 Good afternoon. My name is Lisa Garcia-Bedoya and I am Vice Provost for Graduate Studies and Dean of the Graduate Division here at UC Berkeley. It is my sincere pleasure to welcome you to the 2022 Berkeley Grad Slam Competition. Today's event comes on the heels of Graduate and Professional Student Appreciation Week, a week that seeks to emphasize the contributions, impact, and value of graduate and professional students on campuses across the United States. What could be more fitting than to follow that week up with Grad Slam, an opportunity to celebrate and make visible the outstanding research done by our graduate students? Grad Slam was launched eight years ago by former UC President Janet Napolitano with the goal of sharpening graduate student communication skills and providing an opportunity for the public to learn more about the amazing research being undertaken by graduate students across the 10 UC campuses. This program also enables us to demonstrate our commitment to inclusion in that it asks graduate students to present their research in jargon-free language, especially as a public university. It is our responsibility to communicate our research in a matter that is accessible to a general audience and not just to academic specialists. Here's how the UC-wide Grad Slam Competition works. Each of the 10 UC campuses conducts its own campus-wide competition and selects a finalist to compete in the system-wide event, which will be hosted this year by UC President Michael Drake on May 6. That competition will be live-streamed and we hope you will all log on and join us again to cheer on the 10 UC Finals. Here's the process we use at Berkeley to determine who our campus finalist will be. A faculty subcommittee of the Graduate Council reviewed videos submitted by graduate students from a range of disciplines and selected the 10 semifinalists whose presentations you will hear today. Let me take this opportunity to thank the members of the advisory committee for graduate student and postdoctoral scholar professional development for all of their hard work. We know there were many worthy submissions. Today's presentations have been evaluated by a distinguished panel of judges. I would like to take a minute to thank each of them. First is Wendy Chakuta, an award-winning TV journalist, speech coach, media consultant author, speaker, and naturalist. Second is Eric Stern, who serves as senior vice president at Capital International Investors and also is the chair of the Graduate Division Executive Advisory Committee. And finally, Fiona Doyle, distinguished professor emerita of material science and engineering and who was the former vice provost for graduate studies and dean of the Graduate Division. Our sincere thanks to each of you for volunteering your time to help us with this event. In addition to the panel of judges, several years ago we introduced a new feature to the competition. You, our audience members, will also be judges. You will be invited to select a People's Choice winner. I will tell you more about that later. Now I am very pleased to introduce our keynote speaker, Adelaide Bernard, who completed her PhD in metabolic biology here at Berkeley last spring. Not only is Adelaide last year's campus branch SLAM champion, she also took home second place in the UC system-wide competition. Knowing that she would be unavailable today, Adelaide kindly recorded a message for us to share her reflections on the value of the grad SLAM program. Hi everyone. I'm so honored to be given this opportunity to speak about my experience at the grad SLAM last year and to remind us how important it is as graduate students to make our research accessible to the world. I want to thank the Graduate Division grad SLAM team for organizing this amazing event. I'm super excited about this year's competition. When I started my PhD, one of my peers participated in a nerd night during which he talked about his research on type 1 diabetes in a bar in front of an audience that was enjoying their beers. And I was amazed by how he was able to present all those concepts in a fun and interesting way to a lay audience in a casual setting. I did not speak English very well at the time, but I thought to myself that it would be such a cool challenge to participate in something like this when I felt ready. And at Berkeley I first heard about the postdoc SLAM and then learned that there was also a graduate student version. So although it was very intimidating, I decided to give it a try. I sent my video to the grad SLAM committee and as a non-native speaker I was already amazed to be selected as a semi-finalist. You can imagine my astonishment at last year's event when Vice Provost Garcia Bedoya announced that I had won first place. I was really honored and I have to say I was so grateful to the grad SLAM team and in particular to Wendy Tokuda who told me so much as I prepared for the system-wide competition. Their critical feedback really helped me in perfection in my talk and getting ready for the system-wide event. As a scientist it's easy to become obsessed with the nitty-gritty details, but when it comes to explain what we do to the public, we have to first use more general concepts to frame our research. And second, if there's something learned during the grad SLAM process is that less is more. We have to let go of the jargon and find the right words that won't hurt the data. And as the research is ours, we have this privilege of being able to distill the data and our conclusions for the public, making sure we put out into the world information that is although simplified still accurate. But why should we simplify science? This gets me back to the fact that as graduate students or scientists or academics, it's our duty to make our research accessible. It's clear that there's a huge gap between scientists and society and that gap can be bridged by good communication skills. A good example is what happened with COVID over the past two years. Suddenly everyone on social media became a virologist or an epidemiologist. And nowadays when people want to be informed, they don't go to the library, they hit the YouTube button and watch whatever video shows up first. Everyone strives for these shorter, quicker bites of information and that's not the format in which we make our research results available. And I'm not saying that we should all become YouTubers, but the grad SLAM format teaches us how to reach people in a more accessible way. How to step back and see the general context in which we work and how to make our research attractive and explain it in an organized and concise way that will help people keep listening and most importantly will enable them to remember what we said. Thanks to the communication skills we developed through programs such as the grad SLAM, we can learn how to do our part in putting informative content out there that can help other construct their own informed opinion, start a conversation and make choices based on accurate information. To do this, our universities needs to continue investing in taking their graduate students out of their research bubbles. And I'm very grateful that the grad SLAM program is doing just that. To finish, I want to wish the best of luck to all the contestants. You should be already so proud of yourselves for being part of this event. Thank you. Thank you. In a minute, we will begin the presentations. First, a word about the order of speakers, rather than the usual last name alphabetical order, in an effort toward greater fairness, we have randomized the selection. For the grad SLAM competition, each contestant has three minutes to present their research. According to the guidelines used in all grad SLAM competitions, presentations are judged on intellectual significance, appropriateness, clarity, organization, engagement, delivery and visuals. Points are deducted for every three seconds that exceed three minutes. In preparation for today's event, we had each contestant pre-record their presentation. Please note that the contestants were prohibited from editing their videos after they were recorded. Our contestants are also here with us live in studio. So after each presentation, I'll be asking each student one or two informal questions so we can get to know them and hear about how they came to their research and their future plans. After all 10 presentations, you will be invited to select the People's Choice Award recipient through an online link that we will provide. At that time, we will also invite the audience to submit their own questions for our contestants. After our contestants have answered audience questions, I will then announce the winners. All 10 competitors today will receive at least $300 with the People's Choice, second and first place finalists receiving respectively $750, $1,000 and $3,000. Today's first place winner will go on to represent Berkeley at the UCY competition next month on May 6. Our first presentation is from Ryan Lendl-Cray, a fifth year PhD student in psychology. Let's play the video. Feeling that you're being watched? If so, don't worry, you're not alone. But did you know that even the mere belief that we're being watched can have important consequences for our behavior? One of the most interesting applications of this theory comes from the sporting world and is known as the Home Team Advantage. A variety of factors combine to create the Home Team Advantage, which describes the idea that across a wide variety of sports, home teams are winning games at significantly above chance levels. One such factor is the fact that the vast majority of the fans in that stadium are cheering for the Home Team. But the problem is that there has never really been a way to isolate or quantify the impact of the fans themselves on this Home Team Advantage. Until now, one of the downstream consequences of the COVID pandemic was that sporting leagues were shut down. Seemingly overnight, sports arenas went from the bustling complexes that you see on the left to the incredibly sad-looking places that you see on the right. Two leagues in particular, the NBA and the NHL, had their seasons interrupted by the pandemic. They then invited the team back into a playoff bubble in which the fans were not allowed in order to complete the season. So by comparing the games that were played prior to COVID with the games that were played in the bubble, we