Duration 58:41
16+
Play
Video

Keynote Panel AI in Healthcare

Sanji Fernando
VP Technologies at Optum
+ 4 speakers
  • Video
  • Table of contents
  • Video
Video
Keynote Panel AI in Healthcare
Available
In cart
Free
Free
Free
Free
Free
Free
Add to favorites
19
I like 0
I dislike 0
Available
In cart
Free
Free
Free
Free
Free
Free
  • Description
  • Transcript
  • Discussion

About speakers

Sanji Fernando
VP Technologies at Optum
Matthew DiDonato
Senior Staff Product Manager - Artificial Intelligence at GE Healthcare
Nataraj Dasgupta
VP Technologies at Rxdatascience Inc.
Slava Akmaev
Chief Technology Officer at Scipher Medicine
Shahidul Mannan
Head Of Data Engineering & Innovation at Mass General Brigham (Partners Healthcare)

Sanji Fernando is a senior vice president at Optum, where he leads the Artificial Intelligence (AI) and Analytics Platforms team. He is responsible for developing platforms that support the design and development of leading edge AI models and analytic tools for the enterprise. Previously, Sanji was a vice president at OptumLabs and led the OptumLabs Center for Applied Data Science (CADS). The CADS team applied breakthroughs in AI and machine learning to solve complex health care challenges for UnitedHealth Group (UHG) by developing and deploying software product concepts. CADS pioneered using deep learning to streamline administrative processes in revenue cycle management and developed graph analytics tools to support provider network design, among other innovations. Sanji joined OptumLabs in 2014 from Nokia, where he created Nokia’s first data science team. His team launched the first big data computing cluster at Nokia, using cluster derived insights on user activity and engagement to design new product concepts. Before that, Sanji spent 9 years at Nokia in a variety of corporate roles with Nokia’s Multimedia Division, Nokia Research Center and Nokia Ventures. Prior to Nokia, Sanji was a co-founder and VP of Engineering at Vettro, a venture-backed mobile software company. Sanji began his career in consulting with Viant and Accenture.

View the profile

Over 21 years of industry experience in developing the vision, strategy and execution of analytic capabilities in finance and pharmaceutical domains. Experience at IBM, UBS Investment Bank, UBS Wealth Management, Purdue Pharma, Philip Morris.Core architect of the RWE and Rx Data Analytics solution at Purdue Pharma which was spun out into a new startup, RxDataScience with over $ 3.5M in seed funding. Member of the core founding team and VP of Advanced Analytics.Published author of multiple books on data science, journal articles and research papers in commercial market research, health outcomes, epidemiology and RWE Analytics. Presenter Keynote Speaker and Chairperson at over 20+ machine learning and AI-related healthcare conferences in US, Europe and Asia.

View the profile

Slava Akmaev is the Chief Technology Officer at Scipher Medicine. He is responsible for product development in precision medicine and early discovery efforts in drug development using the Network Medicine platform. He has been recognized as a leader in the adoption of the AI/ML technology in healthcare and drug development and is a frequent speaker at some of the most prolific industry events. Slava is the inventor on a number of issued and pending patent applications and has published more than 30 peer-reviewed articles in computational biology, artificial intelligence and molecular biology. Additionally, he authored book chapters and numerous scientific presentations and posters.

View the profile

Shahidul Mannan is the Head of Data Engineering and Innovation at Mass General Brigham (Partners Healthcare) – largest healthcare organization in New England. He is also founder of Bridge Innovation Institute – an organization focused on education and innovation connecting global talents and entrepreneurs, and founder Principal of MBS Analytics. Shahidul served as the Chief Analytics Officer at Human Health Project. Most recently he commercialized data and AI products at TTEC Digital as Vice President and GM of AI Platform products. Shahidul also led at Dell EMC as Global Head of Big data and Analytics – leading all data assets and building next generation Data Lake and AI driven Analytics across enterprise, and at Altisource, where he served as Vice President of Technology - responsible for all data management, big data, business intelligence and advanced analytics, developing Fintech analytics products and services. A 20 years veteran in Data Analytics and Technology, Innovation management, Shahidul has a B.Sc. in Engineering from Bangladesh University of Engineering and Technology and an MBA in Finance from Baruch College, New York. Shahidul obtained Executive Certifications in Strategy, Innovation and Data Analytics at Sloan Management School in MIT and Tuck School of Business in Dartmouth College.

