 Denise welcome to ODSC India. Thank you very much, James. It's fantastic to be here. Well, we're thrilled to have you here You're going to be giving a keynote presentation tomorrow So I'd love to hear more about your keynote talk give us some insight into it But also tell us a little bit about your background if you want indeed so I am from Australia in fact from Australia's government research agency So which is a really unique position because it sits between academia and the industry in fact Passionate about translating research into product that people can use in their everyday life and we developed Wi-Fi for example So from that perspective, it's it's unique in that we we have that those examples where we see how good research or interesting research has really transformed life and Specifically in the area that I'm in so my area is life science research specifically genomics and Genomics is going to revolutionize medical research or medical care because it holds the information for your future disease risk and now we have the technological power to actually change your genome in order to ensure a healthier life So this is really interesting and exciting time to be in that space Excellent and not to kind of get ahead of your keynote, but what are some of the key Talking points you're going to go into discuss tomorrow Yes, so with the data that really comes out of life science Approaching it with traditional compute is not really possible anymore similarly with Applying traditional Algorithms to it Therefore what we need to do is we need to come up with bright ideas from the community from ourselves in AI and machine learning To really make Applications possible that you know a couple of years ago have not been possible or thought possible for example We now are able to interrogate the three billion letters of the genome to really identify Which genes can drive disease or predict your future disease risk? Which is quite amazing. I think No, that's excellent time It's over the years. We've had quite a few talks around life science at ODSC using predictive analytics to Diagnose diabetes for example predict patient outcomes with some talks recently around using AI for Drug discovery. So how much do you think the has the hype kind of run? In front of the reality of it. No, no, I think we're just starting I think we're just starting to scratch the surface Because there will be more and more data available in fact by 2025 50% of the world's population will have been sequenced. That's a staggering amount So when we typically think of big data disciplines, we think of astronomy maybe Twitter YouTube But genomics is going to eclipse all of that with 20 exabytes new data generated per year and Therefore, you know with that amount of data machine learning, you know That's that's everyone's dream and machine learning to play around with this kind of this kind of data Right, that's the original data science of scale or a big data problem, right exactly so Yeah, and now with new computing power coming online Have you looked much in the hardware side? For example is using using TPUs and other advanced hardware to help solve some of these problems Yes, although I have to say that the advances in Hadoop and Spark Might be able to eclipse that Not because they're more powerful, but because the questions keep changing in life and life science all the time Like once you've solved a certain domain area, you want to move on to the next one therefore building something that is you know based on Accelerators and things like that. We don't have the luxury of actually doing that therefore all our algorithms need to be Disposable so we can develop something specifically for that question Once we solved it move on to the next one And Spark and Hadoop allows us this flexibility Excellent and Is there much employment of open source deep learning tools in the life science at the moment like are using MX net or TensorFlow Is there applications there that you're employing? Yes, absolutely Not necessarily in the area that we're in because our data set is just too Too detailed three billion letters in the genome finding the disease gene which means that We need to have a three billion feature matrix No deep learning technology can deal with that, but in saying that the genomic Architecture is actually quite complicated. It's not only the genome, but there's an epigenome and then all of that is Folded like the two meters of the genome is actually folded into that tiny little cell So all of that complexity is not captured by Traditional machine learning methods, but deep neural networks could actually do that so from that perspective there is a game changer for Specific questions where you have a small number of features for example predicting whether a snip is going to be deleterious which is a mutation in your genome and for that you only need the Immediate surroundings of the genome Therefore looking at the genome the actual sequence the epigenome how it's translated and how it's folded and Having these layers in the deep neural network represented is quite exciting. That is very exciting And do you see a lot of people coming out of life sciences? Seeing the promise of machine learning and data science for life science and start to adopt those skills Is there is there a lot of take up on that? Yeah, although I would say it's not more than in any other discipline because machine learning is going to revolutionize You know every aspect So from my perspective we are you know in the in the midst of it Developing contributing and I think one of the key messages really from my talk will be that as a community We really need to work together Irrespective of what the domain expertise is the underlying methodologies that we develop can be applied to the different areas Wonderful and last question. How are you finding odyssey India? Yes, it's fantastic to be here Simulated already, but you know the amount of education and diversity of domain areas in the audience is quite astonishing in that usually I'm coming from academic conferences where there's a robust discussion around you know, whether the technology that you're playing is actually the correct one and you don't expect to have that kind of Healthy interactions from the audience in a conference like this, but this is actually quite unique and teasing out those really interesting conversations between the speaker and the audience and between the audience actually Wonderful to hear so Denise. Thank you so much for being here for being a speaker and looking forward to hearing your keynote tomorrow Thank you once again. Thank you very much. It's pleasant pleasure to be here