have the perfect opportunity with my research to study the impact of the fans themselves on this Home Team Advantage. So what impact did the fans actually have? Well, in the NBA, we see that teams are allowing significantly more points per game than the fans are absent, thereby performing worse on defense. Interestingly, in the NHL, we see the opposite pattern. Teams are allowing slightly more goals per game, but they're scoring significantly fewer goals, thereby performing worse on offense. These differences may not seem like much, but keep in mind that we're talking about professional athletes. These people train literally their entire lives in order to perform at an optimal level in front of a crowd. And so, we can ask ourselves whether these minor differences in performance translate into the most important metric of all. Namely, did the home teams actually win the game? And the answer is that resoundingly, they did not. In the NBA, we saw a decrease in home team win percentage from 65% when the fans were present, down to 52% when they weren't. This decrease was replicated in the NHL, where home teams won 55% of their games with fans, and only 40% of their games without them. These findings have important implications beyond the field of sports psychology. Teaching students that the presence of others could be beneficial, for example, might be an important first step in reducing test anxiety. And so, in summary, the next time you find yourself doing something that you think you're good at, do yourself a favor and make sure that someone is watching. Thank you so much, Ryan, and now I know there are going to be lots of sports fans feeling totally virtuous, paying for tickets to watch their favorite team. I very much appreciated your presentation. I'd like to start by just asking you what got you interested in this particular area of research. That's a great question, thanks. So think about it this way. If you've ever been somewhere with somebody that you know really, really well, like a best friend or a romantic partner or something like that. And then you get into this context, right? And they behave just completely differently from the way that you would expect them to behave. So much so that you're kind of staring at them and being like, who are you? Who is this? Why is this happening? And this has happened to me more times than I would like to admit for someone who specializes in personality psychology and is really, really interested in it. But that's what got me really interested in the power of the context itself. So I study all kinds of contexts and how they influence the personalities that we express. And I think that context has a much greater impact than a lot of people are giving it credit for. Thank you. And which aspect of your work do you find to be the most fun? Well, you know, the domain of subject and psychology is of course human beings. And so I really, really enjoy the fact that every time I go into conduct my research, I get to meet all these amazing people who have really unique and important life stories and learning all about them and how that's impacted their lives. And really interesting and meaningful ways is something that's super rewarding to me. And do you know what you plan to do when you finish your program? Yeah, I really love teaching. I love academia, basically everything about it. So that's the root that I'm headed. Great. Well, thank you so much, Ryan, for the research and for being with us here today. Thanks. Our next presentation is from Chandan Singh, a fifth and final year PhD student in electrical engineering and computer science. Let's play the video. In the last decade, something dramatic happened to machine learning. Class of complex models, known as neural networks, suddenly made enormous strides in analyze data, including computer vision, autonomous driving, and most interesting to me in science and medicine. These models work by processing huge amounts of data and searching for underlying patterns beneath them all. This is potentially a radically different way to do science. There's no hypothesis for testing. The network looks only in data and uses that to come up with a hypothesis. It's potentially enables discoveries far beyond what we could have just guessed on our own. But there's a problem. Even though network can precisely predict these patterns, you can't explain what it's doing. The patterns are stored somewhere in the millions of parameters of the network. But to answer our scientific questions, we need the ability to read these patterns. Importantly, we need a way to convert a monstrous neural network into a simple form which we as humans can understand. To make this concrete, here's an example I want to show you. In biology, we're interested in several different processes such as how do toxins enter a cell? Here we see a video of the cell being imaged under a microscope where each bright spot represents a potential entry point for a single toxin into the cell. A lot is unknown about this process, and we want to fundamentally understand it so we can design more effective drugs and better interpret disease symptoms. We can try to understand this process by watching cell videos, but it's really hard for humans to just look at and understand them. Again, if a model sees enough of these videos, you can statistically find which signals lead to successful toxin entry. In fact, we spent years building a neural network on this data, and found that we could predict toxin entry better than anyone could before. But we still couldn't explain what the model had learned. After years of thinking and working with scientists across different domains from cell biology to cosmology, we finally came up with the solution. We call it adaptive wavelength distillation. The idea is to take the complexity in millions of neural network parameters and distill them down to something simple. Crucially, rather than just looking at data, we used a hard-earned knowledge of decades of work by domain scientists to precisely specify which key information in the network you need to preserve. We then used this info to convert it into an understandable form known as a wavelength model. The wavelength model is extremely interpretable. This discreet red curve shows all the parameters of the model. Instead of just having a messy black box of network parameters, the wavelength model uses just this curve and makes predictions by scaling and shifting. Now that we have this model, a biologist can understand precisely what's going on. In our case, the rise and fall of this curve corresponds to a single event where toxin enters the cell. It tracks the rapid buildup and dissipation of proteins involved in the process of cell entry. This was awesome, but we wondered, can we trust it with something lost when we went from this behemoth to this little curve? To our big surprise, the answer was no. The smaller model actually improved predictive performance. This is basically because the neural network puts only at the data, but disregards years of scientific knowledge. By using that knowledge to specify the form of the wavelength model, we can get the best of all worlds, a highly predicted model that's also interpretable. Thank you so much, Chandan, for reminding us why simplicity is so important. I'd like to take a moment for us to learn a little bit more about you. I'll start by asking, how did you become interested in this particular area of research? Thanks, Lisa. Actually, I started being very interested in neuroscience where all the kind of mysteries of intelligence, memory, curiosity are stored in this ball of jello in our brain. Luckily, at the same time that I was starting my PhD, neural networks, these artificial models that are loosely connected to the brain were becoming extremely popular and becoming used in all kinds of different fields. It was the perfect time for someone with a computer science background like me to port over ideas that have been studied in neuroscience for a long time and try to understand instead of the brain, this mathematical model of the brain. And it turned out to be amazing because there's all these different things that neural networks can be used for. And if you were to study them the same way you'd study the brain, you can interpret them in ways that allow you to do things that immediately have nice impacts, when the models are biased in certain ways, improve them and try to generally make them better and more interpretable. And what kinds of things can we understand better by making the models better? Yeah, so really machine learning is like kind of nice in that anytime you have a lot of data collected in something in a scientific field, you can build a model of it that predicts well. And basically whatever your application is, like you can ask questions about the model in it. So I just talked about cell biology here, but this project was, you know, done with a ton of different collaborators. We have amazing people, not only in cell biology but in medicine or in cosmology and wherever you can fit a model, you can basically ask these like same kinds of questions. So hopefully, at least for me, the thing I'm most interested in is healthcare, and we can try to see, can we use these models to predict things well, things like diagnoses, but also interpret them, know that we're using them safely and know that, you know, we're not going to make false diagnoses with them. And so which part of this is the most fun for you, which part of the work? I think the most fun part for me has changed over time in the beginning was really like, can we find these like, you know, scientific discoveries, but as I'm getting older, and you know, going through the pandemic and things like that, I really like just, you know, seeing the impacts of your work and like and help people, for example, like right when the pandemic started our entire group, like, completely shifted what everyone was doing and like full time we were just trying to predict kind of like the progression of the pandemic and very directly working with nonprofits to try to prioritize where we should send PPE. I think things like this are like, very exciting for me now. I really want to see like, can we take the research we've done and apply it to real problems, especially in healthcare where it's, you know, there's people's lives at stake, and hopefully using these models we