View the profile

About the talk


Share

AI in healthcare panel this afternoon. We have a great time. Rule of analysts today. Thank you God to the penalties for driving and spending the time with us today. I will briefly introduce the analyst and then I'll let everyone introduce themselves and talk about their introductory perspective on the AI in healthcare what's going on in their organization and with the RN with their heading towards so we have a wonderful group of experts and Executives from

Matthew didonato director at GE Healthcare Technologies rxdatascience ink and Slava at my PTO at the site for medicine and my name is Shahid. I'm the head of data engineering and Innovation at Mass General bring him. I'll be moderating the session happy to be here and I have Begin with the great panel. So as the day goes on. I'm sure you had an opportunity to empty into various exciting as sessions in healthcare. I would just close to everyone's heart and our work

and there's tremendous work that's going on in healthcare after running behind or some time with other Industries retail and financial. Most recently Healthcare has picked up the most and growing at a higher rate than others the Investments and the BC startups. All those are great indicators of that and it's definitely for a good reason with the newer technology cloud and NLP everything coming together. Now, we are in a better stage to unleash the power of data

and then Unleash the Power of insides and build more sophisticated Advanced Energy leading to a TV. Start with an L. Hi, thanks for the laundry Fernando actually think my role has been updated yet. I used to be with Optum Labs. I'm now with Optum proper and I read a eye and Emily platforms for OptumRx, which is really like like you said at a reflection of the maturity. I think of artificial intelligence where we moved out of the research phase and recognize the importance of scaling and

around Healthcare Optum. We are leveraging artificial intelligence and almost all facets of our business and and we've seen incredible impact in in some of the areas that may be a little less sexy as I pay tell people we refocused some of our efforts around administrative processes, like revenue cycle management payment Integrity risk adjustment essential parts of the US Healthcare System. And as we discussed will have to share why we think they're well suited for machine learning,

but but but we think that's a great place for us to start as an industry to go to realize some of the benefits of cheese. Thank you Cindy. Let's hear from Slava. Thank you, sir. I'm slow at Acme if I'm Chief technology officer site for medicine Cypher medicine is a small company in Waltham Massachusetts vehicle stir Boston. I've been in AI space and in health care for almost a decade and it's been really interesting to observe the development and Adoption of AI in healthcare starting with primarily some of the early application for the life sciences were

some of the research was done in using genetic data understanding how to apply machine learning in artificial intelligence for primarily probably biomarker Discovery and Diagnostics the later the industry transition to using a AI more in healthcare building predictive models for patient outcomes for financial operational outcomes and cost-effectiveness, and now it's becoming more and more present in real world. Evidence and other types of Digital Data that we observe coming on the market almost every day. It's interesting to

say the greatest Basenji here because Optimus one of our investors inside for medicine and I think the interest in the industry is really in advancing Precision medicine advancing Precision Therapeutics. Now, we're going to be on Precision Therapeutics where there's a patient stratification before we even develop Target sand and drugs and then we're going to clinical trials with a specific patients the population of a particular disease causing these approaches in both of these areas. I think is very important and we're driving towards Precision medicine obviously

and oncology, but there are multiple diseases have been traditionally Untouched by man sick neurologist and specifically are immune is still sore wide open and we're trying to make progress in rheumatoid arthritis. IBD and other areas in Oroville Excellent exciting stuff and Matthew do not have to go next. Thanks. So I'm Matthew did not a product team with Angie Healthcare building at Rai platform. And when I think about AI in healthcare, the biggest thing that popped into my head is just what a huge opportunity is. I might think similar to the incident over just gave I

ain't driving towards Precision medicine from a number of different angles is exciting Thug Givenchy Healthcare is interested in in the internment as we're collecting all of the right inputs to do that in building the right models, which I think is a big and challenging effort that will span many of the traditional particles within Healthcare in that interim. I think it's important and I tried to Highline just in my car talk earlier. Getting a right and inserting it into work clothes really is key. And that's something that I think it's just really interesting for the interest rate

overall. I think it's exciting to see how regulatory Frameworks are evolving evolving based off of the technology and also the peace it with the pace of which the technology sort of Kim produce results. The question is always how can we go faster certain certainly data collection is difficult in healthcare. Not only because it's I load but because inherently there's patient information and sensitive information there. So my overall outlook on and Health Care is very very positive because there's such a big opportunity but there's clearly

some hurdles that we have to systematically collectively as the companies that are doing it to to get towards that big promise of precision Healthcare. Find Matthew thank you. Couldn't agree more. I think we are still scratching the surface as we are wrapping up in a great day. Around the tip of the data scientists. And then I feel there's a chance to go away. So I just turned off but let me just finish up the introduction and then I'll get to it. So thank

you, cuz I stay the same over 20 years of experience in the field. Working in high frequency trading analytics and about seven years ago. I moved into farm and what I realized or object was that the potential capabilities already built up as far as like those kinds of Technologies and Platforms in the pharmacy sector. So what ended up happening was using some country popular and some investors that has 20,000 startup into commercial entity of its own and academically. I have a masters from Imperial College.

