10 Apr Interview with Jennifer Kloke of Ayasdi
Dr. Jennifer Kloke is the Chief Product Officer at Ayasdi. For the last three years, she has been responsible for the automation and algorithm development for the entire Ayasdi codebase and led many efforts to development cutting edge analysis techniques utilizing TDA and AI. During that time, she was the principal investigator for a Phase 2 DARPA SBIR developing automation and data fusion capabilities. These have led to breakthroughs in the field and several patents. Read her full bio.
PMWC 2018 Michigan taking place June 6-7, 2018.
Q&A with Jennifer Kloke
Q: What need is Ayasdi addressing?
A: Ayasdi is pioneering the application of AI to value-based care by targeting two of the most complex problems in healthcare: population risk stratification and clinical variation management. The ability to create fine grained populations with multi-factorial co-morbidities and then to design precise care process models for those groups will fundamentally change the practice of medicine and demands AI to manage the complexity of the task.
Q: What are the products and services Ayasdi offers to address this need? What makes Ayasdi unique?
A: Ayasdi has a general purpose AI platform that is powered by Topological Data Analysis – a unique framework for machine learning that combines and synthesizes different statistical, geometric and machine learning algorithms. That platform supports the rapid creation of applications and Ayasdi has two such applications in the value-based care space. The first is the Population Risk Stratification application which identifies fine grained patient populations using unsupervised learning. The second is the Clinical Variation Management application which designs precise care process models for both acute or long term care conditions.
Q: What is your role at Ayasdi and what excites you about your work?
A: My current role is VP of Product Innovation. As Ayasdi’s first employee, I have worn a lot of hats from data science to engineering (even a little sales). What we do here at Ayasdi is immensely interesting. It builds upon my PhD in mathematics and engages my passion for computer science. More importantly, however, the work is meaningful. From powering the team at Mt. Sinai’s groundbreaking work in diabetes or UBIOPRED’s pioneering work in Asthma, we are making a material difference in the science of medicine and that is immensely rewarding. Many companies talk about changing the world – we actually have some proof points – with more in the works.
Q: When thinking about Ayasdi and the domain Ayasdi is working in, what are some of the recent exciting developments that is propelling the field forward and how will it impact healthcare and why?
A: In terms of AI, we are on the verge of a radical transformation on how we practice data science. For the last number of years, much of the work has gone into making deep learning more effective – from a performance perspective, an accuracy perspective and an interpretability perspective.
What is happening now, however, is a realization that unsupervised learning needs to occur first. Unsupervised learning allows us to find the patterns and relationships that exist in data – without having to ask questions. Given the size and complexity of the modern healthcare dataset, there are simply too many possibilities for us to iteratively ask question after question. Unsupervised learning provides a principled starting point by identifying what matters in the data without the bias associated with the creation of an objective function.
This will have a tremendous impact on healthcare as the complexity of the challenge grows daily. With patient, billing, omics and other data at our disposal finding the relationships that matter require the ability to understand the patterns and structure of the data – something the human mind simply cannot do.