can make things better. Thank you so much, Candid, and congratulations on being on the tail end of your doctoral program. Thank you. And so we're going to go now to our third presentation, which is from Tanya Govachevich, a second year PhD student in Earth and planetary science. Let's play the video. As of today, we've mapped over 4000 exoplanets in the Milky Way. Many astrophysicists like me are driven by many open questions such as how the planets in our galaxy formed and evolved. Are the interiors of other exoplanets similar to the ones in our solar system? And what makes Earth so unique? We're tasked with determining the interior composition of exoplanets from several light years away. Scientists have come up with creative ways to theorize what these exoplanets are made of, and these methods allow us to categorize them based on their density. Studying the composition of exoplanets offers us insight into the formation and evolution of these planets and even to the planets within our own solar system. We're particularly interested in understanding water-rich exoplanets, or water worlds, because they're common in our galaxy and water is pretty important for life. My hope is to figure out what's inside these alien worlds and add a piece of information on the evolutionary pathways that took place in making the planets in our galaxy. Water worlds have been suggested to have upwards of 20% of their mass coming from water, being less than 0.1% here on Earth. This means that the surfaces of these planets are covered by a huge icy water layer. The ice layers are so deep that interesting chemistry begins to happen between the rock and ice, and they may behave in new ways that are alien to us. The interaction between that rocky material and water ice has only been theorized at these conditions, and I want to investigate the interactions at that rock-water boundary layer. Maybe they'll mix, and if they do, what are the pressure and temperatures needed to get them to mix? Until interstellar travel becomes a viable method of transportation so I can scoop up planet material, I've learned to use computers to determine chemical interactions of rock and water ice at these extreme pressures and temperatures. I'm able to model the chemistry that occurs at these extreme conditions at the atomic level, giving me insight into these planet interiors. Based on my calculations, I found that once rock melts, rock and water do mix. My results indicate that the interiors of our neighboring exoplanets are more mixed and a little more complicated than we think. There may be more oxygen and hydrogen hiding within the mantle of these worlds than we realize. As an astrophysicist armed with powerful computers, I want to be able to tell you the interior compositions of exoplanets, and the only way to do so is to begin researching the chemistry of materials and conditions that are pretty alien to us. Wouldn't it be cool to point up to a twinkling exoplanet in the night sky and know its interior from light years away, or that maybe it might be able to host life if it doesn't so already? A presentation that really seems to be the stuff of science fiction and shows you the great things happening on campus. So as with the other contestants, we're hoping to get to know you a little better. So the first question is, what made you interested in this particular area of research? Yeah, definitely. Thanks for the question, Lisa. So it was when I spent a summer doing research at MIT and undergrad, and we were encouraged to go to seminars given throughout the summer. And the third time I ended up in the planetary science department, I was like, okay, this is maybe what I'm interested in pursuing a graduate degree in. I'm a first generation college student, so outside of my undergrad field, I didn't really know what fields there were to pursue a graduate education in. And again, the third time I ended up at the same planetary science seminar, I just knew that, yeah, I'd pursue a PhD program in that. And then my curiosity, since I was a little, I've been super curious. I wanted to pursue a job that allowed me to be curious. Do you have any advice for other first generation scholars who might be thinking about going into doctoral work? Yes, ask questions. No question is stupid. You can, I have so many questions that I've asked people that like my other fellow first grad, first generation graduate student fellows have also asked and wondered and it just takes one person speaking up to help out a lot more people. And to really reach out and lean on the resources that are available for you and to find your kind of tribe. Thank you. And last question, which part of your research do you find the most fun? Kind of to echo Ryan's sentiment, the networking and hearing about all the people who do really, really cool research that I didn't even realize was out there. And sometimes like, you know, since we've been on zoom a lot just being in awe looking at all the people I get to talk to on a daily basis and kind of like, you know, seeing your favorite author in a video on a zoom screen with you is like, super cool. So yeah, the networking and learning the other research that people are doing. Thank you so much, Tanya. Thank you. The fourth presentation is from Sophia Stephens, who is a fifth year PhD student in chemistry. Let's play the video. If you've spent a summer night camping on the California coast and experienced the dense fog that morning friends, you know that summer is no time to leave your rain fly behind. But have you ever considered what makes your rain fly water repellent in the first place? The answer is that it contains PFAS, a class of water and oil repellent chemicals that have been used in products like rain gear, nonstick cookware and stain repellent materials. When PFAS is added to these products during manufacturing, some ends up in waste streams, which contribute to global water contamination. In fact, PFAS is considered ubiquitous and has been detected in remote regions like the North Pole, in addition to nearly all drinking water supplies. One exposure to PFAS through drinking water has been linked to cancer, endocrine disorders and birth defects. Researchers are trying to figure out the best ways to remove PFAS from water, but PFAS is a challenging contaminant to treat because it likes to float at the surface of water. In salt water, more PFAS accumulates at the surface, which affects how PFAS contamination spreads. For example, when PFAS is concentrated on the surface of water, it's more likely to stick to soil as contaminated water moves through the environment. In my research, I'm investigating how salts increase the amount of PFAS at the water surface in order to better manage PFAS contamination in saline environments like the California coast. So let me show you how salts affect PFAS. PFAS are a type of surfactant, which means they have a water-loving head and a water-repelling tail. The reason PFAS float at the surface of water is to keep their tails dry. So let's zoom in on that surface. The heads of PFAS are negatively charged, so when too many molecules accumulate at the surface, the heads repel each other. But what if we added some positive charge? These salts, which are positively charged ions like sodium and magnesium, can reduce the repulsion between PFAS head groups, causing more to accumulate at the surface. This means that in salty environments, there's more PFAS concentrated at the water surface. Through my experiments, I've quantified how increasing water salinity increases the amount of PFAS at the surface. This is important because if your research are sampling saltwater and freshwater, you might detect less PFAS in the saltwater if you didn't know to look for the PFAS hiding out at the surface. Accurate detection helps us prioritize sites for treatment to avoid further human exposure. So on your next camping trip, I'm not suggesting you did your rain fly, but remember it's PFAS keeping you dry. Thank you so much, Sophia. I will never look at my ring gear the same way again after this. So just again, help us get to know you a little better. Could you let us know how it is that you became interested in this particular topic? Yeah, so when I started research in PFAS, I didn't even actually know that these compounds were surfactants. I started observing super interesting behavior in solution experiments in terms of how I was detecting PFAS compounds. And when I started digging into that and learning about surfactant chemistry and all the solution properties like salinity and pH and temperature that can affect surfactant chemistry, I got really invested in understanding how environmental conditions could impact the behavior of these contaminants in the environment. Thank you. And what is it about your work that you find to be the most fun? Well, I like to say part of the reason I like chemistry is because I also love cooking and I really love the process of working at the bench and measuring things, making solutions, observing things, and just the hands-on component of the research is super fun and engaging for me on a daily basis. So just out of curiosity, what does a doctorate in chemistry help you with in the kitchen? Well, it definitely helps me if I'm trialing baking recipes and there are a lot of surfactants that are actually in food products. So that's kind of a fun application of understanding surfactant chemistry. Do you want to tell us which ones? Egg yolks are a great example of something containing a lot of surfactant. Things like whipped cream and any like easy cheese, any kind of things like that, a lot of those contain surfactants. Not all of which are toxic, I will tell you. And do they, out of curiosity, do they behave the same in foods? Well, they'll behave differently depending on the other components of the food or the food product. But the main thing that surfactants do in food is they help emulsify things. So they help kind of add like a creamy or foamy texture to certain food products. Thank you. Thank you so much for helping us see our everyday things a little bit differently. Thank you, Lisa. Our next presentation, the fifth contestant is from Kimberly Burke, who is a fifth year doctoral student in sociology. Let's play the video. There's a desperate need to address police violence in this country. But scholars are divided on whether the problems arise at the