Ai and machine learning problems with various pharmaceutical organizations and Baltic organizations need to walk different companies that do you end the various types of products. We do from commercial to clinical depression finder, but not be discussed. Perhaps I'll get an opportunity. Thank you. Great. Thank you enough. And I think at some point we'd like to go to hear from you how you would compare your experience and the AI progress that you have seen in the financial sector and now trying to implement similar in the healthcare sector

and I'd like to hear from you. What's the forecast of today? And what are the key drivers or pinpoint that business or organization strategy perspective you're focused on and using AI to augment or enhance that goal ever made that call and if you can also refer training in your organization or in the industry that you find very much relevant for what you're trying to do in your organization and start with Matthew. I guess your perspective on that. I think the first tackled the question on focus of AI within my organization

here. So it might seem Works across all of GE Healthcare and I think they like the Red Dead tadpoles all the different G efforts together a sort of making GE devices which but you know are almost in every hospital across the globe in some form or another making those devices easier to use better images duesing amount of training it takes to use them increasing the consistency between operators. I think making our devices easier to use it is really that the core

of what she's trying to do with a i and build on that eventually towards incorporating our information to the drive towards Precision medicine as I mentioned. I think that there's clearly Driver's both from Clinical needs. I think that people want a better experience and if we think about what we've gotten used to as consumers, if if you think about the phone that you pull out of your pocket and how seamlessly and how included that is used how it gets better over time. I think that's the same experience. We want to

drive with Qi devices. We don't expect regularly that Apple Google consumer-facing companies will always call out that I've been years. We wanted to be in the underpinnings of the software in the experience you're having with the device in and that's what we're trying to drive with you you sort of across the organization as far as any like critical efforts or or things in the industry that. I think you're interesting one of the thing that one of the things are really sore. is evolving

thinking around AI operations within Healthcare and what I mean by that is sort of finding ways to take advantage of the fundamental differences of deep learning and the most recent incarnation of AI has and doing that in a way that is Regulatory Compliance in in make sense what I mean by that is deep learning in particular differs from other forms of AI in the fact it decides what features what data what pieces of information are important or not and that means

Almost always more data is power. Certainly. There is a difference between good data and and bad data and levels of curation. But generally the more information you have the more deep learning can take advantage of that. And so what's exciting mean industry is finding ways to take advantage of that in healthcare that that also keep patients safe and their information protected in private. And I also think there's interesting conversations about model and data drift in and how when you want something Things may be true about the macro environment

of about the information coming in that shift overtime. How do we account for that? So I think that's an interesting mashup of the constraints in the regulated environment that is health care and also the key aspects of AI in its most recent Incarnation and where those two intersecting. I think the industry is starting to really sort of sort out the nitty-gritty details and that will probably take years to evolve into its final form, but it is an exciting place to be Excellent. That is great to hear and thank you Matt. Are you

asking follow-up question and around from an organization perspective. Do you see a GE panel for leveling at the AI and David are investments and initiatives or do you see it's increasing at still think it's an interesting question and I'm I'm going to answer it with a sidestep but I think it's actually the right answer which is why I'm calling it up. So I think that the investment is increasing but not because of a specific groups or individual pieces of the organization. I think the investment is increasing because it is the fundamentals of almost all of the products that were

building now, we should everything should be a high-powered and so as long as we're investing in in growing the business or investing in AI if that makes sense Totally couldn't agree more. I think that is becoming more and more of the norm becoming the various types of new capabilities are applications are becoming data and share. Yeah, you know. Maybe just a little bit of context. My answer, you know Optum is a brand that maybe we may not know but we're a very large Enterprise in the u.s. In the US Healthcare Market serving

patients providers other insurance companies. Not just our sister company United Healthcare, which is a little bit more well-known local Alex state and federal government. And and and so in that regard We Touch almost every part of the US Healthcare System. We today we have an employee more positions than anyone in the US. We have a large ambulatory care business and we are now we have a very large Pharmacy benefit business and I say all that because the

initiatives that we explored. Add a range of things, you know things that you made typically expect us to explore like clinical decision-support genomics things like that are all the way in our research. Our technical teams are looking at how machine learning can be used to improve Ichi all processes in our database sources across our data centers in so in some ways it's very pervasive and we're excited about that. If I think about what's happening in the industry that's interesting to us is that with with with Burger specific examples of

them, but many of them and Anna's we especially look at high transaction High processing workflows like save I recycle metal. We introduced artificial intelligence. The exciting part for Isis is is recognizing now that these are not science projects are real important capabilities in our products and the necessary model operations that we need to do and Implement. I think that I just mentioned things like modern for data jerk retraining ensuring we are evaluating models for appropriateness In fairness things like that

