rectal cancer venous drainage

good evening, i'm don levy, i'm the vice president for research for the national labs at the university of chicago. and i want to welcome you here and say i'm delighted that you are here for the latest version of the discovery series program. we designed the discovery series basically to show off activities, scholarship, science, engineering, that go on at the university of chicago to a wider audience, the audience of students, educators, members of the community, and we're actually quite proud of what we have to offer

in general in this series, and what we're going to talk about tonight. i'm told, it says here, for social media savvy... i am not social media savvy, so i have no idea what i'm about to say, but i'll say it anyway. i'd like to ask you to please share your feedback about tonight's program on twitter using the hashtag #uchidiscovery. i'd also like to particularly welcome parents and students from the northern suburbs who are members of maroon kids which is a group that's sponsored by friends and alumni of the institute for molecular engineering i should add that the group is open to all students in grades 6 to 12

and their parents who have an interest in science. so i'm delighted that you've joined us, and i hope you enjoy the program, and you find it worthwhile. at the university of chicago, we have it's an unusual university, in the sense that we have never had engineering, we're 125 years old, and we've never had engineering here. this is not something i point to with pride, i think the university would have been a better university if we'd had engineering from the word go,

but we're a little bit late, but we are correcting the problem. and it's a particularly important time to correct it. about 5 years ago, we established the institute for molecular engineering and everybody you'll see on stage will be members of the institute for molecular engineering. it is more or less been forced into it, the overlap between the sort of science that goes on at the university and the sort of engineering that goes on in the institute for molecular engineering are so tightly interwoven that if we didn't have this marriage, we would be missing a huge intellectual [?], so this was tremendously important.

the institute has been a huge success: we've managed to attract some absolutely outstanding and prominent world leaders in this branch of engineering 5 of whom you'll see tonight. and this is done in partnership, to a large extent, with argonne national lab where a lot of this work also goes on. so the ime's researchers are trying, as you heard in the video, to translate discoveries in basic physics, chemistry and biology into new tools that address important societal problems. and that's where the overlap between the science and engineering goes on.

i think both sides need each other to really make progress. the head of the institute is matt tirrell, who will moderate tonight's discussion matt is dean and pritzker director for the institute for molecular engineering at the university of chicago. he actually wears 2 hats. both hats are so large that everybody wonders how any one individual can support them all. he's supporting 2 of them. he's also deputee lab director for science at argonne. because of the close relationship between ime and argonne, this was very important, matt's a pioneer researcher in the fields of biomolecular engineering and nanotechnology,

he specializes in the manipulation and measurement of surface properties of polymers, his work has provided new insight into polymer properties and new materials based on self-assembly of synthetic and bio-inspired materials. most recently, before coming to the university of chicago 5 years ago, matt was the chair of the department of bio-engineering at berkeley and prior to that he was at the university of california at santa barbara as dean of engineering. you might wonder why it is he left the warmth of california for chicago. but he actually started his academic career as an assistant professor

in the department of chemical and materials engineering at the university of minnesota, where he eventually became head of the department. so coming from minnesota, the answer is, he came to chicago for the warmth. he is the founding director of the pritzker... the pritzker director for the institute for molecular engineering, and has been an absolute force in assembling the molecular engineering and the group of scientists, the group of engineers, that have positions there. so please welcome matt tirrell to start tonight's program.

[applause] thank you very much. and it's really a pleasure to address this crowd on future science small scale, large impact. we could equally well have called this future engineering, small scale large impact. to emphasize that we're talking about two subjects that are always sitting close together in universities and in real life but actually have a little bit different philosophical approaches.

science is about curiosity and discovery, and understanding how the world works. and engineering is about invention, and design, and changing how the world works. but as has been implied, there's a cyclic relation between the two and we're creating not only the first engineering program at the university of chicago, but really the first of its kind in the world. where we really try to create this intimate connection between molecular level science-- physics, chemistry, biology, mathematics-- and build functional systems from the molecular level up.

you've seen some examples already, and you're going to hear about some examples in more detail, but they range from the future of computing to... healthcare, to energy storage... and the point is that engineering is about taking science into society, and doing useful things for society. the institute for molecular engineering has gone from its inception in 2011 as you've seen now to an organization that now has 15 faculty members signed up, i guess there are only 10 of us here right now, but we've grown in faculty size, and of course importantly, in support staff.

we started a phd program 2 and a 1/2 years ago. we're in the midst of recruiting our 3rd class of phd students, and we have over 60 phd students. and maybe most significantly, we've just launched our undergraduate program this quarter, with a major, including about 30 undergraduate students. so there will be, assuming all these students do well, 30 bs molecular engineers from the university of chicago entering the workplace in the spring of 2018.

so this has become a substantial activity, we also have been enabled by substantial facilities. first of all, we have a new building at the corner of 57th and ellis that really has remarkable facilities, and you saw many glimpses of it in the video. we also occupy a significant portion of a relatively new building at argonne, where many of us work part-time. and these things are crucial to what we're trying to do. so you know, we're starting to look like a full-fledged academic unit but we have a lot more to go.

8 of our faculty members, first of all, have joint appointments at argonne, so there's a lot of moving back and forth. but i'd say we're entering some kind of adolescence right now, we've gotten through our infancy. and our pre-adolescence, and are moving into a stage where we have to plot our strategy thoughtfully and carefully, given what we already have. which wasn't something that i was encumbered with 5 years ago. because we're starting from ground zero. so it's in every sense like a startup organization

it's not a startup company, we don't have financial risk. we have reputational risk. all of us left good universities to come here, so we want this to be a huge success, not only for the university of chicago, but for our own satisfaction and reputation. and the thing that's very similar to a startup is that things change roughly quarterly, i'd say. the kinds of things that were working 6 months ago have to be rethought and re-envisioned, and redone.

at least every 6 months or so. so anyway, that's our story, and that's the organization that brings you the program this evening. we're going to have 4 speakers, the same ones that you largely saw in the video. but they'll be able to spend 15 minutes talking in more detail about the areas that they work on. we're not going to have questions right after the talk, but rather after all of the talks, the speakers will come up on stage and hopefully we'll have

20-30 minutes for questions, ok? so the first speaker is david awschalom who is the liew family professor of molecular engineering, and the deputy director of the institute for molecular engineering. david started his professional career at ibm, and then after a decade or so there moved to uc santa barbara as a professor of physics and electrical engineering and the director of the nanosystems institute the california nanosystems institute there. david is a leader in technology known as spintronics and quantum information

technology, and his work has been broadly influential on the evolution of this field which is growing worldwide. and i'd like to invite david up to tell you about it. [applause] thank you very much, matt. good evening. it's nice to see so many students here tonight. i used to be a student. and probably unlike a lot of you, i had a pretty tough first quarter as a student.