macro level of laws and policies or the micro level of individual officers. My research bridges that gap by asking instead what is the relationship between macro level standards and patterns of behavior across individual officers. The laws and policies that regulate police use of force tend to be vague, often reiterating the constitutional standard that police force should be objectively reasonable. I participated in an apartment's use of force training to determine how they translate objectively reasonable force into practice. I found that this department trains its officers to regard any non-compliance, including passive forms like failing to identify oneself, as a red flag of potential threat, and to neutralize that threat using low levels of force, like grabbing and handcuffing, to maximize the officer's control of the non-compliant person. I call this strategy forceful de-escalation. The department frames forceful de-escalation as a positive strategy because cops are less likely to shoot or be shot by someone in restraints. In this way, forceful de-escalation aims to reduce everyone's risk of being killed in police encounters. However, I argue that forceful de-escalation normalizes aggressive police responses to passive resistance. Moreover, it points to a dangerous belief that individuals should completely abandon their constitutional freedoms and police interactions, regardless of whether or not they are obeying the law. To examine how this department level standard impacts individual officers, I conducted interviews. Existing research suggests there are two types of police officers, warriors and guardians. Warriors are defined as having a militarized war-like approach to policing, and guardians are defined as protectors, more oriented toward communication. So, I hypothesized that warriors would be more likely to endorse forceful de-escalation. However, I was surprised to find that warriors and guardians alike supported the use of low levels of force in response to passive resistance. They also considered any defiance of police directives to be a criminal offense, regardless of the legality of those directives. And they consistently categorized their use of low-level support as de-escalation tactics, rather than reporting them as use-of-force incidents. Most significantly, I mapped each of these justifications onto existing case law, effectively bridging macro-level consequences back to macro-level structures. In conclusion, my research strongly indicates there exists a mutually reinforcing relationship between officers' justifications of violence and the laws and policies that transform them into reasonable safety measures. Presentation. We'd also like to get to know you a little bit. So, if you could start by telling me what brought you into this research. Yes. So, I've broadly been interested in issues of social justice since I was a young person. I think at the age of 13 or 14, I joined my neighborhood coalition as we advocated to remove the charcoal factory and paper mill that were located in my neighborhood, just a block from my house, two blocks from my middle school. And through that activism, I learned about environmental racism. I learned how to connect the personal to the political. That drove me to get a women's studies degree as an undergraduate. And I was fortunate enough to build a career combining my passion for social science and social change at an organization called the Center for Policing Equity that focused on racial disparities and gender inequities and policing outcomes. So, while I was there from about 2013 until I started at Berkeley, I was coordinating projects, serving officers, leading focus groups between officers and community members, tackling these really tough issues around racism and our criminal justice system. So, pursuing my own research at a doctoral level is a natural next step for me. Thank you. As someone who studies political engagement and communities of color, I know not all 13 year olds are getting involved in community activism. So, I'm curious what led you there or what brought you to start on this path? I have to attribute that to my mom. Honestly, I wish she could chime in. I think she's watching. I want to say she had us picketing when I was still in diapers. So, she has always been driven to volunteer work to champion for the needs of marginalized groups. And so it's kind of been in my blood. Yeah. And I know I've talked to everyone else about fun, but your topic is really profound and potentially I think a difficult, a difficult one to address. I guess I'm wondering if there's joy to be found in the work that you do. Joy. You know, that's not a feeling I've had much of the last few years. So, thank you for asking that. I think that I am moving towards joy once again, and creating spaces on campus for myself that really talk about fostering love amidst tragedy and building support and connections amidst crises. And in that I'm able to refocus and have a renewed sense of energy and commitment to this work, which is very difficult. So yeah, I think in connection with the other people who are advocating for change, that's where I find my joy. Thank you so much Kimberly for this important work and I hope you continue to move toward finding a sustainable practice for addressing really difficult issues in society during your career. Thank you so much. Our sixth presentation is from Coleman Thompson, who is a first year master's student in civil and environmental engineering. Let's play the video. It's estimated that by the year 2025, 1.8 billion people will be living in areas of the world affected by absolute water scarcity. One possible solution to this crisis is desalination, turning seawater into drinkable freshwater. My capstone team is using genetic algorithms to design solar powered desalination plants. A desalination plant consists of two components, an energy component and a desalination component. The energy component uses solar panels to draw electrical and thermal energy from the sun, storing it in batteries and large tanks of water. The desalination component consists of a reverse osmosis and forward osmosis system. My particular area of research was in the thermal modeling of the system, tracking the heat flow and temperature everywhere in the plant every hour for an entire year's worth of weather data. We need to do this because we need to know whether or not our plant design is viable. Do we have the right number of solar panels, the right tank mass, and so on and so forth. Each time step, the first thing we do is calculate how much the temperature of the thermal storage tank rises because of the heat pulled in from the sun. The next thing to do is calculate the amount of electricity we need to raise the temperature of the water up to 85 degrees Celsius, which is where we need it for the forward osmosis system to work. The last step is to calculate the properties of the heat exchanger. The heat exchanger uses many panels of metal to transfer energy from the thermal storage system into the forward osmosis system. We need to know the area of those panels so that we can inform our clients on what size and type of heat exchanger to purchase. The thermal modeling component of the system is only one cog in a broader machine. For us, the mean of the model is the genetic algorithm, which is what we use to actually optimize and find ideal plant designs for our clients. The genetic algorithm begins with a randomly generated sample of plant configurations, random numbers of solar panels, osmosis membranes, tank masses, and so on and so forth. We then evaluate this sample for two things, their viability and their cost. We take the cheapest viable plant configurations and use their parameters to randomly generate the next generation of plant designs. Then we repeat this process over and over many, many, many generations until finally we reach an ideal design that produces a certain amount of fresh water at an ideal cost. We can then hand this design off to our clients, who potentially will save a lot of money with our tool. As well as potentially produce lots of drinkable fresh water for millions of people for years to come. Thank you so much Coleman. I know we can all appreciate the importance of drinkable water and thank you for your great presentation. So again, we're going to take a little bit of time to get to know you and want to know first what brought you to this area of research. Yeah, so it really wasn't one individual thing it was more like a confluence of different things. This particular project represents an intersection of a lot of my different interests. So I'm interested in the environment and environmental policy and how we as a society can actually solve climate change like actually how to do that. I'm also was a physics major in undergrad. So the thermal component of this project was really interesting to me. And then I'm also really interested in machine learning and simulation using code. So it was kind of those three elements all together that really drew me to this capstone project. And which part of it did you find the most fun? I think the part that I really enjoyed the most is the problem solving aspect. It's just really satisfying to me when I'm working on this code and sort of slaving away for hours trying to figure out. Hey, you know, is this going to work and trying to solve all the little issues that come up? You know, why are our numbers not quite right? Why isn't the aren't the heat exchanger properties quite correct and gradually building and building and building until we finally get to a point where we have a solution that makes sense and a feeling there is just so satisfying. Yes, those feelings are rare but really important in research to keep folks going. Definitely. Do you have a sense of what you're planning on doing once you finish your program? Yeah, so I'm actually going to go to work with the company that we're working with right now, Optomy. So I'll probably be continuing to work on similar projects. I may continue to work on this project depending on how close to finalizing it we can get by the end of this capstone. But that's sort of a general idea of my future trajectory and after that I have really no idea. So you mean you're going to be working on continuing this work with the company that makes this product? Yes, and probably other different kinds of projects too. In