important now that these models and these Solutions are becoming a core part of how we go to market in offer our services to the range of customers we serve Excellent. Thank you for sharing that make sense. I would move to Strava. If you want to share your knowledge that we use at Cipher for Discover. It came to us from Northeastern University and scold Network biology. It's a platform that analyzes if you will dig experimental biological data and come up comes up with the inside that is relevant for precision Therapeutics and precision medicine on top of that. We

analyze historical molecular data and willing get together with the surf top of topological data said that we get on that road biology and that leads to the discovery of patients are populations just use subgroups and driving towards Precision Therapeutics. The biggest challenge that I think we have in our space is data that is essentially molecular, right? Panda falling on on another Rush sentiment about dating later. I tell you I've had some experience or Training Day to you as well. When you look at the trading data, it is so precise

and it's so accurate, why'd you can get trade information? That's per second and then you get the exact date of that has no their ability in it. It's really really well set up for neural networks are deep learning all the types of statistical analysis in machine learning and AI when you go to lifestyle choices Life Sciences going to begin to look at Mass spectrometer data, for example, or you're looking at the next gen sequencing data on an Expression data. It is so noisy compared to what you see in digital world or trading data, right? It is noisy. It has

biases in his batch effects. It has other types of deficiencies that are really difficult to to deal with when you are working with an algorithm like Laurens Road, because you you you know, you're a trained algorithm to recognize features that are irrelevant to the clinical outcomes or to the desire to bite through these things in the AI and machine learning is a is in the center of our product development. We use it for everything that we do. We just recently launched a product and rheumatoid arthritis. That's cold

plasma Ray. It's a diagnostic test that uses an algorithm specifically random Forest. Calculate the score for a particular individual patient to whether they are likely to respond to no respond to the infant that I am feeling of therapy incredible application that is practical that we use on a daily basis now patients have access to it and its uses machine learning algorithm everyday calculating the score out of multiple variables are features on top of that as I talked about therapy. We begin to unravel this

process of precision Therapeutics where instead of going and initiating clinical trials that are looking for evridges in hundreds of patients, whether we see an effect of a very small size and potentially we have twenty 15% of people responding to the therapeutic and we need to enroll comments of people to actually observed statistically. We take a different approach. We identify patients through biomarker liquor signatures, and then we use that those signatures to enroll patients in our Phase 1 Phase 2 trials that are more likely to respond to therapy in the way. We can

dissect a complex diseases such as for example, IBD disease surf take away patients that are not likely to respond to our investigation of therapy and really minimize the size of clinical trial so it will focus on 15 20 30 individuals. We know that there are more likely to respond and let the effect size of the trial that where I know he's going to be pretty significant with small number of patients all in all our approaches are nested in AI machine learning. We use random Forest as I mentioned. We use neural

networks. We use Bayesian approaches as well and other types of statistical learning methodologies that are commonplace in in the space. I can see that Facebook. I know you're so nice other phone providers side and a v work with our clients right now organizations and you know, many of the participants Year award for dawgnation Cespedes SS what I said, they just had like off to him. I could be a Tetra so In terms of in the song the recent project done. Nothing to start a standout one is on vacation finder. It's a very critical topic, especially when it

comes to rare diseases because you know, if you think you're going to have met you one or two patients out of a million patients who qualify for the free of disease, it's a non-trivial problem and typically in that's something that we do in conjunction with the HR records as well as structured data, like claim status is for a rare disease company in the price of the value of finding patience is extremely high and finding those patients, even when the patient then says don't know they have the rare disease it started for challenging know if it's a sensitive

and balance problems. So that's one of the second one is Adverse Events. So predicting what kind of address even swallow a sequence of Adverse Events and that's typically from clinical trials data sets. Give me all the same to other ones who know their there. They're more on the scientific side, which is Prince's. There's something we could do called quantitative systems pharmacology and maybe like very briefly mentioned that is based on the preclinical state instead of testing a new drug on a patient. It is possible to model the patient using systems of all, these are

ordinary differential equations. It's a very sophisticated method but it started becoming an industry-standard. So those are some of the ones that they've been involved in this a real world project at very large pharmaceutical organization. It's not a lot of ending words and actually bad Reddit war going on in on the organization's. Thank you all for sharing app for the next question. I would actually start with natraj him. So we got his video of question is that you see in the industry

as we are trying to Mark Delong managing the big data and filling out new innovation of which Ai and then running into specifically your work area organization. What would be the top three or talk to your pee that you would identify as a biggest challenge and how you're going about it? Thank you for the question is a very pertinent question, especially when it comes to this topic in the healthcare sector, right? Because the guy came from Finance data was available like I coming to Pharma and then He's such a huge problem. So I