not so much because of the coursework, that was ok, but because...well, because i discovered television. and that was something that i didn't have the chance to see too much growing up. and in particular i was captivated by star trek. reruns of star trek, i'm not that old. reruns of star trek. and in particular these interstellar voyages, led by a courageous and emotional captain, a cerebral and somewhat logical first officer, a science officer, and i thought 'finally, a role model!'

but what was remarkable about star trek, particularly to those of us working in this field that matt mentioned, is that it was a remarkable predictor of new technologies. you know, in star trek, even in the late 60s, people had universal translators. they had mechanical devices they could talk to and exchange languages with both aliens and foreign visitors. and today we have these applications on our smart phones. quite remarkable. they had personal access devices where they could take data from all manner of sources and have it readily available on the desktop, and today we have that in the form of

tablet computers. where all of you can access information from the world wide web, trivially, for a few hundred dollars. it really was remarkable that people thought about this. in more recent episodes star trek brought up the idea of holograms where you could have virtual realities. right? you could bring characters from the past and visit them, particularly important characters. and today, well, we have michael jackson.

and we also have virtual realities, of course. there is an expanding presence in our lives. but one of the most appealing technologies, and what i want to talk about this evening briefly, is the transporter. arguably the most exciting technology, where we dissolved into our constituent atoms, sent over vast distances, and then reconstructed hopefully together.

well, that's a work in progress, and that's a bit about what i want to talk to you about this evening. i mean, it really is a work in progress, in some very exciting direction. and it's because when you think about matter, and particularly matter at the atomic scale, on the scale of individual electrons the world of quantum physics becomes very important, and it's quantum mechanics and quantum science that dictates our behavior and our interactions. it's very different than our world today. and it's because the length scales are so different, that we don't see it today.

we're used to human length scales, on the order of feet and inches we think about our blood cells, down to the fraction of a human hair diameter down to chromosomes, then you get to molecules, and dna. which themselves are composed of individual atoms and electrons. and for quantum technology to really make things work, we need to understand the behavior of individual atoms and electrons quantum-mechanically and see if we can build technology based on that. and that technology will be very, very different than what we have right now.

so to put this in perspective, today's technology of course uses atoms, and we actually take them and we assemble them with remarkable precision. we build beautiful lattices of materials, beautiful single crystals. and we're really, really good at fabricating devices that all of us benefit with today. now, to give you a sense of that, if you look at the upper left-hand corner of the screen, that's a single crystal of silicon. it's taller than i am. a couple of feet in diameter.

and they're just beautiful materials. we've gotten incredibly good at this, and they're sliced into wafers, which you see in the upper right, and then devices are fabricated from them in clean rooms that you saw in the video, on the lower left. and they're used in our technology very well. but for quantum technology, we're taking a very different approach. we're going to take a step back, and instead of trying to remove all the defects from the materials

that have been so important to today's technology, we're going to put them back in. we're going to take beautiful single crystals, and we're going to make them defective. and the reason that they're defective is that these defects are missing parts of a crystal, are ways that we can hold individual electrons and play with their quantum states. oddly enough, these are experiments that could have been done decades ago, and people simply didn't look.

so it's the defects that are very important. and those of you who have bought diamonds, for example, are very familiar with defects. different colors of diamond are because there are different defects in there. and in fact diamond is a semi-conductor, like silicon, and you can build quantum systems out of diamond as much as anything. so the theme of the next ten minutes or so is the fact that defects are good, and semi-conductors in a sense are like people: it's the defects that make them very interesting.

[laughter] i don't know how that got there. so we're going to take a beautiful single crystal, and we're going to riddle it with defects. and the picture that you see in the upper right is a semi-conductor that we've actually taken a very pure material filled it with defects, and light is glowing from these individual spots, each one just one or two electrons that we can control

and control their quantum mechanical properties. and it's these types of quantum properties that are going to form the basis of future technology. so what are these properties? matt mentioned in his introduction the idea of 'spin.' and electrons have a charge that we're familiar with today it's how our electrical circuits work, but each electron also has a quantum property called its spin we typically don't take advantage of today. and the spin of an electron is just what you might think.

the electron spins around this axis, just like the earth spins around its axis. or a top spinning on the table which can also process. the electron has this property as well but in a classical system, the earth spinning or particles spinning, the spin can be in one of two states, right? it can spin clockwise or anti-clockwise. or like a bar magnet, it can be up or down. in the quantum world, this spin could be in all different combinations at the same time an infinite number of possibilities.

and you don't actually know which one is it. it's a very, very different way of looking at nature. so what do i mean? i mean classically if you take a single particle, say a little refrigerator magnet, we're pretty comfortable with the fact, right, that it can point up or it can point down. in the quantum world, this little spin is pointing in all different combinations all at the same time, and we don't really know what to do with it until we look at it.

and the moment we look at it, we know what position it's in. and every time we look at it, it can be in a different position. so you might think that's actually pretty annoying. it can be. but it's actually the basis of a very exciting technology, and so this is an example of some data that students take in the ime, just down the street from us:

where they're looking at a single electron looking at it in this quantum super position of states, on the desktop, at room temperature, and that red curve shows this electron's been pointing up, down, and everything in between. it's very, very sensitive to its environment. and again, you might think that's a problem for technology. you don't want things to be very sensitive, right? you walk around with them, you want them to be very stable. but in the quantum world, that's actually a real opportunity.

and let me explain what i mean: it means that if you have one of these quantum states, you have a remarkable sensor. it's a sensor with extraordinary precision, and if you bring it next to a very small object, like a molecule, you have the opportunity to move this around, and get the exact structure of the molecule. something that's very, very difficult to do today. and to put that in perspective, some of you may have seen magnetic resonance imaging systems in hospitals,

where at the human scale they work by taking very large magnets orienting the spins in your body, and then watching them relax. and because different spins relax at different rates, based on their environment, you get an image. and the image you see like this is at the human scale. now imagine you could do images like this at the molecular scale. things would be very, very different.

and here's an example of how much this has changed in just the last 2-3 years, going from dozens of orders of magnitude of resolution down to a handful. by taking these quantum states, bringing them very close to molecules, and determining the structure and functional relationships of these molecules with a quantum probe. and you can see why that would be very, very important in everything from drug design to new materials discovery. we'll talk about that just for a few minutes in just a second.

students here in the department of chemistry, as well as in the ime, have come up with some clever ideas of taking these quantum probes and putting them in very small pieces of material, and putting them in living cells. what you can see on the left is an example of a quantum state in a human endothelial cell that's alive. and this quantum state acts as a type of beacon, looking at the electric, the magnetic, and the thermal fields in a living cell and sending the information out to the observer. and in principle should also be able to watch information moving

in and out of a single cell. and the reason i thought these were interesting examples to show is it would be very difficult, if not impossible, to do this today with classical technologies. now, there's a second quantum aspect of physics that's pretty interesting, and this is sort of the california version of it: matt mentioned some of us were from southern california, "two bodies, one soul." and what that means is you can build a type of state

a piece of information, and put it into a pair, two or three electrons and you can put it in such a way that this information is entangled. you can't read it from one particle or the other, it has to be the pair. it's literally put into both of them. and this has a very peculiar property: once information is put into the pair, you can separate this pair. and you can separate it over very large distances. over a city, over a state,

over a country, and between planets even, it doesn't actually matter. even though they're separated, and there's no physical connection, they're connected. people found this very, very hard to believe in the early days of quantum physics. in fact, even though there's no physical connection, you measure one, it affects the other. i see you don't believe this. that's reasonable.

