 Our next speaker is here to share with us a message that is both incredibly timely and vitally important. IBM Research recently announced that it's releasing new AI and cloud-based resources to help research and medical communities better understand COVID-19 and accelerate discovery around a potential treatment. Though some are still in early research and testing stages, IBM is making them available now due to the urgency surrounding the coronavirus pandemic. For more on this story, joining us now is the co-director of Science for Social Good from IBM Research. Let's listen. Hello everyone and thank you for joining me today for this virtual session. I know that we imagine doing this differently being in San Francisco together in the same space, but we find ourselves in the midst of this unprecedented human catastrophe and hoping for solutions, looking for ways to end this pandemic. But in a way, this story, this talk is about big challenges and it's about how they sometimes present us with big opportunities to think differently and also opportunities to use data and artificial intelligence to do some real good around the world. And in a way, this story starts many years ago, seven years ago, I believe in 2013 when I was leading our data science team in IBM Research. And you may recall that it was the height of the data science craze. It was the hottest skill in the world. It commanded huge salaries. It was transforming the industries and the way we live in the world very much to what's happening with artificial intelligence today. Anyhow, it was all great, but my team and myself included, we kind of felt that each that need to kind of use your skill for some time that maybe goes beyond just industry problems, things that might mean something for all of us and for the entire world, things that not always get the best skills and the best resources available. And anyhow, as we were thinking about how we can get engaged and what are the kinds of things that we can work on, another epidemic happened. It was August 2014 when there was an Ebola emerged in West Africa. And we read about it, we learned about it and we thought, well, gee, this is really a great opportunity. This is where we can really put our skills to good use. So we reached out to our colleagues in academia, colleagues in other organizations and companies. And we started to organize things like hackathons and data dives and created mobile apps to collect data about hospitals and treatment centers and perhaps even hoping to do better disease forecasting. But the funny thing is that when I look back at that time and when I think about what we did, I realized that we actually didn't do much. We didn't do much with Ebola and we didn't really help in any way, shape or form. And there are two reasons for that. One is because quite frankly, the big challenges of this world, the big problems of this world, they're not going to be solved in hackathons and weekend events. Hackathons are great to schmooze and meet new people and pick up a new skill, but they're really not great for creating solution. Another thing that was painfully obvious to us is that we actually did not quite understand the problem. And if you think about us, the average scientists and researchers and engineers, in a way, we are very blessed. We don't know what it means to be poor and what it means to be hungry and what it means not to have access to clean water or medicine or to be discriminated against. And then it's very hard to find solutions for problems that you don't understand or you don't experience. So we thought about it and we decided to do things differently. We work in IBM Research. It has 12 labs all over the world with more than 3,000 scientists and engineers with fantastic skills. So we decided to tap into this talent and actually do good in a more formal way. So we created this program. It's called IBM Science for Social Good, where we partner with NGOs, with public sector agencies, with social enterprises, and we listen. We try to understand what is it they're working on? What are the kinds of challenges that they're facing? And we try to find the problems where actually technologies, new technologies could make a difference. And then we partner with them, we scope these projects, and we do it together. Interestingly enough, our very first project coincided with another epidemic, Zika, and we teamed up with the researchers from Cary Institute for Ecosystem Studies. They are disease ecologists, the people who hunt for diseases, try to understand diseases. And we teamed up with them because it was obvious that these global emerging epidemics are a completely new problem for this world. We're seeing these diseases that we've never seen before. So we wanted to understand this a little bit better. And one of the first things that we learned from our friends at Cary was that most of the modern diseases are of the animal origin. What it means is zoonotic, right? The virus resides within an animal. The animal doesn't get sick because they have natural immunity. But then another transmission mechanism happens like a mosquito or another animal and the virus gets transmitted to humans and the disease emerges. And what's really interesting with Zika is that even though we understand this mechanism of spread really, really well, we don't actually know which animals are reservoirs of the disease, the carriers of the virus. So in this project, we teamed up with Cary Institute to use machine learning and apply it on the data about the zoonotic diseases that are known and data on animals, the primates that we know and their characteristics like body length, the number of offspring and things like that. And to use machine learning to try to predict which animals around the world could be reservoirs of Zika virus. And we did that actually were quite successful and we identified a couple of new carriers. And what it does, it actually allows you to be able to then create these risk maps to be able to say, okay, these are the regions for high risk of disease. This is where we think the new epidemics might happen. And you might be even able to do targeted animal surveillance and control to prevent the disease from happening. So this was the first project, we've been doing this for five or six years now. And you can see here on the screen, it's our project map. It's really a set of wonderful stories. We did things like happening blind and illiterate people navigate the world around us, help use machine learning to inform policy decisions, help food pantries be more optimized with cognitive supply chains. And I think it showcases this world of possibilities. It showcases how inspiring these problems are and these difficult problems are. And they make us think big and they make us come up with completely new and never seen solutions. Let me give you a couple of examples. Opioid crisis. Opioid crisis is one of the biggest health challenges here in the United States. And the sad thing with