think data governance is one of the biggest challenges that you know, if we come across want to see if they'd Access Data sharing data restriction. So not only you know do not have access to certain data says friend since they fit CHR records of the pharmaceutical organization. A lot of times. You might not be able to get get the actual doctor's note data-sharing to even after you have access to the data. Can you share it internally with other organization members and overall data restrictions, you know, like all the process of starter catalog and catalogue in the

data brother because when a large organization have so many different date of vendors nobody iqvia Flatiron truven marketscan all these different date of Ender. Under so many different types of data sets the organization ustellar proper make a repository of this data sets was supposed to be able to the right people are able to find find the right. They just said anyone that has resources. So 1 is machine learning and AI resource where you have a certain level of Competency, which requires let some deep learning understanding of the

quantitative fundamental principles of AI where is and then have a C-section with biosciences. So there's other phone but it was and get them together to work together is one of the biggest and just one side wants to give me the very basic level one side wants to use Python. The other one is just ask for all their work, which is totally understandable, but I'm into a Big Lots too much of the main once and I'll say they taxes and resources would be the talk to in my mind. Fixing perspective. Thank you. I think certainly we hear anything

around specially having HIPAA Phi and all the compliance regulations coming or having all those are good reasons are much more are tame. constrained in some ways and careful about. Yeah, it's interesting and he mentioned the Flatiron we all know about Foundation medicine and this a relationship that they've developed with flat iron to get access to data that combines molecular data and Healthcare data from EHR systems for patients in oncology. I think that's a great example where the industry is moving towards what we're trying to do at Cipher is to

create the same type of structure in autoimmune disorders. We are accumulating observe industry-leading databases and data sets of molecular data in rheumatoid arthritis and IBD and we are linking the data with some of the EHR and claims data that is present already in healthcare by working with multiple organizations on the healthcare side primary physician organized Healthcare systems and Data Partners, you know the number of companies that links or if there are the betta together and in in in in in

significant numbers right by putting these types of data sets together. Maybe it's registered a that maybe these data coming from EHR systems. Maybe it's a molecular data coming from genomics analysis would believe that by integrating. These data says together in analyzing it on mass and we're talkin about not just you know, 200 300 patients were actually looking at things of thousands of patients integrating get together and using Advanced machine-learning Matheson AI putting it on

biological Pathways and understanding how real for molecular biology and some of them Electric Feel lyrics driving real outcomes that we observe in the clinic. Then that allows us to create interventions and do a lot of the Do it between relation to work that used to be done in laboratory. So we can actually do simulations in perturbations in our system and understand how specific targets can affect patient population. I was just leading to change in the phenotype would be Countess really are for the rebuilding the system in autoimmune diseases and I

think in the next year or two will be able to identify most promising and and relay it targets the high likelihood of elevation in deep foundation in biology in multiple diseases in in the baby in space. Excellent. Thank you one when I dad is like I think we all hope and intend that organizations are transformed by that we can think and ever do what we do differently and maybe more effectively in NM or performing Lee and so one of the things that I think is a big challenge for all of us.

It is working through. What is the change management for a large organization to understand and embrace mnadopt artificial intelligence and machine learning some of these more advanced capabilities all of it. Like, you know when I see someone in a technology background, I've had a healthy dose of skepticism. Maybe when I hear the word change management and And I realized how important it really is because you know, we're not running into Roblox for people are not saying I don't want to use our person tells a machine-learning are this is not something one. It is more the

opposite of folks are excited and engaged across our Enterprise to do this and but now we we want them to be up skills in a way to understand when it's appropriate when it's well-suited to use something machine learning. When do you have the data when you have labels prefer to do things going to supervise Rd Manor whenever possible and Four Oaks with a non-technical background, how can we give them sort of the Baseline set of skills so they can understand that and be great partners with us or with anyone there working with in designing and

transforming their own businesses. You know, one of things are responsible for is an AI learning program. that we run out of our team in like a better set of skills that we moved Beyond in adopting some more advanced machine learning neural network using some of the reinforcement learning Concepts that are emerging realize that are the product manager is operational leaders the clinicians to give them maybe not the same set of skills, but enough for them to engage with our data scientist with our machine learning engineer so we can have a

really productive conversation almost where we all can bring our subject matter to these together and really understand the problem facing two things like regulatory change and things like that but do that together and so like I said was a healthy skeptical change your answer it now, I realize it's It's Central to to Our Success at least and I'll come today. Excellent find X and Y and it's always the transformation comes with people process and Technology changes your to marchello and I have the right balance