einstein didn't believe it either, and in fact, when this first came to be, you know, he had a very famous phrase: spukhafte fermwirkung i'm sure you've heard of that. but you've probably heard of the translation: spooky action at a distance. that was literally the translation at the time. people find it very hard to grasp this. it made no sense-- how can you communicate when there's no physical connection?

a student from our group moved on to become a faculty member at netherlands, delft university of technology and very recently did that experiment. he put information into a pair of electrons, moved them over a kilometer apart, in delft and did the measurement. he measured one, which in fact affected the other.

this is reported a few months ago in the new york times 'spooky action' is real. it's remarkable, it's something we don't see in our world, at our length scales but is common at the scale of individual particles. interaction without a physical connection. i see you still don't believe this. ok. so how could you take this type of science and move it into a technology? and one thing you could do is take a commercial material

and paul will be talking about this in a few minutes, materials like silicon, silicon based structures, fill it with defects, like i just talked about, and try and create states where the electrons trapped in these defects are entangled with nearby nuclei of atoms that also have spins. make entangled pairs. so students here have done this very recently, actually just a few months ago. and they built thousands of these entangled spins.

in commercial wafers of semiconductors, something that people really never thought you could easily do, but it turns out people just didn't try. and the reason this is very exciting is that each of these entangled pairs acts as a type of quantum device. and all of these quantum devices serve as a framework for a future quantum machine. it's the quantum machine that would enable you to do models of quantum systems that even with infinite computing power that's available today it's just not possible to model systems like this. you cannot model quantum systems with classical machines.

so these are called quantum simulators, and people all over the world are working on this. and the reason is a very exciting one. if you can understand matter at the very basic levels, you should be able to bring phenomena that we've seen today into the ambient world. for example, room temperature super conductors. superconductors are materials that carry electricity with no resistance. these exist at lower temperatures. if you could bring that

to room temperature, we would be in a very different world. you would be able to convey electricity with no resistance, and distribute power. you would have levitating trains and levitating transportation, because you could generate magnetic fields that oppose each other at room temperature. it would be a very exciting discovery, and in principle, it might well be possible, if we could understand materials at a fundamental level. and perhaps more importantly it would give us options in pharmaceutical design. being able to design drugs whose targets could be well-decided,

and testing done through complete simulation. it would obviate the need to do a lot of tests. you could calculate every possible functional interaction from a drug with a quantum simulator. and this is the type of future technology that a lot of us think about today. so some of you are wondering, probably, where is your transporter? it's a reasonable question. well, quantum information, and this idea of entanglement, offers a pathway to move information, called teleportation. and teleportation is a pretty interesting phenomenon:

if you look at these two purple atoms, as i just showed you, you can build them in a way that they're entangled, even kilometers apart or planets apart. what's also true in principle is you bring a 3rd object into this entangled pair the entanglement acts like a type of superhighway, and will transmit the information instantly to the next particle. it will make a complete copy instantaneously, even though there's no physical connection. you don't believe that either.

but just again, a couple months ago at nist, the national institute of standards and technology in washington that's been demonstrated teleportation over 60 miles. they teleported a particle, essentially between the university of chicago and the university of notre dame. yeah, right, it's amazing. and it worked. so now you're asking again, where's my transporter? there are a few experimental problems here. first of all, it doesn't work most of the time.

so it only, ok, so it only works once every 10 minutes. so we just talked about the number of atoms in a person. right? it's a large number, so to actually move a person from one place to the other, assuming you could do it, you'd literally die trying. i mean, with these estimates of roughly 10 trillion ages of the universe the universe's age is about 10 billion years, so that's a very long time. so it's the sort of thing you would not want to try at home.

but it's a very exciting time for us. i mean, what's becoming very clear for a lot of us, both at the ime, and around the world who work in this field is that we're building technologies now that approach single atoms. and when you build a technology with single atoms, the laws of quantum science determine the behavior. not our classical world. so quantum engineering is very much becoming a reality. we believe that will enable new designs and discoveries of materials

both for practical applications and pharmaceuticals or medical applications and because of the unusual nature of these quantum states, it also provides nearly tamper-proof security for quantum communication. something that's already being driven by startup companies right now around the world. so we believe there are a lot of fundamental physical and biological systems that we can probe and learn more about with quantum sensing and when you put this in perspective, when the field of semiconductor science was just starting,

people never thought that there would be personal computers, global communication, and yet when we think about now quantum information processing what's really exciting is we actually don't know what's ahead of us in the future. so thank you very much. all right, forgot a piece of equipment. our next speaker is supratik guha, who is a professor in the institute for molecular engineering, and also director of the nanoscience and technology division at argonne

he joined the university of chicago in 2015. his work is on semiconductors and oxides and other materials for computing architectures, for sensors and for other applications, including energy applications. he had a 20 year career at ibm before joining the institute for technology at the university of chicago and in the last several years has served as the director of physical sciences research at ibm. and i'd like to invite supratik up.

thanks matt. big data in data analytics today is largely dominated by computer generated data. and that's either transactional in nature, or social in nature. but as we go forward there is general consensus that there will be a large body of data that will emerge that will be physical data. meaning data that comes from sensors that measure the physical world.

so what can we do with this large body of physical data? we could analyze it in the way we analyze computer transaction of data we could run it through statistical packages, and look for patterns. and that would certainly be very helpful. and useful. but we could also take this data, given that it is physical data, and run it through physics models to learn a lot more about the environment

that they're measuring, and give us a lot of useful information. so what do i mean by that? to give you an example, i'll show you a scenery of waves in an ocean hitting against a rocky outcrop. there's a very nice description that richard feinman, the famous physicist gave in a little poem he wrote at the time that he accepted an award many years back, where he talks about these two sort of independent bodies waves in an ocean meeting a hill or a mountain of some kind

and then the two interact to form a surf. so suppose we wanted to predict the formation of this surf. so we could take a bunch of sensors, and lets say these sensors can measure the velocity of the water, and the pressure of the water at a certain point and we took a whole lot of these sensors, so we have a dense number of these sensors and they're measuring the velocity and the pressure of the water and let's say we measured the shape of this rocky outcrop, and its angles and so on, with a great degree of precision and then we took all of this data

