Machine Learning for Medical Diagnosis

Dr. Gary Fogel, CEO of Natural Selection Inc from United States participates in Risk Roundup to discuss Machine Learning for Medical Diagnosis.


Medicine, medical care, disease care and healthcare are rapidly emerging to be a human — machine collaboration.

At this point, we see both humans and machines performing tasks at which they are good at.  While this symbiotic relationship is still at an early stage, as machine learning and deep learning systems develop and evolve, it is expected that machines will increasingly assist humans with those tasks at which they are not very good.

So, what are we humans good at?

We, the humans, are good at processing information from our senses. We are also very good at perceiving human emotions. However, we are not so good at remembering things, searching for and organizing data– and we are also not good at correlating and reasoning data as well. This is where machine learning systems will add tremendous value. Machine learning systems will make physicians, providers and practitioners faster and smarter in their diagnosis. As a result, this will reduce uncertainty in their decisions, thereby reducing costs and risks and saving valuable time.

This is welcoming as there is an increasing concern that manual medical diagnostic practices, that are driven largely by human intelligence, are no longer sufficient to effectively perform complex disease diagnostic tasks on its own, in a timely and cost-effective manner.

From across nations, there are numerous reports emerging that machine learning has convincingly penetrated complex processes of medicine, especially medical diagnosis. As machine learning seems to be on its way to transforming the world of medicine and medical diagnosis, it is changing the fundamentals of not only disease diagnosis and care, but also healthcare.


When Artificial Intelligence (AI) powered, consumer-facing, disease diagnostic applications that everyone can download onto their phones have started going mainstream, the fundamentals of medicine, medical care and healthcare seems to have changed forever. It is therefore important to evaluate-

  • Whether we are witnessing a surge in initiatives on not only AI initiatives but machine learning initiatives for medical diagnosis
  • What trends are driving the deep machine learning revolution for medicine in general
  • How would an intelligent machine or non-human intelligence bring fundamental transformation of medical disease diagnosis


Across medical diagnosis, large amounts of data are being generated and transferred. Thus, it is getting difficult for human intelligence to monitor broad or specific medicine threats with the currently available tools. This is probably one of the reasons why potential risks of disease diagnosis go unnoticed. As a result patients’ lives are at risk.

Amidst that, machine learning promises to help physicians and practitioners make accurate disease diagnoses. Not only that, machine learning also helps to choose the best medications for patients and predict re-admissions. It also helps identify patients that are at high-risk for poor outcomes.

While on surface, machine learning help improve patients’ health while minimizing costs, it is important to evaluate-

  • Why is there a need for machine learning driven medical diagnosis?
  • Without the help of machine learning system, can the doctors or medical practitioners resolve medical diagnosis issues in a timely manner?
  • What does machine learning promise medical practitioners?
  • Why does clinical diagnosis still rely mostly on doctors’ expertise and intuition?
  • What needs to be done for the machine learning technology to be accepted by medical practitioners?


How do we humans determine the effectiveness of machine learning on medical diagnosis? Do we have specific tools that can help us evaluate the effectiveness of machine learning? It is important to evaluate-

  • How does the emerging intelligent medical diagnostic system that is being built on machine learning platform, perform better in disease diagnosis than the human system?
  • How effective are the medical diagnostic models built by machine learning approach/initiative?
  • What is the underlying technology and algorithm fundamentals?
  • Are the intelligent machines for disease diagnosis effective for both communicable as well as non-communicable diseases?


It is believed that despite the advances in machine learning, the heart of the disease diagnosis system will still be a team of human experts, that is doctors. Physicians and practitioners will go through the results of the machine learning analysis and ultimately identify and handle disease diagnosis risk incidents on a case by case basis.

It is important to evaluate-

  • How widely accepted are the intelligent machines for disease diagnosis?
  • Are we confident that machine learning over human analysis for disease diagnosis is trusted by the medical community?
  • Will the human component that is a doctor or healthcare expert still play an important role in disease diagnosis in the coming years?
  • How will medical systems across nations make use of machine learning?
  • Will doctors be willing to accept the conclusions of an algorithm without understanding how it achieved those conclusions?


It is believed that in the coming years, many roles of doctors won’t exist. This is because machines will be able to do many of their tasks. In addition, because of the ability to transform data into knowledge, many areas of medicine will also get disrupted. It is therefore important to evaluate what jobs will disappear due to machine learning.

It is also important to understand-

  • What will be the impact of machine learning to not only the medical disease diagnosis, but to disease prevention, treatment and overall healthcare?
  • Which areas of medicine will show quickest transformation with machine learning?
  • How will patients notice the impact of machine learning?