for what do you see with this event Collision Gary just mentioned that you are also moving towards more data democratization or sub service the more and more, you know experts across the Enterprise can participate in this conversation and transformation. I think the answer is a qualified. Yes, and I think you probably know it just as well as I do and I think when we think about Dana democratization, we want to enable teams to lovers day or whenever possible

but we're in healthcare and we have a mission to serve members patients providers across the industry. So, you know, I think when I take my day to take reservation, I think we we want to make data accessible but safely accessible with the right governance and control it. Sometimes that takes more time for us to specify and plan out how we wanted design and remodels, you know, maybe with maybe we don't have the same velocity is another industry, you know, maybe social media or retail. They don't have new and important restrictions in a regulatory regime

suitcase as well. I think it Healthcare. We have a DNA for being real stewards of the data that we were asked to to use with on behalf of all these important constituents. So it's absolutely the right thing to think about who I think we do it much more thoughtfully. At least. I know we do a document and I'm sure we do it across the street. It's alright. Thank you since I've moved to Matthew. Door. So, I agree I think with everybody's answer here. So I think it changed and the ability

to find solutions that can be taken out but my customers as well as internal organization died completely agree with everybody else that touched on data. It is in every industry data is stored the bottleneck for a iTune element. I think all the Privacy restrictions that we've talked about our important that the only other take that I would add on top of that which again I agree with is we've got a number of data silos and systems within Healthcare in certainly Solutions like fire

hl7r. I hope you break down walls between those data silos and allow stick connect patient information across the different clinical systems that are you To provide their care but I do think alike. The reality is is there's a lot of existing software out there with long life cycles. And I think this is going to continue to be something that as AI developers within Healthcare sort of finding ways to efficiently circumvent that the walls that are built up in, you know, maintaining privacy and things like that, but

funny what it connect the right information to build AI is going to be really important on top of that not everything structure. Right if we look at report reports in a clinical setting are you generally unstructured and certainly a I can actually help here like an LT and solutions that actually can take unstructured information to make a structured can be useful in the end you get Solutions like attending Stanford has done a number of Publications about week labeling associated with an LP. I'm reading the reports and then applying sort of some

level of curation to today. I think there's an interesting take on the reality of a lot of existing software being out there and it having long life cycle. We're going to have to find the right Technologies to to overcome that the only thing I would add on top of all of that is there is a growing investment right at you last year Alone 4 billion dollars of the oven vest from venture capitalist is coming into Healthcare depending on how you count that. The number could be even higher. I do think that

they create the Paradigm or we've got a lot of Point Solutions and we got a lot of Point Solutions because that's kind of how AI works right now. We don't have generalized there. We have a eye that replicates and predicts Things based off of examples that scene. We're going to have a lot of companies with a lot of opportunity. He creates a lot of churn and purchasing decisions that we will have to take on as an industry. I do think the concepts of marketplaces make sense here, but they also come with their own trade-offs and

caveat to it. I think it's an interesting Dynamic to to keep track of which is not only how do we build these Solutions but How do we tackle sort of the business side of them as well? Xnxn fine. Thank you. I think you have a question on this topic to send you his mansion with regards to a I right. So I think a lot of the work that we do false in the realm of machine learning so to speak, you know, cuz technically I should be able to synthesize something for instance in

drug Discovery. It's not there in the painting dataset and you're coming up with a new synthetic compound has an example right now in terms of you know, what is angio to mention. I think it's important to bear in mind that we have to be judicious about what to use and when to use them in a lot of times you get a lot of information just said during a simple average which can be very insightful. But what tends to happen is that when a data sent its game today the Sciences the first reaction is let's prize support Vector machines around them for something fancy. Which might not be the right

side of Roi in East End Ave done a lot of kaggle competitions and you can spend the money on improving accuracy by 0.01 but it doesn't have the Bushnell in a cost-benefit paid off. So this week and second is knowing what to use for instance in Alexa is better than you say something like optimization and optimization is their field in and of itself in or if it's something to wait in a healthcare studies where there's field survey reports until 1 then so basically I think it

is very important to know what to use for a given problem within the time constraints and secondly what kind of theoretical basis to use or the other problems. Excellent. Excellent Vine any other follow-up questions from venomous on this? Thank you think we can move on to our next topic which is we all are going through the Panama Canal being in healthcare setting up our services capabilities how to optimize in operations in some cases how to best reduce capacity management how to best about capacity management

and what not. But if you hear your experience and kind of stories from your organization's what is the impact and what if any from the data analytics side or right side that has fun out of this crisis? start with Sanji Yeah, you know we have so many amazing data scientists and and statistician and Allied professionals across out them. You saw a real Groundswell of of of of work and effort and an InFocus across our Enterprise and and and I think maybe an hour I was going to like,