and we plugged it into some equations that come out of fluid mechanics and we had a powerful computer, and we solved all of this we could, today, simulate and predict the way these waves would form, or the way the surf would form, right? now this is, of course, not a simulation, it is simply the photograph that i've taken and textured, as you can make out, but you get the point, right? but this is essentially what is called a cyber-physical system, and if we look at it in a little more detail,

it consists of sensors that you can... which are part of a platform and you can screw these sensors in and out like lightbulbs they could be measuring various things they could be satellite data that measures, say, groundcover or agricultural coverage it could be sensors that measures pollution in cities it could be sensors that measures your electrical power consumption as a function of time

and tries to determine by looking at little blips how many refrigerators you might have. it could be any kind of sensors, right? now the data comes into a platform, and then there's now enough advance in wireless technologies that this data can be uploaded with high efficiencies to the cloud, and there's enough computing power today and memory is cheap enough today

that you can then calculate various things using physics models and statistical analytics models to make determinations and recommendations and conclusions that you can push to your smartphone so this is a cyber-physical system. and you can use that information to better your environment or something. so what is that something? but before iget there, let me talk to you a little bit about the limitations

of these systems. these systems today are possible. and largely made possible because the power that exists in your smartphone today is more, the computing power is more than what you had 20 years ago on a graduate student's midsize server. so you can do a tremendous amount of computation today. yet these systems today are [?] they are not efficient enough, they are not cheap enough and the main limitation are the sensors.

they're just not cheap enough, not good enough. but this is how computing was, desktop computing was 30 years from now. and as we go forward in the future, you will see that these systems will become seamless, and they will be ubiquitous in our lives. so where would they affect us? it can have major consequences for things of global impact and i'd like to give you a couple of examples: one in the area of water quality,

and one in the area of agriculture. and i'd like to point out to you the need for developments in sensing technologies in giving you these examples. so lets start with water, and i will go to the ganga river in india it is one of india's 2 major rivers about 1/3 of india's population is affected by the ganga and they live in its basin. the problem with the ganga is pollution.

it is one of the most polluted rivers in the world. and cleaning it up is one of the major objectives of the current indian government. there are about 140 drains, or what are called nullas that feed into the river. and they bring with it fecal bacteria, organic matter, and things like fertilizer runoff and industrial pollution. now just to give you a sense the fecal coliform concentrations in the ganga river run from somewhere between 1,000 to 1,000,000 per 100ml of water.

now the safe, permissible limit for swimming in a lake in vermont is about 300 of 400. so that gives you the sense of the level of pollution. now let us say we were able to measure fecal bacteria all organic matter, using a measurement known as biochemical oxygen demand and today there are techniques for measuring these, but to be able to measure them accurately these tests take at least a couple of days but let's say we were able to make sensors, ok?

that could measure these instantaneously and they were cheap, let's say they cost about 100 rupees, so a couple of dollars then we could be measuring this river with high density, with high resolution and this data, if it could be sent cheaply up to the cloud it could be analyzed and we could use things like computational fluid dynamic models and so on not just to predict the flow in future, of these contaminants, but also to do things like inverse modeling to back-predict and try to find out sources of contamination.

say it could be a sewage treatment plant that wasn't working right or say it could be an industry that was not following regulations for dumping pollutants. so there could be so many different things that could be done if only we could make these systems, these sensing systems cheap enough so that they could be usable worldwide. so water quality is a major area where sensors and sensing systems and cyber-physical systems will make a huge difference.

the second example is one of agriculture. now as many of you are probably aware, water is overused in agriculture. ok? it's not measured accurately, farmers kind of guess as to how much water they need to use, and this is a global phenomenon. about 70% of fresh water consumption in the world is due to agriculture now this is an experiment that scientists in ibm

and yellow wineries did and i was fortunate to have played a small role in this experiment while i was at ibm. so the role of this experiment was to try to find out if one could improve yields and increase the water efficiency, and therefor reduce the usage of water. so as a test bed, about a 10-acre plot of vineyard in lodi, ca, was chosen this vineyard was...

we divided it up into little tiles that were about 30 x 30 meters and we rigged up the irrigation in such a way that the water delivery could be individually controlled to each tile. so now during harvest season the team regularly monitored satellite images of this control region and these satellite images are free, they're available from the us government you can open an account and download it

it's available at 11 color bands, and you can use this data to basically look at the greening of the vineyard canopy. basically look at a chlorophyll map. so then these scientists analyzed this chlorophyll map they got climate data, weather data, etc. and then they calculated exactly how much water should be delivered to each of those little tiles. and when they did that, over a two year period two harvest seasons

they found that the yields and water efficiencies went up by about 10-20% and this was huge! this was a wonderful result because conceptually, all that was done was very simple we used free satellite data to calculate the amount of water and then dose it accordingly. now let us say we were able to have sensors that could go into the ground, that measured things like dissolved nitrates that measured things like soil moisture or water potential

that measured things like plant disease then we could similarly calculate much more accurately the amount of water fertilizer and pesticide that would be needed and we estimate that could be savings of up to 2-5x savings of these resources. so what is the problem? a lot of these sensors are available today. but they are again very, very expensive. they cost a sensing platform today, here will cost you somewhere between $800-$2,000.

so even for a high-value crop like wine-grapes, you cannot afford to put more than maybe one sensing platform or maybe a couple every acre. but you need them to be there about 10 meters apart, because that is sort of the length scale of soil variability. so these sensors need to come down in factors of about 10-100x in price for them to be usable, but once that happens it will change the agriculture. so again, it points to the need for sensors.

so how can we improve these sensors, make them more cost-efficient, and more power-efficient? and the answer, i believe, lies in the area of nanotechnology. nanotechnology has been around for about 25 years or so and a lot has been learned in the field, but i believe sincerely that the calling card of nanotechnology will be in what it does for sensors. ok, and the reason is very simple:

over here, the picture you're seeing here is a nanoparticle of platinum nanomaterials are essentially very small chunks of materials that are just a handful of atoms, maybe, you know, 10s of atoms or 100s of atoms or at the most maybe 1,000s of atoms. and if there's one message i'd like you to go home with it's that nanoparticles have many different properties but the most important property

is that its uniqueness lies in the fact that its property is determined by the environment that it is in. for instance, if i take a bulk material, if i take this table, its property is determined by the way its atoms are bonded, and so on. but when a particle is very, very small how it behaves is determined entirely by the environment it is in. and this is important when you are trying to make a sensor, because you want something that senses the environment. and that is why these materials are ideal for sensors.

so how would we make a sensor out of these nanomaterials? well, they interact in different ways. with light, with magnetic fields, as you saw in david's talk with pressures, and let me just give you one small example here again going back to water, and in this case, let's say, determination of dissolved oxygen in water

so this is a technology that is already there it's relatively new. you can buy products. it essentially is using something called a fluorofor, which absorbs light, in this case it's absorbing blue light. and it emits red light. now if there is dissolved oxygen in water, and if this fluorofor is dipped in water, then this oxygen quenches, or kills, this light.