It is important to identify and evaluate the current applications of machine learning with respect to medical diagnosis-

  • Where can we apply machine learning/deep learning?
  • What machine learning algorithms could be used for decision making and finding appropriate medical diagnoses in a short time?
  • Where is machine learning used currently with respect to medical diagnosis?
  • How many doctors/hospitals are actively using machine learning application for disease diagnosis?
  • Do all medical diagnosis solutions leverage the same level of machine learning capabilities?
  • Apart from human disease diagnosis, what are some of the other intelligent machines applications for the disease care/healthcare should we be expecting in the coming years?
  • How will be the next generation medical diagnosis solutions look like?
  • How are practices, physicians, providers and practitioners using machine learning today to solve advanced problems like chronic diseases?
  • How can machine learning empower disease care/healthcare professionals in the fight against communicable/non-communicable diseases?
  • What are some current machine learning healthcare applications?
  • Does the medical community see the real potential of machine learning?
  • How many companies are developing machine learning based solutions for medical diagnosis?


Disease diagnosis seems to be a relatively straightforward machine learning problem. It is important to evaluate that, when technology exists for coding both symptoms and conditions, and there are large data-sets of training data available, why is machine learning not more widely used for medical diagnosis.

To apply machine learning to communicable/non-communicable disease data, it’s important to understand the different value of the data that will be used to build machine learning models. It is important to evaluate what are the laws governing healthcare data, and the availability of healthcare data in general.


It is important to evaluate-

  • What’s the biggest obstacle to achieving meaningful benefits of machine learning for medicine? How do we get past that?
  • What are some legal constraints of putting medical decision power in algorithms?
  • What are the obstacles for use of machine learning in medicine?
  • How to achieve cooperation and collaboration?

As we speak, machine learning/deep learning and AI are transforming the disease care / healthcare industry. While they improve patient outcomes and change the way doctors think about providing disease care as well as health care, they also bring complex challenges and security risks that needs to be managed effectively.

It is important that we understand the risks associated with the machine learning for medical diagnosis.

For more please watch the Risk Roundup Webcast or hear Risk Roundup Podcast

About the Guest

Gary Fogel is Chief Executive Officer of Natural Selection, Inc. He has spent most of his career as a data scientist applying computational intelligence and machine learning to medicine, industry, and defense. He is also Adjunct Faculty at San Diego State University, and co-founder and managing director of Theragence, Inc. He is currently Editor-in-Chief of the journal BioSystems, serves on the Administrative Committee for the IEEE Computational Intelligence Society, and is a Fellow of the IEEE. He received his B.A. in biology from the University of California, Santa Cruz, and Ph.D. in biology from the University of California, Los Angeles.

About the Host of Risk Roundup

Jayshree Pandya (née Bhatt) is a visionary leader, who is working passionately with imagination, insight and boldness to achieve “Global Peace through Risk Management”. It is her strong belief that collaboration between and across nations: its government, industries, organizations and academia (NGIOA) will be mutually beneficial to all—for not only in the identification and understanding of critical risks facing one nation, but also for managing the interconnected and interdependent risks facing all nations. She calls on nations to build a shared sense of identity and purpose, for how the NGIOA framework is structured will determine the survival and success of nations in the digital global age. She sees the big picture, thinks strategically and works with the power of intentionality and alignment for a higher purpose—for her eyes are not just on the near at hand but on the future of humanity!
At Risk Group, Jayshree is defining the language of risks and currently developing thought leadership, researching needed practices, tools, framework and systems to manage the “strategic and shared risks” facing nations in a “Global Age”. She believes that cyberspace cannot be secured if NGIOA works in silo within and across its geographical boundaries. As cyber-security requires an integrated NGIOA approach with a common language, she has recently launched “cyber-security risk research center” that will merge the boundaries of “geo-security, cyber-security and space-security”.
Previously, she launched and managed “Risk Management Matters”, an online risk journal and one of the first risk publications, publishing “Industry Risk Reports of Biotechnology, Energy, Healthcare, Nanotechnology, and Natural Disasters” over the course of five years. Jayshree’s inaugural book, “The Global Age: NGIOA @ Risk”, was published by Springer in 2012.

About Risk Roundup

“Risk Roundup” is an “integrated strategic security risk dialogue” for nations: its government,  industries, organizations and academia (NGIOA) in cyberspace, geospace and space (CGS).Risk Roundup is released in both audio (Podcast) and video (Webcast) format and is available for subscription at (Risk Group WebsiteiTunesGoogle PlayStitcher RadioAndroid, and Risk Group Professional Social Media).

About Risk Group

Risk Group believes that risk management, security and peace walk together hand in hand. Though security is related to management of threats and peace to the management of conflict, risk management is related to management of security vulnerabilities as well as management of conflict, and it is not possible to conceive any one of the three without the existence of the other two. All three concepts feed into each other. Risk Group believes that the security we build for ourselves is precarious and uncertain until it is secured for everyone across nations. Tradition becomes our security-so if we build a culture of managing risks effectively it will lead us to security and security will lead us to peace!

Copyright Risk Group LLC. All Rights Reserved