What you need to know about data fluency and federated AI

• Read original article here

Sharecare is a digital health company that offers an artificial intelligence-powered mobile app for consumers. But it has a strong viewpoint on AI and how it is used.

Sharecare believes that while other companies use augmented analytics and AI to understand data with business intelligence tools, they are missing out on the benefits of data fluency and federated AI. By using federated AI and data fluency, Sharecare says it digs deeper to find hidden similarities in the data that business intelligence tools would not be able to detect in health settings.

To gain a deeper understanding of data fluency and federated AI,Healthcare IT News sat down with Akshay Sharma, executive vice president of artificial intelligence at Sharecare, for an in-depth interview.

Q: What exactly is federated AI, and how is it different from any other form of AI?

A: Federated AI, or federated learning, guarantees that the user's data stays on the device. For example, the applications that run specific programs on the edge of the network can still learn how to process the data and build better, more efficient models by sharing a mathematical representation of key clinical features, not the data.

Traditional machine learning requires centralizing data to train and build a model. However, with edge AI and federated learning combined with other privacy-preserving techniques and zero trust infrastructure, it's possible to build models in a distributed data setup while lowering the risk of any single point of attack.

The application of federated learning also applies in cloud settings where the data doesn't have to leave the systems on which it exists but can allow for learning. We call this federated cloud learning, which organizations can use to collaborate, keeping the data private.

Q: What is data fluency, and why is it important to AI?

A: Data fluency is a framework and set of tools to rapidly unlock the value of clinical data by having every key stakeholder participate simultaneously in a collaborative environment. A machine learning environment with a data fluency framework engages clinicians, actuaries, data engineers, data scientists, managers, infrastructure engineers and all other business stakeholders to explore the data, ask questions, quickly build analytics and even model the data.

This novel approach to enterprise data analytics is purpose-built for healthcare to improve workflows, collaboration and rapid prototyping of ideas before spending time and money on building models.

Q: How do data fluency platforms enable analysts, engineers, data scientists and clinicians to collaborate more easily and efficiently?

A: Traditional healthcare systems are very siloed, and many organizations struggle to discover the value within their data and unlock actionable trends and clinical insights. Not only are data creation systems and teams isolated from data transformation systems and teams, but engineers and data scientists use coding languages while clinicians and finance teams use Word or Excel.

The disconnect creates a situation where the data knowledge is translated outside of the programming environment. The transformations between system boundaries are lossy and without feedback loops to improve an algorithm or the code. Yet, all stakeholders need early and iterative access to the data to build health algorithms effectively and with greater transparency.

The modern healthcare stack facilitates the collaboration of cross-functional teams from a single, data-driven point of view in Python Notebooks with a UI for non-engineering partners. Building AI models can be time-consuming and expensive to build, and it is essential to hedge your bets by getting early prototype input across domains of expertise.

Data fluency provides an environment for critical stakeholders to discover the value on top of the data or insights and in a real-time, agile and iterative way. The feedback from non-engineering teams is immediate and can help improve the underlying model or code in the notebook instantaneously.

Each domain expert can have multiple data views that facilitate deep collaboration and data insight discovery, enabling the continuous learning environment from care to research and from research to care. Data fluency works with cloud-native architectures, and many of the techniques can also automatically extend to computing on edge, where the patient and their data reside.

Q: Why do you say the future of analytics in healthcare is federated AI and data fluency?

A: Traditional analytics in healthcare is rooted in understanding a given set of data by using business intelligence-focused tools. The employees using these tools are not typically engineers but analysts, statisticians and business users.

The problem with traditional enterprise data analytics is that you don't learn from data; you only understand what's in it. To learn from data, you have to bring machine learning into the equation and effective feedback loops from all relevant stakeholders.

Machine learning helps surface hidden patterns in the data, especially if there are non-linear relationships that aren't easily identifiable to humans. Proactive collaboration at the data layer provides transparency into how the models or analytics metrics are built and makes it easier to unravel bias or assumptions and correct them in real time.

Federated AI and data fluency also address the barriers to data acquisition, which are often not technological, but instead include privacy, trust, regulatory compliance and intellectual property. This is especially the case in healthcare, where patients and consumers expect privacy with respect to personal information and where organizations want to protect the value of their data and are also required to follow regulatory laws such as HIPAA in the United States and the GDPR in the Eurozone.

Access to healthcare data is extremely difficult and guarded behind compliance walls. Usually, at best, access is provided to de-identified data with several security measures. Federated AI and the principles of data fluency can share a model without sharing the data used to train it and address these concerns. It will play a critical role in understanding the insights within distributed data silos while navigating with compliance barriers.

The privacy-preserving approach to unlocking the value of health data is crucial to the future of healthcare. The point is to improve healthcare machine learning adoption and understandability to drive actionable insights and better health outcomes. Federated AI goes beyond traditional enterprise data analytics to create a machine learning environment for data fluency and explainability that enables the training of models in parallel from automated multi-omics pipelines.

Twitter: @SiwickiHealthIT Email the writer: bsiwicki@himss.org Healthcare IT News is a HIMSS Media publication.