now you could buy 50 cents worth of light emitting diodes and detectors and therefor very quickly measure this level of quenching or killing of the fluorescents and determine how much oxygen is there. this is a very simple idea, it's a product, it's one of the early products of what i would call nanotechnology and you will only see these types of products improve and become cheaper and more efficient over the years. so where is the future headed?

and for this i would like to give the example of the boquila vine. this is a vine that is found in chile and it is a very interesting plant when this vine grows on any tree, the leaves of this vine try to mimic the shape of the tree, of the leaves of the tree that it is growing on. now this is amazing, because the vine cannot see in the common way that we define seeing. so obviously it is sensing something in a distributed way

and then changing the shape of its leaves. this is sort of the ultimate cyber-physical system. it is adaptive-sensing, it is distributive sensing, and in many ways this is the early form of intelligence. we look at today's cyber-physical systems and push it out, you know, 20, 30, 40 years from now, that's where our systems will be, they'll be distributed computing systems, and you will start seeing the first things of what we would call intelligence. with that i would conclude my talk. thank you.

thank you very much, supratik. our next speaker is paul nealey who is the brady dugan professor of molecular engineering. paul is a pioneer of a process that is being increasingly evaluated and used in industry for fabricating integrated circuits and storage devices known as directed self-assembly. he'll tell you a lot more about this, but the idea of self-assembly is that one uses a kind of information content built in to the molecules so that they self-organize into the patterns that you want to build

in integrated circuits and storage. paul joined the faculty for the institute for molecular engineering in 2012 having previously been the schumacher professor of chemical engineering at the university of wisconsin. let's welcome paul to the stage. [applause] so good evening. it's a great pleasure to have the chance to tell you a little bit about our research here at the institute for molecular engineering.

so all of you should be aware, especially if you're over 40, or 45 years old that you're actually living through a technological revolution. we call it the digital age. so there's some images up here on the screen of products that you're familiar with that you use everyday. but you need to think about the way that these products have changed the way we live. the way we communicate with one another, the way we generate, store, save, use information

and how it's brought the world to be more a global society. so another way to illustrate this point is to show you this advertisement from 25 years ago in 1991 from radio shack of all these electronic devices that were available for sale and all of those devices are now contained in your smartphone. amazing, just 25 years ago. so really revolutionary, and i think it's worthwhile taking a few minutes to think about how different your lives are now than they were 25 years ago, because of these developments.

so there's less obvious things that have come out of the availability of faster and more powerful integrated circuits this is a picture of mira, which is a supercomputer at argonne national laboratory. and there's so many things i could mention for how its changed our world, but this was a list i found on the web, for the 9 super cool uses for supercomputers. so evaluating, predicting earthquakes, for example. recreating, or understanding the origins of the universe, etc.

so again, all of these developments, this technological revolution that we are experiencing, is directly related to being able to make and fabricate more and more devices called transistors on one surface. and this is known as moore's law, it's really an observation that he made in the late '60s and early '70s that it seems as though the number of transistors that you could make on a cpu, on a computer chip, was doubling every two years. and so if we think about that ad for 1991,

there are approximately 1 million devices on a computer chip and in 2016, 25 years from that date, there is now 5 billion. ok? exponential growth over that time period. and of course that's what's responsible for being able to do all of these great things, even that we've heard from david and supratik. ok, so to bring that home a little bit further in the manufacture, here i show actual images of devices at relative scale for different years. and so if you're going to double every two years, that means you're going to quadruple

the number of devices every 4 years and if you think about it, you shrink the area of a device by 2, that means you can fit 4 devices where you could only fit 1 device previously. so if you look at the images for 2004 and 2008, you see that they roughly shrank exactly by a factor of 2. ok so i want to explain to you a little bit more about how semiconductor devices are manufactured, so you can understand our research a little bit better. so i included a few cross-sections of devices here of modern computer chips.

the devices that i'm talking about are weighed on here, and they can't even see them, and then these are all the layers of wires, or interconnects that are made to make the devices function. and so just from looking at one of these cross-sections you can see, one, that you're patterning at very small length scales, but that you make these devices in a layer-by-layer fashion perhaps 25 to 30 different layers, and so each layer has to be made so that it connects properly with the one under it and connects properly with the one on top of it.

so you have to align these patterns perfectly, that's called registration, and then if you think of the numbers of layers, if there's even 1 defect in any layer the device doesn't work, and so perfection is absolutely necessary for these devices to be made. so again, iwant to get down to the actual process of making this device, so let's look at, say, one of these layers of wires here that are called interconnects and these are wires of copper that are separated by an insulator a dielectric material and to make just that one layer,

would require this series of steps: we'd deposit a dielectric on top of the device layer, we'd coat that with a polymer called a photoresist, and now i want to explain to you the process of photolithography we'd expose that photoresist to light, and where that polymer is exposed to light it changes chemistry. and then if we bake the material a little bit, it changes chemistry a little bit more, and then if we plunge this material into an aqueous solution, basically water, that exposed material will dissolve much more faster than the unexposed material

and that allows us to pattern the sacrificial layer. then we could deposit copper everywhere polish off the copper where we don't need it and we'd arrive then at our one layer of copper wires in this structure. so then to make the overall device, we simply repeat that series of steps again and again and again maybe 25 to 30 times

to arrive at our fully integrated circuit. and so obviously, again, to reiterate, you need to be able to do this process at very small length scales you have to be able to position each pattern perfectly that's called registration, and you can tolerate no defects. ok, so why can't we simply make these... oh, one more bit of perspective if we think about, david showed you these wafers and these ingots,

they're now either 300 mm or 450 mm in diameter, and we can pattern hundreds of computer chips on just one of these wafers so if each computer chip is 5 billion devices, and there's a 100 of them on the wafer we've patterned more devices on one surface than there are people in the world by a factor of 100. and my colleagues at intel tell me this is the marvel of the engineering world that there are times when every one of those devices is perfect. can you imagine patterning that many perfect things all at once?

it's unbelievable. ok, so why can't we just continue to make smaller and smaller patterns? the problem is that as you do this exposure of the resist, the photoresist through the mask, you can't simply keep reducing the slit, or the aperture that the light goes through, because as it approaches the wavelength of light, it undergoes a process called defraction. so you could make the slits smaller, but you would not be able to make smaller features. ok? so for many, many years,

it was very easy to go to smaller and smaller devices simply by reducing the wavelength of the light so that you could get the smaller apertures and smaller features. but that strategy petered out in about 2000-2005. so instead i'd like to pose a question to you: what would you do to go to smaller features if you had a tool that could only give you, say, 80 nanometer periods but you're at intel and your boss says that you need to make 40 nm period structures a very difficult challenge.