the same, in the same field. Yeah, in the same general area. Well, thank you so much for the work and for making sure that we all have clean water to drink and for the great presentation. So our seventh presentation is from Justin Lee, who is a third year PhD student in metabolic biology and nutritional sciences and toxicology. Let's play the video. Let's address the elephant in the room. It's 2022 and SARS-CoV-2 and the COVID-19 pandemic is still persisting and here's what it looks like. New variants and breakthrough cases threaten to overwhelm our hospitals and especially for underserved communities with limited healthcare resources. COVID-19 continues to be a public health nightmare. Now, vaccines have played a huge role in preventing severe COVID-19 illnesses, but we still have limited universal therapeutic options. So we start to ask, what's the next breakthrough to end our ongoing arms race with the virus? My research tackled this question by first taking the existing DNA and RNA technology that was even used to develop the vaccine and then modifying it to specifically target the virus. My products are known as locked nucleic acid, antisense, oligonucleotides, or just ASOS for short. Now the name might be a mouthful, but the idea behind them is simple. Here's how it works. SARS-CoV-2, the RNA virus shown above in red, enters our bodies and has a specific structure that allows it to hijack our body's machinery to become its personal copy or to produce tons of virus copies for more infection and spread. The ASO that I developed shown below in blue can be inhaled through the nasal passages, which delivers the treatment directly to our airways and lungs. Once there, the ASO binds to the viral RNA and neutralizes it by disrupting the critical component that the virus relies on to make more copies. To show this treatment design works, I took COVID-19 effective mice and treated them with my inhaled ASO and saw that our treated mice had almost undetectable amounts of virus by PCR tests compared to untreated mice. My findings prove that my ASO effectively jams a key part of the viral copy machine and stops the virus from making more copies of itself. What's exciting about this development is that because my ASO targets such an important structural element that the whole SARS-CoV-2 family relies on to function, my one ASO is adaptive and effective against all virus variants, including the evasive Delta and Omicron variants. In fact, we can use the growing knowledge about the coronavirus to modify the ASO to target other important parts of the virus to allow us to have a quick and strong response to ongoing and future outbreaks. On top of it all, the ASO is stable enough to be transported without heavy refrigeration, so it can be widely distributed to high-risk communities with less developed healthcare infrastructures. So now imagine a mild case of COVID-19, but my inhaled ASO would help reduce symptoms and speed up recovery, which would be a game changer to ease the caseloads off hospitals everywhere. So let's just take a deep breath in together because I firmly believe my inhaled ASO is the complementary tool to combat this pandemic. Thank you so much, Justin, and I'm sure I speak for our panelists and members of the audience and that we really, really hope you're right. And that it was exactly the way that you're saying it does, but we'd also like to get to know you a little bit. So could you tell us a little bit about what brought you to this research? Yeah, thanks for the question, Lisa. I think like many people across all academic and scientific fields, our lives and works are really upheaved by the changes brought about by this COVID-19 pandemic. And of course we recognize that COVID-19 has brought about many dire situations and statistics in the world, but really the best we can do is really take this as a learning experience as a lot of people have. And so I came to my graduate career with the intention of studying metabolic biology, but upon the arrival of COVID-19 and like with many other professionals, I was able to take my knowledge and field expertise and apply it to understanding the complications of the virus and really become this kind of self-converted Corona virologist. But even just beyond scientists, this pandemic has taken policymakers, healthcare workers, economists, students, frontline workers and taken all their respective expertise and kind of applied it to really fight and understand this pandemic as a real big societal effort and really turned everyone in all different professions and walks of life into Corona virologists and their own respective rights. Thank you so much, Justin. I think you speak to what I think many of us think maybe the positive about the pandemic, right? The degree to which people have come together in community in so many different ways. I appreciate this is really important work, but which aspects of it are the most fun for you? Yeah, I think there's kind of two parts to it. So first, I think working at with this pandemic, it's really working at the front edge of science and learning as we go and really a kind of attribute that to kind of the idea of this thrill of the chase where, you know, as a little bit of an adrenaline you really don't know where you're going and you don't know where this road is going to lead. You wake up every morning and there's breakthroughs happening left, right up and down, just pulling almost like rabbits out of hats like magicians and so that's what makes it really exciting every day. So beyond just this intellectual excitement of these works and studies, a more of a technical kind of fun piece of this is that because we're working so closely with the virus, as you can imagine, we have to take numerous precautions to protect our health and so when we go to the virus, we have to wear these almost like astronaut like suits with the big helmet our little air purifier packs and we're all wrapped in. And so it kind of reminds you which is real funny because once upon a time I actually wanted to be an astronomer go to space being astronaut. So in a way I got to fulfill that dream in a very different way I didn't expect it to work this way but this way I'm keeping my feet on the ground and fulfilling it that way. And you get the outfit right at least. No we do not get we do not get the outfit just in case there's anything that accidentally got on us we keep that and keep that in very safe containment areas. Thank you. We get fun pictures. Yeah, there you go. Thank you so much Dustin for your presentation and for a very important work. Thank you. Our eighth presentation is from Sarah Harris, who is an eight year PhD student in German. Let's play the video. In the freezing winter of 2019 Chicago woman Candice Payne anonymously rented hotel rooms for over 100 people experiencing homelessness. Her life saving actions made news worldwide but in German her act of kindness was reported as though a man had done it. This wasn't an intentional misrepresentation, simply the result of the generic masculine, a linguistic practice in which masculine words like brother, father or he refer to all people, especially when gender is unknown or irrelevant. We have it in English too, like a researcher should present his work. In German, the generic masculine is found in nouns which come in feminine and masculine pairs. English has a few of these like actress and actor, but most of its words for people are gender neutral. Researcher doesn't mark gender for example, it's the his that makes it masculine. But in German the majority of nouns for people are gender specific. Though feminine words only are used for women, masculine words can be either gender neutral or specifically referred to men. The likelihood of picking one interpretation over the other can be put on a scale like this. What we're taught is this black pointer on the left, that the generic masculine makes us think of anyone, regardless of gender. But studies in English show that it's much more likely to make people think of men. In short, the generic masculine isn't generic at all. This has significant social, economic and even legal consequences for women and non-binary individuals. It influences what jobs people apply for or think they'd be good at and they can make media less accurate. Though these social effects are well researched, the linguistic causes behind them are less clear. For example, do the properties of a language affect how its speaker interprets the generic masculine? Specifically, if we know that English is here on our scale, where is German? Are its speakers more or less likely to interpret the generic masculine as male? My research answers this question by examining how the two languages talk about gender and applying existing knowledge to those differences. For example, remember all those gender-specific nouns German has? Studies show that children learning a language with this property learn to categorize people by gender at a younger age. And that makes sense because these children are thinking about gender much more when they speak, compared to English-speaking children. My research finds that the more a language makes its speakers think about gender, the more likely they are to read the generic masculine as referring to men. Compared to English, German has more properties that make people think of gender, meaning that a German speaker is more likely to interpret the generic masculine as male. This is significant not only for German, but for any language with properties that make people think of gender. The generic masculine influence is how we see our world, but we understand it not only when we look at the social effects, but also their linguistic causes. And my research helps us do just that. Thank you so much, Sarah. As a Spanish speaker, now I'm really worried that I'm probably on the upper end of that scale. And you did a beautiful job of laying out the significance of these linguistic practices. But I'm curious for you personally what it is that brought you to this topic. Yes, thank you. You know, when we teach German, and I've been teaching German for a while, one of the first things we teach students is how to say that they're a student. We want them to talk about themselves. But that choice implies gender. Are you, you know, our students saying it's been student, it's been student in my male student or my female student. And