you know As it as someone who's more folks on machine learning. I think there's an issue a question as you can imagine all of us like hey, can I help you? Or how can we help with? What can we do today? I can this help magically solve some problems with a big ass the availability for the ventilator crises we had if I'm really honest what I struggled with his that it was a very hard thing and I don't think I personally was able to contribute in a way that was meaningful in solving problems

because we didn't have a lot to learn on you know, like this is so new. That was so much uncertainty. I'm sure there's someone smarter than me who can think about it on supervised method that might help in this instance, but some of the things we relied on are well-founded epidemiological models that look at the spread of disease some of the optimization. Algorithms that Niraj mention that I've been in place, we studied and understood for many years. And so in that regard it was it was the best of times worst of times. We're all the season try to do a

lot to help contribute, but I think it also reinforce that sometimes these methods aren't the right answer, you know, and in some circumstances we can't help by applying, you know, a really fancy neural network with multiple layers and and use auto mail to restructure it because maybe we just don't know what to learn from and how to learn from that. So that's a very personal story and I'm sure there are others who could have figured out how to do that. But I felt like we had to rely on the things we knew about the reviews for years decades and

that was a bit eye-opening for me. Next time now thank you. And then you share with the story. I think we're faced with covid-19 playing with you know on the front line as well as in the operation that capacity management. And one thing we did from each side is we had on a road map and planning to be more real time in our operational business intelligence and getting the insides and various. That is something we jumped in quickly because there was so much noise as you mentioned on capacity management for people to ventilators to bed. So we put out the new

architecture and bought all the rear end and putting the plumbing with Asian 7 streaming coming from happy and pretty quickly spun out pretty robust capability to monitor almost in a minute by minute where everything is and where the next Focus needs to be. Tremendously helpful in managing the resource constraints that we have all been going through. So if it's in the best of times and worst definitely with so many deaths in suffering but opportunity to expedite some of the Innovations and standing up capabilities like that. I wanted to

hear from Matthew your observation experience are your organization's certainly if you if you look at GE in and there's two big segments, there's like our our manufacturing division in the devices we produce and then also the the software that we produced on top of that and then really proud to be part of an organization works when there were lack of ventilators and things like that. We were finally found ways to increase production and make sure that we met the door or help to meet the world needs now. I also think that there

were cases. We've got a solution and a business line with color Command Center works at analytics in that has been helpful for customers that have had that keeping track of their their covid-19. It's than patients that have planning for their vagina to swell pulling in like population health information and making sure that sort of hospitals are ready for surgeon in demand and things like that. So we did find a couple ways to address the problem, but but I think I agree. With the general sentiment that a couple things isn't always the right solution, right?

If you're a hammer everything is a nail and I try not to be an AI hammer and sometimes the simplest solution is the better solution, but it was very interesting to me when highlighted was certainly the rate at which we can build things with a eyes really really fast, but that bottleneck of understanding what to build and having the right data to do it can be long and so even if training and a data scientist take a week a day an hour, you still got to know the right thing to stop and get I have the right to her and didn't do it. And in this case. It's hard to know our

Radiology images predictive or diagnostic but we don't know like this is a brand-new area for us to look at if they are do we have the right information and in the right Keurig identity to actually go to solution and it's highlighted to me that for as fast as this technology goes there still bottlenecks and we still can go as fast as the world changes so we should try not to do that AI hammer in find that the right solution for the right problem. Excellent inside and I think I couldn't actually as it was listening to you

reminded me of famous code that I always like when you're trying to solve a problem you spend 90% of the time understanding as to sing and analyzing the problem and the 10% for actually solving it and I think that's a perfect a framework for data science and data-driven AI Solutions. Scariest I think I mean, I agree with everyone in a day is like over doesn't really have a president. You know, it's a very unique situation where you have a two prong Supply shock and a dementia. And in other ways to question those dogs better than you had. This is colicky Nation Supply shock, you know,

it's a shock that causes the hardly expected demand going to the indirect effects of the job loss reviews consumer spending and so on. So anyway, I mean that's besides the point. But basically when you don't have a president when there no models that have been done for her and any situation. I mean, I just feel like maybe two or three papers Witcher based on H1, N1 long years ago which talks of a situation like this because it's just so you need you know, what happens in these circumstances is a switch to basic research, which is what has been traditionally successful and

it's really not that much of a time for experimentation like I need to move fast but he is overtime in there with me different methods that will come up like capacity planning and someone but I just want to reflect on a slightly different sort of Aryan farmer that is that has adopted to this which is only found in marketing and digital marketing. Especially from the commercial research site. Right? So those are in Ms. Paint, I'll send in Marathi positioning data side. They would understand basically since the Redskins will visit the doctors and so on promotional mix modeling has