it turns out the concepts for doing this are not so difficult. so here is one strategy: oh, sorry... called litho etch litho etch, where you'd simply do this process twice and you'd have a machine that could so perfectly move that wafer back into position that you could put a 40 nm line right in between the two 40 nm lines at 80 nm pitch. that would be called litho etch litho etch. that's actually in production. the second strategy would be what's called self-aligned double patterning

turns out that you can very controllably conformly coat structures. and control the thickness of the structure very, very well. so in this strategy you conformly coat that first resist structure, and then use the conformal coatings on the two sides, ultimately to define the pattern at double the density of the original pattern features. ok, so finally, to get to our research, we've been developing a process called directed self-assembly and here we have the same tool that could give you, say, at 80 nm pitch pattern but instead of doing those other processes that i just described,

we use the tool to make a chemical pattern. and then the idea is, could we simply deposit a material, a magic material that responds to that chemical pattern that would assemble into these structures at higher density than the original pattern to allow you to get to smaller and smaller features. and this is the principle of self-assembly. and this would be extremely useful, because not only could you use it to double the resolution for example, of the pattern, but you could potentially use it to triple and quadruple and even to go to higher factors of density multiplication.

ok, so the magic materials that we use, these self-assembly materials, are called block copolymers. there are two kinds of polymer chains connected at one end by a covalent bond. the two polymers aren't that... exotic. one is polystyrene, like the cup that you get on an airline and the other is pmma, the plexiglass that's in your screen door. but if you make two molecules, and you bring them together, and bind them at the molecular scale, they spontaneously form these self-assembled structures

at very small length-scales, at the molecular length scale. and in fact they make the structures that you see, different shapes and very uniform sizes, at the length-scale of 3 to 50 nm, which is what's very difficult to achieve in the process i was just describing by traditional lithography. so does it work? the answer is yes. so here's one of those magic materials that wants to make line and space patterns lamellar structures

on one side is where we haven't told it what to do, we haven't directed its assembly, and it makes this random, what we call a fingerprint structure and when we give it direction on a chemical pattern, you can get it to perfectly assemble into these line spaces that are then useful for manufacturing. to give you a little flavor of the magic of this assembly here are some molecular simulations from my colleague wanda pablo: that shows you a very different process than what i described as photolithography, where now you're depositing this material and letting it spontaneously assemble with some direction

into the structures that are then useful for intended applications. ok, so the biggest challenge in implimenting this strategy in industry has been to show that you could reach these incredible levels low levels of defects that are required for manufacturing computer chips and to give you an idea of how incredible a constraint that is, you need to be able to pattern perfectly, and only have one defect in a 10x10 cm area so if we were actually to take images where we could see the structures that we make, we would have to look at 1.6 billion images and find out whether there's not more than just one of those defects

in any of those 1.6 billion images, that's the level of defectivity thats required. well it turns out we had to ... this is one of the hallmarks of the institute for molecular engineering is finding the tools and the skills that you need. the actual first graduate of the institute, paulina rincon delgadillo, was one of my students that went to imec, a consortium in belgium to be able to utilize the specialized tools to find that answer: could we reach the defect levels that are necessary using that strategy that i described? and the answer was yes, and here are some images then of...

these are 14 nm lines and spaces, or actually these are 10 nm lines and spaces so ahead of what's necessary for manufacturing at present. perfect, and over the areas of, that i was describing as needed. so to give you one more perspective of scale, these 50 lines and spaces are about a micron wide, and if i were to compare that entire 50 nm line and space to the width of a human hair, it would be just that little spot that you can barely see as it shrank to the length-scale of the hair. ok, so our current research is to understand the 3-dimensional structure of these assemblies, so unlike traditional processing, they can be 3-dimensional

and actually the position of these lines and spaces can change as you move along the structures. and so this is utilizing specialized tools at argonne national laboratory to do what's called transmission electron microsco-tomography to image these structures. we're obviously hoping to go to much smaller length-scales than even what i've showed you this is our best effort to date where we quadruple the resolution of the original lithographic tool and are assembling 8 nm features on a 16 nm pitch.

so i'll leave you with one final thought. this technology that we've developed is actually relatively mature and is actively being commercialized in the context of semiconductor manufacturing but now we're beginning to apply these concepts of directed self-assembly in other areas these same, you saw that the chips were 3-dimensional, one question we ask ourselves is can we use this same strategy to assemble 3-dimensional structures for this type of manufacturing? a second question would be could we use these strategies for materials for energy applications? the answer turns out to be yes.

where we make ion-conducting materials for membranes that are used in fuel cells and batteries, for example. and then the sensors that supratik was talking about we use chemical patterning and nanoparticles to create structures that you can't make otherwise, and these assemblies of nanoparticles have incredibly sensitive sensing capabilities, even single molecule sensing. so with that i'll stop, and turn it over to the next speaker. thank you very much paul our fourth and final speaker is melody swartz, who is the william ogden professor

of molecular engineering. melody's work will be a significant change from some of the digital technology things that one has been hearing about melody works on understanding the lymphatic system and transport in the lymphatic system and its role in immunology, or in immunity, i should say. and how one can address the lymphatic system to attack all cancer and other immune disorders. melody joined us in 2013... 2014, actually.

having been for a decade before that the professor at l'ecole polytechnique federale de lausanne in switzerland. melody, please join us. thank you. thank you, it's really a pleasure to be here. as matt said, this is going to be a big departure. i am often asked two questions when i tell people i work on cancer first is, why is that molecular engineering? and the second is, why haven't we cured cancer? i mean, think about all of the billions and billions of dollars and effort spent in the last

several decades on cancer research, and we still don't really seem to have a cure. what i'm going to talk about today is immunotherapy. this is potentially a cure for cancer, it holds, many people believe it holds the key to curing cancer but this is something that we've been working on, and we still don't know enough as many other people working on it to actually use it to its full extent. despite the promise of new immunotherapies that are coming out, cancer is still really winning. as you all know, cancer is one of the biggest killers of people

that are not accident-killed, and this actually claims hundreds of thousands of people's lives every year. deb was my friend since childhood and had breast cancer. she thought she was cured. a few years later had some headaches, and found, they found cancer having come back in her brainstem. and she was, embraced her fate, as you can see from this tapestry she made, and succumbed to it just last year. many of you, if not most of you, know people, friends and family

who have been... had their lives cut short, early by cancer. now that's not to say we haven't made a lot of progress, there's tons of progress with all these billions of dollars we've been spending. here's a list of some really incredible milestones in cancer research that have made a huge difference. some of these are treatments, going from surgery a long time ago when they finally discovered anaesthesia radiation more than 100 years ago to kill tumor cells and chemotherapy, the first of which was mustard gas.

most chemotherapies, up until recently, killed rapidly dividing cells that was a hallmark of cancer. more recently, very targeted drugs are being developed to kill cells that carry a specific mutation that was known to be existing in particular cancers. so these are becoming more and more patient-tailored, and more fine tuned. ok, here's some pictures. some of these milestones have come in early detection. a lot of this is in imaging techniques, but also some biochemical assays.

