OpenAI's ChatGPT system has sent the topic of artificial intelligence through the roof.
But so many professionals across industries, including healthcare, do not truly understand how AI works – especially how the different forms of AI work.
Further, there are a variety of acronyms floating around out there in the tech space: AI (artificial intelligence), ML (machine learning) and now FL (federated learning). But what’s the difference between them, and how does each relate to healthcare?
To get a primer on this important subject, Healthcare IT News talked with Ittai Dayan, CEO and cofounder of Rhino Health. Rhino Health is a vendor of a platform designed to enable developers and researchers to analyze data, create AI models and deploy them.
Ittai is the author of a highly diverse clinical federated learning study, EXAM (EMR CXR AI Model), published in Nature Medicine last year.
Q. What is AI, and how is it used in healthcare today?
A. Artificial intelligence refers to the ability of machines to perform tasks that would typically require human intelligence, such as visual perception, speech recognition, decision-making and language translation. AI systems can learn from experience, adjust to new inputs and perform human-like tasks without being explicitly programmed.
In healthcare, AI is being used in a number of ways to improve patient outcomes and streamline medical processes. For example, AI-powered diagnostic tools can assist physicians in identifying diseases and conditions based on symptoms, medical history and other patient data.
AI algorithms can also be used to analyze vast amounts of medical data, helping to uncover new insights and treatment options. Additionally, AI can be used to develop personalized treatment plans, monitor patients remotely and improve the efficiency of clinical trials.
AI is helping healthcare providers to make more informed decisions, improve patient outcomes, and provide more efficient and effective care.
Q. Now, let's drill down. What is machine learning, and what can it be used for in healthcare?
A. Machine learning is a subfield of AI that focuses on the development of algorithms and statistical models that enable computers to improve their performance in a specific task. In contrast to traditional programming, where rules and logic are explicitly defined, machine learning algorithms are designed to automatically improve their performance by learning from data.
There are different types of machine learning, including supervised learning (labels define the ‘ground truth’), unsupervised learning (no labels), and reinforcement learning (the machine learning algorithm learns from ‘experience’), each with its own strengths and weaknesses.
In healthcare, machine learning is being used to improve a wide range of processes and outcomes. For example, machine learning algorithms can be used to analyze vast amounts of medical data, such as electronic health records, to identify patterns and relationships that can inform the development of more effective treatments.
Machine learning can also be used to develop predictive models that can help healthcare providers to anticipate patient outcomes and make more informed decisions. Machine learning is playing a crucial role in advancing the field of healthcare by enabling more precise, personalized and effective treatments.
Q. What is federated learning, and what are its healthcare applications? How is it different from machine learning?
A. Federated learning is a distributed machine learning technique where multiple participants each have their own data, and the model is trained by aggregating updates from these participants without sharing the raw data.
In other words, the data remains on the local device and only the model parameters are communicated to the central server for aggregation and updating. This approach enables organizations to preserve privacy, security and data ownership while still taking advantage of the benefits of machine learning.
Federated learning and machine learning are related, but distinct, concepts. Machine learning refers to the development of algorithms and statistical models that enable computers to improve their performance in a specific task through experience.
In contrast, federated learning is a specific type of machine learning that enables multiple participants to collaborate and train a shared model without sharing their raw data.
Federated learning can improve machine learning models in healthcare by enabling the use of larger and more diverse datasets while preserving privacy and security. Some key ways in which federated learning can improve machine learning models in healthcare include:
Improved data diversity: Federated learning enables the use of data from multiple sources, including hospitals, clinics and patients, providing a more diverse set of data to train models on. This results in models that are more generalizable and better able to make accurate predictions for a wider range of patients.
Enhanced data privacy and security: By keeping the data on local devices, federated learning ensures that sensitive patient data is never exposed or shared between organizations. This helps to protect patient privacy and security and can increase patient trust in the technology.
More transparency and trust: Federated learning enables data ‘custodians’ to maintain control over their data, and provides a simple way for them to enforce contracts and ensure transparency across the full ‘lifecycle’ of data.
Q. Please talk about your EXAM federated learning study and what healthcare provider organization health IT leaders can learn from it?
A. The EXAM study was a research project – led by myself and Dr. Mona Flores, Nvidia’s global head of medical AI – that was published in Nature Medicine in September 2021. The study demonstrated the feasibility and benefits of federated learning in the healthcare domain.
A model was developed using local data as well as data across a federated network for predicting outcomes of patients that turned up to the emergency department with respiratory complaints.
The EXAM study proved that federated learning can enable hospitals to collaborate and provide federated access to data without compromising patient privacy and security.
The study showed that the federated learning approach was able to improve the performance of the predictive model, creating a global federated model that was better than any local model, and that proved a high degree of generalizability to unseen data in a subsequent validation study.
Thus, this demonstrated that federated learning has the potential to transform the way hospitals collaborate to improve patient outcomes.
The results of the EXAM shows that there is a way to overcome some of the major challenges associated with data sharing in healthcare, such as privacy, security and data ownership. The study provides a roadmap for how healthcare organizations can use federated learning to improve patient outcomes while still preserving privacy and security.