obviously that's not working for everyone. We have non-binary students. We have, we have so many situations in which the gender that we're teaching students to speak doesn't actually represent the lives of their living. And then we look at the generic masculine where we say, hey, this is just a shortcut. Don't worry about it. But studies show these really serious effects. I was interested in that and I wanted to see how can we bring these two languages together? And that's what my research is trying to do is look at this one thing in these two languages and go, what can we see when we put our minds to it? And I think I speak for many people and that I think folks totally appreciate what you're saying, but don't really know how to fix it. So I'm curious if you have any ideas about, how do we manage, right? How do we create space for that non-binary or non-gendered space? The first thing to do is realize how often gender is being used in our lives and our languages in the first place, even in English. We have a lot of opportunities now to use they and that's been really great. One of the things is noticing what does inclusive language mean? 50 years ago it meant saying he and she. Nowadays we know we could be more neutral and that would actually include a lot more people. So look to what language, what words you use a lot in your language? Look for other opportunities to say person instead of, I'm happy to see the men and women in front of me. Can you say person? Can you say welcome everyone? That's what I do to my students. I try to keep gender out of it unless I know the person's gender and we're referring to it. Thank you. Can I get a note from you from my editor for the journal when I used they and they tell me that that's not correctly correct, but I'm kidding. Thank you for saying that because I think there's a choice all of us have to have, which is how important is it to be correct if we're hurting people? What is more important to you? And I'd rather be wrong and a kind person. I totally agree. And the last thing I want to know though, is there any aspect of this work that is really fun for you or which are the parts? Sing language change. Like I said, 50 years ago we had completely different solutions to this problem that we still have. For me, what I do is I try to keep on top of what language people are using and why they're using it. So for right now, that means I'm watching Queer Eye Germany. On Netflix, which has been great out of the fab five, two of the participants are non-binary. They use different types of pronouns and that has been excellent for representation and it's fun homework. I'm always in favor of fun homework. Thank you so much for your work. And also happy birthday. So we have to acknowledge that and just appreciate you and pushing us to think about things that we kind of take for granted in how we speak. So thank you so much. Thank you. And our, I'm sorry, a lot ninth presentation, I'm sorry, is from Jamie Simon, who is a third year PhD student in physics. Let's play the video. Humans have dreamed of making machines that think as we do. And with modern machine learning, we're finally getting close. Rain inspired algorithms called neural networks can now automatically do things like transcribe your voice, translate your text, caption your photos and more. Neural networks are everywhere in tech these days and you use one whenever you talk to Siri or navigate with Google Maps. However, there's an embarrassing elephant in the room becoming more important all the time. Nobody actually understands why these amazing algorithms work so well. To understand the problem, let's contrast the work of a machine learning engineer with the work of a civil engineer who wants to design the bridge to bear the most weight. They might start by generating several candidate designs. Using Newtonian physics and material science, they can then predict the performance of each candidate, make modifications if needed and then choose the best one to go and actually build. There are scientific principles for bridge design and it would be crazy to just build them all and see which one worked the best. Now suppose our machine learning engineer wants to design a neural network that can distinguish pictures of zeros and ones with as high accuracy as possible. There are many, many types of network that might do this task, but there are no scientific principles that would allow one to actually predict their performance or choose the best one. These algorithms are black boxes and an engineer often has no choice but to just try them all. This is really slow and expensive and when people do come up with clever new algorithms. Most of the time they're just guessing. Compared to bridge building machine learning is pretty much alchemy. My dream is to transform it into a science. In a step towards that goal, I derived a new set of simple physics like equations that predicts the performance of certain types of neural network. My equations not only predict performance but shed light on why it was good or bad and how it might be improved. The black box of neural networks is illuminated and we can choose which network type would be best without having to run a single one. Here's an illustration of my results on the real task of distinguishing zeros and ones. Now neural networks learn from examples so I've plotted a network's error against the number of example images against to learn from. As the dots show and as you'd expect the more examples to lower the error. The blue curve shows the predictions of our theory and the match is almost perfect. Though we're proud of the success we still have a long way to go. Our theory only applies to simple network types so we're working hard to extend into those used in practice. But still, it's hard not to feel like we're uncovering something fundamental like the laws of physics for neural networks and this makes the work quite exciting. Humanity has already worked wonders with neural networks. So just imagine what we'll be able to do once we finally understand them. Thank you so much Jamie I'm going to quote you of machine learning being alchemy. Very interesting presentation and interesting work. How did you become interested in this particular area of research. Yeah, so thanks so much Lisa. The most compelling thing to me is just the fact that we have these amazing machines these computer algorithms that can do things like, like learn to understand your voice or recognize your face. And they're transforming society every day, and we don't understand how they work. Like to me this is such a compelling fascinating mystery that I can't not think about it. I encourage everyone to, you know, next time you use Siri or talk to Alexa, like, take a moment and think about how how fundamentally mysterious it is even to the people who made it. I think this isn't widely appreciated but I'm like should be regarded as like big scientific mystery. And so it's a combination for me. Well, let me hear with a combination of a fundamental interest in discovering what's going on. And also the understanding that, like, good solutions will really affect the future of technology in the next few decades. Thank you. And which aspects of the work you talked about how important it is, which aspects of the most fun for you or do you enjoy the most. Yeah, so there's a really wonderful type of moment that occurs in technical scientific fields every so often, rarely but excitingly where you've done all your theory and made your predictions and think you've thought everything through but then you run the experiment and you see if you were right. You see if you succeeded and your predictions work out and I've had a few of these moments in the last few years and those are always the most exciting when I, you know, generating plots like we're on my third slide where theory matches experiment and you've succeeded. So getting it right. Not so much the getting it wrong part. Yeah, I mean, you have to get it wrong 10 times before you get it right but like, you know, Yeah, I do think it's important to tell people about that part though because we tend to only present the getting it right. Yeah, but how much we have to get it wrong before we can get it right. I think it's important. Yeah, yeah, sure. I've had lots of like terrible theories before arriving at ones that worked so the road to success is paved with failure. Another very good quote. And with that, thank you will end by thanking you Jamie for your great work and we will turn to our 10th and definitely not last but not least and final presentation is from Tina the Josh, a fifth year PhD student in comparative biochemistry. Let's play the video. What does this picture remind you of our good old days when we were meeting and creating each other with big smiles, even traveling and enjoying ourselves on this planet. Unfortunately, SARS Covid virus attacked us. And now it has been more than two years that this uninvited guest has put a mask on our beautiful smiles. Even though vaccines are available, the infection is still spreading. Do you know why? Because SARS Covid virus is very, very clever. It is pulling the human body by changing its appearance in the form of five different variants. This leads to a decrease in the effectiveness of vaccines over the time. Here's where my research comes in. I am using a unique approach to kill this virus by attacking its crucial protein, the pain like proteins, also known as PL Pro. PL Pro is present in all the Covid variants and it acts like a heart of the virus. Without PL Pro, virus cannot exist. So my goal is to find a way to stop the heartbeat of the virus for the treatment of Covid-19. To do so, I prepared PL Pro in my lab and developed a tool to detect those compounds which could kill its activity. I screened a chemical library which consisted of 115,000 molecules against it. The most exciting part of the story is one compound was able to inhibit SARS Covid viral replication in the cells. Importantly, unlike many others, this unique compound has low toxicity. Now, my next step is to inject this compound into mice and develop it into an oral drug for humans. I am proud to say my discovery of a unique anti-viral compound which inhibits PL Pro and SARS Covid viral replication will surely help us to win the war against SARS Covid virus. Thank you. Thank you so much, Tina. And while, of