if we can catch cancer before it becomes really obvious, we might be able to treat people at an earlier stage, and get them before metastasis has occurred. to treat the cancer. ok, so this has actually caused a huge change in mortality rates because it's able to catch the cancer before it spreads. and then of course there's been a lot of technologies in better prevention, starting with, i would say one of the first... i remember this when i was a kid the dragon lady, i don't know if you all remember the dragon lady, but really, you know, starting with saying ok

hey, smoking causes cancer it took, actually, decades for that to become a public... to get out in the public, to the demise of the tobacco industry, but they kept finding new targets. but that actually helped change the actual rates of cancer death by preventing it from happening in the first place. similarly, gene testing. genetic testing for certain mutations that are known to cause cancer, so for example the brca mutation in breast cancer, that leads some people, like angelina jolie for example,

to get a mastectomy preventatively because the chances of getting the cancer will be very high. and then finally cancer vaccines are starting to come in the picture. there are a few cancers that are known to be driven by viral infections. one is hepatitis b driving hepatocellular carcinoma another is human papillomavirus driving cervical cancer. and so we know that cancers arise from chronic infection and so the idea in cancer vaccines in this sense are really preventive by not preventing the cancer, but preventing the viral infections

from ever happening in the first place. so a person would never develop cancer. and in fact, if every single kid today got vaccinated for hpv, we would just eradicate cervical cancer, it wouldn't exist anymore. it would be like smallpox. so all of these things have had an impact. childhood cancer deaths have gone down over the years, even though the incidences remain the same, or even increased. some cancers in adults you can see have actually decreased a lot.

these are cancer deaths, not incidence rates on the right. like stomach cancer. but you can see here, this is from 1930: lung cancer you know, there was a lag, because after the campaign, after cigarette companies were required to put smoking kills on there, it took another couple of decades to see the change in cancer incidence. but you can see it's leveling off, and in men it's decreasing. but still, you see a lot of these cancers have not changed in their killing rate of people like pancreatic cancer, liver cancer, colon and rectal, everything.

there's so many cancers that continue to have had very little impact from treatment. so why haven't we found a cure yet? is it just a moving target? are we just shooting at a moving target? well, the answer is, of course we're shooting at a moving target. cancers are very complicated, evolving, heterogeneous monsters that evolve and grow very rapidly they're not just one type of cell. they're a whole society of cells

we can develop drugs to kill some of those cells, for example, in a lot of chemotherapies you might only kill the cells that have a certain mutation, but then there's always going to be a few cells that don't have that mutation. they're going to come back and form a new tumor. you might kill cells that are rapidly dividing, but there are quiescent cells that just sit there they wait out the therapy, and then they form a new tumor. or you can kill only part of the tumor, but not all of it,

because it's not fully effective. there's just so many ways that the cancer can escape therapy. in the immune system, actually does take care of a lot of killing cancer, normally. we have immune surveillance that recognizes when tumor cells, when cells become tumor cells, when they become neoplastic. and they actually come and attack these tumor cells. so this happens all the time, that's why we,

people who do not have cancer, don't have cancer: because this is taken care of, as it should be. when tumors become cancerous, then they have escaped immunity in most cases. and they can escape immunity in a lot of different ways. ways in which we're still trying to understand. they can hide from the immune system. they can mask the things that make them different, so that the immune system doesn't recognize them.

they can take the immune system and actively suppress it with chemicals that they secrete. or, they can even directly kill t cells that are coming to kill it. so they can even do direct warfare with your immune system. so the whole hope of immunotherapy is to activate the immune system to kill cancer. so even though we know, we're learning more and more every year about how the cancers escape immunity, we're still trying to turn these cells around so that they can do what they're supposed to be doing,

in killing these tumor cells. the reason it's considered a potential cure is because once you get this immunity then any future cancers that pop up, any metastases or silent cells that have been dormant for a long time, they come back, your immune system will just take care of it again. but in the current immunotherapy trials, there are hundreds now we still have very heterogeneous response some patients respond, but many don't. and we still don't fully understand why.

but we do know that the immune system is a lot more complex than we appreciated even 10 or 20 years ago. you can look at the most basic sense of self and non-self our immune system has to be primed to be able to kill pathogens like bacteria, viruses, worms, protozoa and ignore self things, like your own cells and your own proteins. but there are so many things that we have to also ignore that's foreign. food. gut microbiota.

and things in the air, like pollen. at the same time, our immune system has to be sometimes able and ready to kill your own cells. when they're infected, and when they're diseased, like as in cancer. and so the immune system actually has so many mechanisms to try to balance these two things. when we take immuno-suppressant drugs, like to prevent graft rejection or something, we can get cancer.

because the immune system is suppressed, and it gets overly permissive in a totally generic way. when we do the opposite: if we try to over-activate the immune system and this is one of the side-effects of immuno-therapy we can develop auto-immunity, because your entire immune system is a little bit hyperactive. so your immune system, when it's functioning in a healthy way, you really have to balance these two things. now my lab has been interested in the lymphatic system for a long time and we always wanted to understand why the lymphatics

are involved in tumor metastasis. why do breast cancers and melanomas always spread first to the lymphnodes? and we know that the lymphatics do expand around tumors we weren't asking about immunity when we started working on this 20 years ago. but the questions we've been asking, and the more we understand about this lymphatic system that's very poorly understood, the more we realized that actually it does play a central role in helping the immune system balance and make decisions

about immunity and tolerance. and tumors actually hijack that. so what we've discovered, and this is very bird's eye view, we've really discovered that one of the main roles of the lymphatic system and lymphatic vessels in particular, are to help instruct the immune system to distinguish safe vs. dangerous they're constantly exposed to everything that they're draining and they take that information and instruct the t cells what to do with it. because of this, tumors can hijack that.

they can hijack the function of these lymphatic vessels to promote immune tolerance and escape the host immune system. and we're still discovering all the different ways in which it can do that and there are multiple ways. so if you think about the lymphatics, we usually think of them as transporters they're transporting... not the same kind of transporters that dave talked about they're transporting all of these secreted factors the tumors are secreting all kinds of things. they're transporting it all to the lymph node.

they're taking it up themselves, they're giving that stuff to immune cells in the tumor micro-environment so they're taking the specific information, ok, this is a blue tumor cell. this tumor cell is... whatever. you know, specific information about the tumor and then telling the immune system what to do with it. unfortunately, because of the tumor hijacking this whole system the lymphatics are telling the immune system 'hey, that's totally normal it's safe, just leave it alone.' so half of our lab has been working on really trying to understand

the role, the basic immuno-biology of the lymphatic system, and what it's doing in tumors and in general but the other half is trying to work on ways to manipulate that as engineers. to target that. normally this lymph node, for example, is cut out when there's surgery because it often can contain a metastatic cancer and so our idea is to, instead of cut it out, to actually target it to use it to its full advantage it is immune suppressed, it's allowing the tumor to grow,

but it has all that information. it's an information goldmine. meaning it has all the information about the tumor, the specific details of which proteins are being expressed and secreted. the only bad thing it's doing is telling the immune system to ignore that information. so if we can instead target danger signals to that lymph node, we can maybe turn those t cells around, and get them to actually go kill the tumor instead of ignore it. so to do that, we've been working in collaboration with the hubble lab on nanoparticle vaccines.

