“DeepCoder” may signal the beginning of the end for IBM’s Watson.

Recently a group of researchers from Microsoft and the University of Cambridge announced the development of a program called DeepCoder. Deep Coder is a combination of an Artificial Intelligence System and machine learning system that can write its own code. Currently, the system is only able to write relatively simple programs, but the developers envision a day when anyone can simply tell DeepCoder what they want and it will then write the code to accomplish the requested task within seconds and without error!

The implications of this technology are simply profound and absolutely disruptive!

Today, much has been said about the potential of using IBM’s Watson supercomputer to solve challenging societal problems across diverse sectors such as security, finance and healthcare. However, to get the benefit of Watson, you have to be able to pay IBM to use the platform. According to recent reports, you will likely have to be able to pay a huge amount of money! (See MD Anderson benches IBM Watson…) Microsoft’s DeepCoder on the other hand, offers the potential, if placed in the hands of consumers, for almost anyone, anywhere and at anytime, to recognize a need or problem and have a working solution developed for it “On Demand”. As the authors point out, this is not meant to replace current infrastructure or personnel, but to augment and enhance current abilities.

In the health sector, the opportunities are almost endless, particularly if we think across the entire care continuum from wellness, to diagnostics and therapeutics, through recovery, chronic disease management and end of life care. Through this system, every problem essentially becomes a “Big Data” problem for which the appropriate code can be written to address the problem!

For busy clinicians trying to understand the potential implications of a new drug for a patient in the office, rather than scheduling a second visit with the patient, prior to which he/she would review a number of textbooks and other digital/print resources in order to make an evidence based opinion, DeepCoder could, write a program that searches the over 26 million research studies available in the National Library of Medicine’s PubMed database, determine which studies apply most closely to the current patient, synthesize the findings from the most relevant studies and print a one page synopsis and recommendations for the patient regarding the best initial dose, on the desktop printer sitting next to the doctor, for his review, faster than the patient can get undressed.

For the aging senior who is diabetic, hypertensive and lives alone, her adult son, who lives in another state, could tell DeepCoder to find the highest quality senior care day program, that offers door to door transportation, meals and is willing to dispense medications, that has an immediate opening and fits within his budget. Since both he and his mother have an Amazon Echo, he could simply tell Alexa what he needs and she then tells DeeepCoder.  A few seconds later, once the program is written, Alexa could then use it as a new personalized “skill”, execute the task, complete the online application after her son’s approval, schedule the in-person interview and send the required down payment to the Senior Care center in advance of the meeting.

While these examples are exciting, they are not yet possible. However it is also true, that is likely only a matter of time before these or something even better become reality in healthcare!

 

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s