course, I think it's obvious to all of us why this is important to be researching, I'm curious if you would be willing to tell us a little bit about the path that led you to this particular program and this particular project. Thank you, Lisa, for your question. So my father is a patient of retinitis pigmentosa, which is a genetic disease and it doesn't have any treatment. I realize that there are so many diseases which doesn't have treatment and science needs so many people to make it advance, right? And we have seen that Covid-19 really needs people to make it work for the treatment. So the unmet need of Covid-19 brought me in this project and I'm very happy to contribute towards the human welfare through this Covid-19. Thank you. Thank you. And thank you for raising the issue of rare diseases, those diseases where no one is doing research because they're relatively uncommon. Do you think that's something that you want to move toward as you move on in your career beyond this project? I'm happy to contribute in any kind of disease in which there is requirement, irrespective if it is rare or very common. I understand I have a very personal emotional touch with rare, but I really understand that everybody is in the need so we can help anytime to anybody through the science. Thank you. And last question for you is just which aspects of the work do you enjoy the most? I enjoy the most is analyzing the data, you know, when you are done with your protocols, you are done with your experiments. Now you have the data and you can make conclusions out of it. And then you can tell to your professor, oh, we got this data and now this is time to move forward, you know, with new experiments. So that's the most exciting part I feel like for me at least. Thank you. And it sounds like you echo some of the other contestants in that joy of discovery, right? There's nothing like it when you just come up with an answer you can move forward on. So thank you so much for your research and for your great presentation. And that was the last of our presentations. I hope you can all appreciate why I think it is such a privilege to get to work with Berkeley graduate students every day. They are all amazing. And we're so appreciative of all the folks who submitted presentations and to all of our finalists today. And now members of the audience it is your turn. It is time for you to vote. And while you are voting, you will also have the opportunity to submit questions that I will ask of our participants. Please refer to the URL on the slide to cast your vote for our People's Choice Award and to submit a question for one of our contestants. If you are viewing this through our website, you can find these links above in the live stream. We'll have a 10 minute break to allow you to vote and write in your questions. So let's say a 12 minute break just to make it easy. So we hope you will join us back here promptly at 425, and please be sure to go to the URL, submit your vote. Remember the People's Choice Award winner receives props and a bigger cash prize which for our graduate students is really important to give them how expensive it is to live in the Bay Area. And we want to know what you thought and who you thought did the best job in terms of their presentations. Thank you. Welcome back everyone. And thank you so much for voting and for submitting your questions. I'll now ask the panelists, your questions to get their answers and our first question is for Ryan will be going in the order of the presentations just so that our contestants are clear when their turn is. So Ryan, can we explain why traveling teams and in jail only to an empty stadium now have a 60% of winning. I think 60% chance of winning I think. Do you predict any additional factor is involved. Yeah, that's a really interesting question. Thanks. The one thing that complicates the home team advantage quite a bit is that there's a large degree of factors that contribute to it. It's not just the fans and the fact that we were able to remove the fans. Well, technically we didn't the pandemic did right but the fact that that was able to remove the fans completely is sort of the best naturalistic manipulation that we can have. There's also all of these other factors right like the, the bubble for example the playoff bubble was played all in one stadium. Right so that also removes the idea that, you know, maybe the home teams are three way teams are more familiar with the stadium. Right, but in terms of the NHL being so much lower than the NBA which I think maybe I'm not understanding the question, but that's that's sort of the way I'm interpreting it, I think it has something to do with sort of the period in the game in which the fans are able to get involved. So when you think about the NBA right and you think about when are the fans most vocal. Well it's when the visiting team has a free throw, right and that is a very clear opportunity for the fans to just be as loud as possible and you know, distract as much as possible. And when the home team has a free throw they're just completely dead silent. And then the NHL there's not really a set point at which, you know, other than face offs and media timeouts where the game really stops. And so the fans are kind of a more continuous presence as opposed to oh I have to be loud really right now. I rarely have one specific factor and I apologize it's not a great answer to the question. But that's sort of where I'm thinking in terms of the differences and I'm trying to harsh that hash out that argument a little bit better actually as we read up this paper. But that's that's where my line of thought is on that one. Thank you Ryan for helping us think about the context and the rhythm of the game right and how it is related to your research. Our next question is for Chandan. Can this method be used to simplify other networks, or is it application specific. Yeah, so this method is specific to neural networks so not like you know all generic networks that come up in different areas but neural networks themselves are quite generic it's basically any place where you have the ability to collect kind of training data so you have inputs and outputs to a system. So I talked about cell biology with our collaborator go cool but we also have other places where we applied the exact same algorithm basically some cosmology you can kind of collect you know inputs and outputs from simulations and medicine you can collect inputs and outputs from patient data. And it's pretty generic to all of these things. That's where it's saying like this specific type of model doesn't work all the time and no model really does. We're building in a lot of this domain knowledge that works in contexts where waylets are really useful for different things. But I think this general route of trying to take you know really large models and summarize them into small ones is really promising, especially over the last decade where things you know, I've gotten better predicting this rise in complexity has been you know really expensive and makes it really hard to tell when somebody's going to fail. So if you're in some high stakes environment whether it be policy or medicine, I think this kind of thing will become more and more popular, where we try to replace this big model with this small model, and we can go back to humans and say, you know, can we actually use this in the real world when we're going to make decisions for real people. Thank you, Shannon. Very helpful. Our next witness for Tanya. Do you use fluid mechanics and boundary layer mechanics to understand the water ice and rock ice interactions. Good question. So this is just probing the atomistic interactions. So at the fluid, if I was to apply this to fluid dynamics I'd have to perform quite a bit more complicated calculations to determine transport properties but it is something that I look forward to doing in my future research because that actually lends itself to helping us understand the interiors of planets and how, you know, this mixing effects maybe the mass and radius and what we're looking at when we look out into the vast, you know, sky of exoplanets now that we have. Okay, now I have to ask you what your dog's name is. Rocco, when I heard from some family members, he made a guest appearance barking in the background. I was showing his support. It's what they did very enthusiastically. Yes. Thank you, Tanya. And this speaks to not only the brilliance of our students for the brilliance of our audience. So please keep submitting your questions for our panelists. We're going to move on to Sophia. I have heard that PFAs is are found in the oceans all the way up into the North Pole. Is that true? Unfortunately, that is true. PFAs have been used in so many consumer products and so many applications that they have found their way into just about all the water out there. And that's one of the major challenges in treating them. And of course, depending on, I mean, part of the thing that my research contributes to is trying to understand how to treat them in different types of water, like ocean water versus fresh water. Yes, sadly, they're even in the North Pole. Thank you. And sorry, I cannot thank PFAs. Appreciate it. So our next question audience question is for Kimberly. Do you anticipate extending this project to collect data from more police departments? Yes. So this was almost a pilot project and I will be expanding it across multiple sites for my dissertation level. Yeah. But given the logics that are kind of inherent to policing, I suspect that we will be seeing some similar trends and justifications of force with some variations, of course, in places where they have, for instance, eliminated resisting arrest as a standalone charge. But yes, I will be expanding this research. Thank you. Our next question is for Coleman. Our audience member is wondering if there is any approximate timeframe when this amazing process can be brought to market. So I actually am not completely sure about that. That's probably would be a really great question for my CEO. But the main hurdle, I think, is really to me that, first of all, you have to find people to finance these plants. One of the main differences between our plants and plants of the status quo is that we will probably be needing a higher investment costs. Unlike contemporary plants, we need to produce our own energy, whereas