty and I think telling answer and that is a very simple answer. Nobody cares. Well, he said in a mild way, but roughly nobody cares enough about poem translations to give me enough data point. Had you given me enough data point, I would have made it very poetic. So now it's hard to refute that and maybe it's not refutable now. So to what degree will human creativity or artistic taste or sense of humor and so be indeed replicable by codes? Well, I don't know, but I do want to thank all of you for a really intellectually vibrant conversation. We only have one hour, which we have more. We will have more in the coming years. And I want to thank all those who are chatting in the chat box in real time. Sorry we couldn't get to all of your questions. And to all five new Armstrong distinguished visiting professors to Purdue Engineering. We hope to see you a lot here in West Lafayette. Not only virtually as we're doing this hour, but physically in person. Thank you so much everyone. Thank you. Thank you. Take care. See you next time. Bye bye.