opioid addiction, it happens in a very benign way, like you have a surgery or you have a pain incident, and you go and you get this bottle of pills. And sometimes you get it for three days, sometimes you get it for five or seven, it's not really quite informed. And what we tried to do in the Opioid project was to actually analyze very complex healthcare data claims and prescriptions using causal modeling machine learning, very fine causal modeling algorithms to kind of tease out these stories of addiction, how people get addicted and why and what kind of prescribing behaviors are more or less risky. Because if you can do that, then you can come up with informed healthcare prescription guidelines, you can inform the policy, you can even help identify which people are more or less vulnerable. Another example, a lot of drugs, generic drugs actually have a potential to be repurposed to treat cancer. So in a project with the NGO called Cures Within Reach, we partnered with them to create a new natural language processing models and helping models that could read scientific publications and actually identify candidates or identify generic drugs that have a potential to be repurposed for cancer. In another project, we teamed up with United Nations Global Pulse to study and analyze and model hate speech and understand how it influences or might influence events around the world. Recently, we worked with NGO called Citilink from Cincinnati. They are a provider of social services to underprivileged individuals. So what they do, they have like a very holistic approach of actions, such as financial counseling, healthcare counseling, education counseling. And we work with them on analyzing the data of their interventions, many years of data, again, using causal modeling to understand which intervention are more or less successful and which combinations of interventions actually lead to the best outcomes. So again, every project is a fascinating story. So if you look at this map, it's like so many wonderful stories that can be told with technology and with data and with AI. But I also want to be realistic here, because I don't want to sound as if we are solving these problems. They are big. They are going to take many years, sometimes even decade to solve. But we are doing these tiny little steps, providing little building blocks of bigger solutions, creating reusable assets that others can use or build upon to eventually move the needle. One interesting thing when I talk about this map and I show this work to other people, they ask me, is this a new form of philanthropy? I think about it and say, no, no, this is not about philanthropy. This really has to be in the very fabric of how we do work and how we run businesses. And here is why, because not only did we do good, but it's also about the fact that this world is faced with some really, really difficult problems. And when we think about them, we need to think outside the box. And when we do that, we actually end up being better scientists, better researchers, better engineers, and we eventually help our companies make better products. And in that process, we end up also being better human beings. So that's quite of a return on investment, as they would say. Another important thing is that when we do that, we actually end up coming up with potentially even entirely new application area. So let me show you another example. Here, on this slide is an example of what we call generative modeling, right? Artificial intelligence can create things, it can create images, it can create music, sometimes it can create fake news. And that's all very kind of impressive, right? But the question is, why would you want to do that? Would it be possible to create things that are perhaps a little bit more useful? So we thought about it. And we said, okay, how about we try to create generative models that can create new molecules, new drugs, new compounds, so that we can actually discover or create things slightly faster and you know, cut it on these very costly research and development processes. So here is another example. The story of antimicrobial peptides. If you look at on this slide is one of the projections of the World Health Organization that antimicrobial resistance is going to be a number one leading killer on this planet by 2050. And creating new antibiotics takes a really long time and it's very costly. So we sat down and we worked on it for probably over two years. And with the goal of creating an artificial intelligence framework that can create peptides that have antimicrobial properties. So on this little helix that you see on the screen is the first peptide that we created. And we simulated and we developed and actually tested in lab. And it turned out to be really working great. Which kind of brings me to to where we are today. And the COVID pandemic. And as soon as as the crisis happened, I think the very first thing that that occurred to us was well, gee, can we now use this to try to create potential drug molecules for COVID. And we did it. It took us about a week or two to fine tune the algorithms to to apply to a different problem. And we ended up creating over 3000 entirely new molecules, small molecules that can be used in drug development. We released them publicly, we gave them to the world in a web page that has a very nice beautiful molecular explorer that research can use to study these molecules, compare them, see their properties and eventually with a hope that one of them may end up taking us to the drug. So if you look at this picture on the screen, this is lots of little bubbles, we kind of started with, you know, hunting diseases or finding which animals are disease carriers. We went on to try to predict which new diseases or which epidemics are going to happen to repurpose drugs to treat cancer to all the way to create entirely new drugs. And what you see is actually a footprint of an entirely new application area, we call it accelerated discovery. It's a thought that one day, we will be making discoveries and doing science with the help of AI and computing in the way that we've never done before. And this brings me to the end of this presentation. I think we find ourselves in this interesting moment in technology when we are beginning to use more and more of this thing we call artificial intelligence and we're beginning to think about how to put it to use in our businesses, in our industries, in our personal lives to make them more efficient. But on the other thing, how about we focus this on things that really matter, things that matter to all of us? And how about we directed the development and the evolution of these technologies towards things that can really move the needle for this world? Because when we do so, and I hope that I kind of painted this opportunity, when we do so, we will end up developing new solutions, better solutions, solutions that truly represent the future of technology and the future of this world. Thank you for listening.