contemporary plants usually just buy it from the grid, which generally means producing a lot of fossil fuels, which is one of the advantages of our approach. The downside of that is we need to build basically a giant solar farm in order to power a plant, which means a very large initial capital investment. And so then the other problem, obviously, is also that you need a lot of land. So again, that would probably be a really good question for my CEO. I can't give you like a direct timeframe in terms of years. If I had to like make a spitball guess, I would say maybe like 10 or 20 years, but I'm not completely sure. So yeah, but that's a really, really important question. And it's something that I think this in this project, we will eventually have to be able to address I think how long will this actually take to build, but that's really something on the way more technical and then we've been looking at right now. Which is just, can we actually even make this thing viable, which is the question that my research is immediately addressing. Thank you. Justin is next. Are probably more skeptical audience members are worried audience members are asking if there are any side effects in the mice. Yeah, that's a great question. So, as we've done with our studies now and we use the mouse model to kind of try to mimic some of the disease pathology we see in humans we have not seen any really adverse side effects with this treatment. And I think it's attributed to some of the modifications that we make on these asos so very similar to that of the vaccine we've kind of modified that to make sure we don't have any adverse body events. One of the biggest, I guess, worries that you would have with treatments like this of this nature are degrees of inflammation and we haven't seen any severe degrees of inflammation that would adversely affect any patients or in our mouse model. They're helped after before and after this treatment, but great question. That's very good to know. Thank you Justin. So Sarah, the question is, have you collaborated with other language linguists regarding gender and words and the generic masculine. If so, what are some examples and where are the similarities and or differences. Thank you. Yeah, great question. So the life of a humanities PhD is pretty lonely. I'm mostly I'm mostly working on things by myself but there are opportunities like the Berkeley Language Center to learn about I just went to a talk about, you know, how our language teachers, especially in high school and younger, working within working with our students about non gendered language and this talk had people from Russian and German and it was really interesting. But, you know, we look at something like, even if you, you know, watching content, though, you know that there's a to Bruno and our Tia PEPA, right. That information is there. And I think the more that people talk about it, not just not just linguists and scholars but everyone talk about, you know, what does it mean to have all this gender in our language to have to say aunt and uncle is there. I have a word that's a neutral term for that. The more that they that people talk about it and more that they share things that I've worked for them in their language. I think that's going to make us stronger. I need to I need to talk more with other people. I only know German, Dutch and English. So I don't know much Spanish or like Arabic these are languages that are also interesting for this so I'd love to know more and to talk to other people about it because they're going through the same things. Thank you and I have to give you extra brownie points for the income of reference. And a full that into grad slam Berkeley I appreciate that. Next question is for Jamie. Intuitively, how is it that physics equations that describe nature model human made neural networks. Yeah, so that's a great question. It's not the same equations. It's not like we've just taken equations that describe, you know, Newtonian mechanics, or you know fluids or something and then say I it's the same equations that work here we had to derive new ones. But what makes it physicsy is that if you look at an equation like f equals ma that's simple and a human can understand, you know what the parts are and get an intuitive sense of how the different inputs to the equation interact and how they you know lead to interesting behaviors. They're interpretable and that's what makes a good physics equation. So when I said the equations we've derived our physics like. That was what I meant like we have a equations that are new but simple enough that someone can see them for the first time and get intuition for what's actually going on. Thank you. Last but not least Tina, your question is, are there other essential cells in the body that contain PL pro that will be affected by appeal pro inhibitor. Nice question, but the answer is human body does not contain PL pro it's a viral protein which is specific to coronavirus. And if you are putting in drug against PL pro. It will be only killing yourself. It will be only killing viral infect virus virus, I would say, so that it won't be effect. It won't be killing human cells. Thank you. So, please, audience members take the opportunity to pick the brains of our brilliant students. It looks like we don't have so we could do another round of questions but we don't have questions for everyone. So I think we're going to put our contestants out of their misery and and move on to the awards stage. As soon as I hear the People's Choice Award winner, just have to take a moment in the chat. Maybe I'll ask does any contestant want to add anything or say anything about their work that they that they think people should know. Do you want to raise your hand they can spot the our wonderful AV team that has been helping us out today can spotlight you. There's no contestant saying that now we don't talk about Bruno is in their head. I appreciate that as a native of the fan. Any thoughts. Brilliant students. No. Okay, so we're just going to wait one minute. I wish we had some kind of a drum roll that we can give for the awards piece. Actually, we could also have our contestants ask questions one another. Yeah, sure. Sorry. I have a question for Coleman who said he was using genetic algorithms to which are a form of machine learning algorithm to design a desalination plants for use, you know, industrially. The algorithm starts by choosing a random plant, like a random plant design and then making it better iteratively. So like how bad are they at the start. They are pretty bad. We try to design them initially within like certain parameters. So like right now the form the genetic algorithm is actually slightly evolved since I recorded the talk. So now we are starting with a plot of land and trying to optimize the configuration within that plot. So the two things that we are really looking for with the initial population are, does it fit in the plot, and does it run out of power over the course of the year during the winter. And as long as it fits those two things then it's fine, initially. And of course we're optimizing from there. So initially the plants are quite a bit more expensive and inefficient. Usually by the end, we are producing about four times as much water at depending on the size, anywhere from half to a quarter of the cost and maybe even making a profit within 10 or so years. So yeah, they're pretty bad to answer your question, but they get a lot better. And so now, if we can have a, if everyone can imagine a drum roll in their head or still sing, we don't talk about Bruno. I'm very pleased to announce and well first I want to say you are all winners, you are all amazing, your research is amazing. And our but our people's choice winner is Tina Bajaj. Congratulations. Yay. Tina do you want to say a few words. Thanks to everyone who wanted for this. Thanks a lot. It's my pleasure to become one of the semifinalists as a UC Berkeley graduate student. Thank you. Thank you, thank you. So now I have our second place winner. I'm opening the envelope. I am pleased to say that the second place winner is Kimberly Burke from the sociology department for her project. Congratulations Kimberly would you like to say a few words. Thank you. Thank you to everyone on the grad slam team this has been an amazing learning experience and it's the best thing as a researcher to be able to share your work. Thank you to family and friends tune in. First time they've heard about some of this stuff so just thank you for the opportunity. Congratulations, we look forward to reading the final product. And now the moment everyone has been waiting for our first place winner. And our campus grad slam champion is Justin Lee from the program and metabolic biology for his project jamming the cards. Happy machine. Congratulations. Thank you. Thank you to my family and friends and I of course I would be remissed if I didn't acknowledge the contributions of a senior postdoc in our lab she zoo for this undertaking my PI under snar as well as the rest of my lab members lay Melissa, and of course thank you to the grad pro and grad slam committee and judges and the backstage crew today for their support to put on this amazing event. And I'm just absolutely honored to share this award among such a diverse and accomplished field of graduate students here today so thank you all I'm truly honored and I guess go bears. Absolutely. Justin will represent Berkeley at the UC wide grad slam on May 6, and will compete against the finalist from the other nine UC campuses. As we wrap things up here I'd like for us to give a special round of applause to all the student contests contestants sorry who are all amazing and did such a great job today, sharing their research with us. I'd like to second Justin's point of thinking the AV staff that's done such a great job. I'd also like to thank our judges, our keynote speaker, the advisory committee for graduate student and postdoctoral scholar professional development, and the amazing staff members in the graduate division who worked so hard to make this event possible, especially Linda von Hanna and Ali Gleason. And lastly, I just want to thank all of you for being here for being part of this event with us. And for all you do to support graduate education here at Berkeley, please join me in giving everyone just a huge round of applause. Congratulations all.