God totally new update. You know, how do I get messages across to position because Stanley no, no, no, no, no no longer in call planning stage. I'm no longer making visits to the positions, but I have this method to recharge to optimize my message message. So I think in that regard what we have seen across these different organization that huge push towards omni-channel. As well as areas such as such as digital started shifting and there would be some lasting effects that will

injure of the Prevail after the crisis. Sadlier connect more. Thank you. Thank you for the inside. This is been great discussion and we could go for hours. It's a panel and I think we're getting started. So I would like to go around and hear from you in general and fuel closing statement. And where do you see organizations and your organization and the industry going forward in this very exciting space and a reminder to where are mentoring station with the speakers that is going to be available for healthcare

Finance automated world leadership all those topics. You think you have listed for some with the speakers of that conference. So please jump into that after this station that starts at 3:55. Eastern time at 12:55 p.m. Pacific time to just wanted to put that out there. I'd like to hear your thoughts are closing statement and parting thoughts on the industry starting with you Sunday. No, thank you is it's been great panel. And and I think everyone on the panel is has been

his really shared a lot of great expertise. You know, my thoughts are are probably destroy brought. I think you've heard a lot about the promise and you know, the limitations of artificial intelligence today. But if I look out over a 10-year span and an impact on Healthcare, I'm optimistic and hopeful that we're going to see tremendous not just change for change today. But but hopefully real value being driven out of this capability wear for the first time we can take complexity and maybe make it

make it work for us in a way that successful an important but I'll take time, you know, it'll be take time for us to do that dog flea and safely, you know, this is probably a marathon not a Sprint, but I'm going to be part of it and and Discussions like this validate some of the things that we've thinking about but also recognizing that you know, it's not a hammer as it doesn't solve everything in text a lot of work and time and effort to to get this right for everyone, but I'm going to be part of that conversation. Excellent. Thank you Cindy

great insights and thank you again for joining looking forward to continuing the conversation. I would like to hear from Matthew. What are your thoughts and thank you again. And I want to thank everybody involved with the conference first. It's fantastic and I'm excited to have the opportunity to just be with everybody today might take away would-be. I really am. Optimistic and excited about the advancements in healthcare going forward I think will make for better lives for all of us. It will make for a more efficient system

overall it in the world will benefit and it will take time. But but I'm very optimistic. I think my only asked would be to those that are listening aren't as familiar with the underpinnings of a I just spend some time looking at it. It is something that is going to drive almost every industry going forward and it is not foolproof are Pisces and bad models in in it is not a Magic Bullet always and so the more knowledgeable you are the more you understand and be able to weigh in on sort of how the

world is evolving. So, that would be my ask. Excellent, excellent advice and certainly as we say that's coming or we are going through and there's tons to happen and it's going to be reflected every respect for life and love to do at the pace of innovation in healthcare. Certainly. Thank you Matthew and we want to hear from not Ranch. Scary and make it quick. So, you know my hope is perhaps in a 10 years 20 or 30 years from now the next time there's a crisis we going to have an AI machine that's going to

find a box within a week. Maybe not that big but someday in the future, you know, like so when I joined from financing to Pharma, I saw it where there was no deep learning 2013 but deep learning was there like your network was down there on GPU. So there is there has been some catch up but it's come a long way and you know, like it's the development of the evolved Evolution that we're witnessing. We're in a scientist or making more and more easier for AI or machine learning understanding is benefits and limitations and I think it's sort of a journey in very early

stages and I'm quite confident, you know, like communist hands when the next Crisis comes it's going to be much more the Outlook. Will you much more promising than it is to Daiso? Is August a toll open in that? That's my Southern believe in? thank you for the wonderful insights and that we that I think

Cackle comments for the website

Buy this talk

Access to the talk “Keynote Panel AI in Healthcare”
Available
In cart
Free
Free
Free
Free
Free
Free

Ticket

Get access to all videos “Global Artificial Intelligence Virtual Conference”
Available
In cart
Free
Free
Free
Free
Free
Free
Ticket

Similar talks

Nataraj Dasgupta
VP Technologies at Rxdatascience Inc.
Available
In cart
Free
Free
Free
Free
Free
Free
Hudson Mahboubi
Sr. Manager Of Data Science at Workplace Safety & Insurance Board
Available
In cart
Free
Free
Free
Free
Free
Free

Buy this video

Video

Access to the talk “Keynote Panel AI in Healthcare”
Available
In cart
Free
Free
Free
Free
Free
Free

Conference Cast

With ConferenceCast.tv, you get access to our library of the world's best conference talks.

Conference Cast
561 conferences
22100 speakers
8257 hours of content