these nanoparticle vaccines are designed in a lot of different ways so that we can really control the immune response. they are designed to have surface molecules that look like viruses or bacteria. and so the immune system, when they see these things, they really think that they're actually danger signals. they're dangerous. and so they start this program of ok, this is a pathogen, we have to attack it. and then they can go and actually kill the tumor. so by targeting that particular lymph node,

we can really get a big response. and to show you some data on a mouse, this is a very aggressive melanoma we implant it, and within two weeks, in white here, you can see the tumor grows very fast, and it kills the mouse if we give our vaccine to an uninvolved lymph node that means a lymph node somewhere else that still has t cells but it's not rich in all that information from the tumor. we see that, so this is in the contralateral side, the opposite side

it takes at least 5 to 7 days before those t cells actually get up to speed because they have to get educated from scratch. they get educated from scratch, and then they start to kill the tumor. but because of that time, and there's a little bit of tumor cell killing by the t cells in that time, the tumor can kind of re-adjust and mutate and escape that immunity, and they grow back. instead, if we target the tumor draining lymph node with that same exact vaccine we can see a very quick response, within 2 days the t cells are starting to kill that tumor and they kill it completely, and it never grows back.

so in this way we can take basic information about what the lymphatic system is doing, and how it works, combine it with molecular engineering and nanotechnology, and come up with a completely new concept in treating potentially treating cancer, or aiding immunotherapy. another approach we're working on is taking the fact that the lymphatic endothelial cells themselves are actually instructing the immune system safe signals so these are basically inactivating, or suppressing t cells normally. what we're doing is trying to deliver, with protein engineering approaches and nano-scale material approaches, we're trying to turn those lymphatic endothelial cells

into activating dendritic cells that can activate t cells to be killer t cells instead. so it's again similar. same information, you just do something different with it. so to summarize, i think cancer immunotherapy holds enormous promise. it could potentially cure a lot of cancers, if not all, someday. but it still has a long way to go in fulfilling that promise. there's a lot of things we still don't fully understand, there are a lot of things we need to do to make it effective. by looking at one part of how a system like the lymphatic system can help regulate immunity, and at the same time designing therapeutics

that target those, we think that this could be a way to potentially facilitate cancer immunotherapy much better. and i also wanted to point out that while it was mentioned in the beginning that engineers take basic science and translate it into technologies that are used, and this is often the case, there are also many examples where engineers also use technologies to understand basic science and that's the case in this work here. so i thank you for your attention, and looking forward to the questions. please come up, i'd now like to invite all the speakers to come up

and we're going to take questions. so are there any questions? q: dr. swartz, i was wondering, i saw that you mentioned the nanoparticles that mimic parts of viruses and bacteria, why is it that you use the nanoparticles instead of just, maybe, the proteins from bacteria or viruses directly? is there an advantage? are they easier to make? or are they more effective in some way? well, if you use a fully synthetic system, you can control it completely and so you can put exactly the molecules you want in exactly the right concentrations. and we can make, the key here is size, as well.

size surface chemistry, things like that. the size is where you can control where you're targeting. so by making them ultra small, like 20 nm, 30 nm, we can really target lymphatics very effectively. q:thank you. q: hi, i had a question about super computing. so in 1997, ibm created deep blue and it beat the world's best chess champion. and recently, google's deepmind alphago beat the world's greatest go player which is a feat that many people said wouldn't happen for many decades

because of go's complexity. what struck me as interesting was that many analysts said that the go program was not just following certain algorithms but it could think on its own, and had a certain level of autonomy given this recent milestone in super computing, how far away do you think ai being self-aware would be? that's on you, supratik. excellent, thanks. yeah, there was the chess game, and then don't forget the jeopardy game

that the ibm computer beat, and then there's the go it's a tough question, and i don't really have a quantitative crystal ball here with which i can tell you exactly when we will assume a level that our computers will assume a level that we will say is intelligence because that itself is a kind of a vaguely defined line. but what i will point out is what you need to compare is how much energy do we spend in our computer when in plays against, let's say, a human. see, a human brain, right, roughly consumes about 20 watts some people a little more, some people a little less.

[moderator: mostly more in this room] now if you look at the machines, i don't know about the go program, but if you look at the machine that beat the humans at jeopardy, that consumed about 80,000 watts, i believe. so it's a huge factor more, and that, i think, is where the innovations need to happen that we need to look at new forms of computing that are more adept and more power efficient at solving problems that require intelligence and thought and qualitative decisions rather than high-precision mathematics, which is what our computers were originally designed for.

q: thank you. q: this is a question for each one of you. why are you working here instead of somewhere else? what is it unique about here that makes you want to work here? [laughter] let's start with paul. boy, there's an awful lot of reasons to mention, but this was an opportunity that was unlike any other that i'd ever heard of, and have not heard of since. so to start something completely new, to think about a new paradigm

of engineering research and education, was just an opportunity that was impossible to pass up. and the level of enthusiasm of the administration and the fellow faculty and other departments and areas of the university was unprecedented. so it was a very easy decision. yeah, i mean, i would echo that sentiment, and also point out that a lot of the areas you've been hearing about today, the distinctions between science and engineering are really blurred. and so to make progress, you also need to train students in a very different way.

and have students work together in a very different way, and there are almost no universities i'm aware of that are starting programs like this where the distinctions between normal disciplines are gone. so in a way it's a pretty exciting experiment, to try and train students to work very collaboratively to attack very big problems in very different paths. so for me the reasons were very simple, really. and i went through this process last year, because that's when i joined the university of chicago. one reason was just the quality of the colleagues and the diversity of their professional experience this really appealed to me.

the second was, i'm also a joint employee at argonne, and the proximity of the argonne labs, and capabilities that are there, and the colleagues at the argonne labs there are very few places where you can gain from the strength of a university and a large, major doe, department of energy national laboratory.

Drop 5lb of belly fat in 3 day [1-2lbs PER DAY!]

What if I told you there is one simple tweak you can implement today that will force your body to get rid of stubborn, unwanted belly fat everyday for the rest of your life... Would you be interested in that? Good news is...



You can literally lose 1-2lbs of belly fat every single day with this!



This is a new diet “trick” from a european doctor that can help you lose 1-2 lbs of belly fat per day! So...

Check It Out Here And Lose 5lbs