AlphaZero is now the undisputed champion of Go and now of chess. It recently battled Stockfish, the former chess computer heavyweight, and in that series of 100 matches, it won 28, drew 72, AND LOST NONE.
Lets hear that again. AlphaZero, the deep learning computer originally designed to play and beat human players at Go, the ancient board game, has recently been redesigned in a couple ways: 1) to take the original game rules AND NO HUMAN EXPERIENCE as its starting point, and 2) now can receive the rules for almost ANY game (in this example, chess) as its starting point. Then the programmers set AlphaZero to play itself AND LEARN THE STRATEGIES of the game by brute force and whether each strategy led to a victory or defeat.
AlphaZero, having spent time playing itself millions of times and having discerned and taught itself the principles of chess, it only considered 60,000 moves per second instead of 60 million by Stockfish. It played smarter and faster.
“AlphaZero had the finesse of a virtuoso and the power of a machine.”
But, can it teach us its insights? No. Perhaps the most troubling paragraphs in this article is:
“What is frustrating about machine learning, however, is that the algorithms can’t articulate what they’re thinking. We don’t know why they work, so we don’t know if they can be trusted. AlphaZero gives every appearance of having discovered some important principles about chess, but it can’t share that understanding with us. Not yet, at least. As human beings, we want more than answers. We want insight. This is going to be a source of tension in our interactions with computers from now on.”
I am both heartened and disturbed by this. Heartened in that AI is on the launch pad to apply itself to all kinds of human challenges that have been difficult to solve until now. Disturbed also; how long will AlphaZero and its contemporaries need human insight and input before it’s always-accelerating capability outstrips our brains’ hardware and our ability to keep up and be relevant?
CMIO’s take? I have no take. I’m gonna wait for my auto-correct from Siri to get smart enough to finish writing this post.
There are several online supplements: additional specifics about how we conducted the program (30-60-90 day planning meetings, agendas for the 2 weeks of activity, etc), and the actual pre and post-intervention surveys.
Congratulations to Amber Sieja, Katie Markley, Jon Pell, Christine Gonzalez, Brian Redig, Patrick Kneeland, co-authors on our published article in Mayo Clinic Proceedings this week. I’ve spoken of some of the details on this blog, so I’ll let the paper speak for itself. Nice to be recognized! Coming soon: a video by Dr. Sieja explaining some of the highlights of the paper.
This article speaks for itself. Having visited the Verona, Wisconsin campus of Epic’s Intergalactic Headquarters, with its dozens of campus buildings, its massive Voyager Training Hall, and its equally massive 11,000+ seat auditorium (the largest in the state), it is an impressive site. Equally impressive is the young talent that this company recruits out of college, how quickly they learn the necessary lingo and communication skills of health professionals, and dive into the constantly evolving code-base of the Epic EHR. Privately held, this company is NOT at the whim of the shareholder, which I think is a key success factor in its market dominance.
Yes, there are tons of criticisms for this and every EHR. Yes, there is a national outcry of physician burnout, directed at EHR’s in general.
Instead, you’ll read here, that the guiding vision of the reclusive CEO and founder, Judy Faulkner, holds true today; have fun, do good, recruit smart talent, improve patient care.
CMIO’s take? Having been part of the Epic community since 2010, I’ve seen lots of innovations and improvements and look forward to many more.
Thanks to Dr. Lum for her persistence and clarity of purpose. She led a team of physician leaders and IT staffers through a complex process, resulting in a first-of-its-kind online tool:
The ability for patients to complete their Advance Care Planning and upload the results using an EHR (electronic health record) patient portal.
Why is this so important
Very few patients, much less US adults, have an Advance Care Planning document (including Advance Directive, Living Will, 5 Wishes, No CPR, or any other document that records the wishes of the patient in regards end-of-life care)
It is difficult for patients to find the form, to understand the form, complete the form, return the form, AND THEN to have clinical staff file or record the form and its wishes in a way that is easily accessible by both healthcare providers as well as patients themselves.
Patients without any advance care planning documents risk the possibility of receiving unwanted care at the end of their lives; since patients are so often non-communicative as their health deteriorates, unless the doctors and nurses have a clear statement from the patient, we must assume that the patient wants CPR and other aggressive measures, even if it borders on inappropriate. This tragedy can easily be avoided with easily available documents expressing the patient’s wishes.
And now, at UCHealth, patients can avoid this risk! From the privacy of home, patients can now log in, launch the Advance Care Planning module, complete the questions, and even upload any signed documents in regards to their wishes, into their online Patient Portal, in a permanent storage location easily accessible by the patient (or their permitted proxy) or any of their healthcare providers at UCHealth.
This is pretty cool, and a big leap forward. In fact, even without any publicity to patients (the module just appeared in the portal), over 1000 patients recorded a NEW advance care plan in the first month after this module went live. And, patients continue to sign up at a constant rate.
We hope this continues at UCHealth, and more importantly, that other organizations start doing this as well. Only 36% of US adults (studied in 2017) have an advance directive; meaning that about 2/3 do NOT!
CMIO’s take? We have a lot of work to do, people. Let’s get the rate of advance care planning much closer to 100% of US adults; we never know when we’re going to need one, and by then it is often too late.
Interesting. OMRON, maker of the blood pressure measuring device that I recommend to patients, is moving into human-augmenting AI. Purely as a demonstration project to showcase technologies, they built a Ping Pong robot that will play with you to:
Rally with you
Assess your ping pong skill
Assess your emotion from facial recognition
Use the “net” as a screen to tell you what it is thinking and doing
Coach you to be a better player, using what it knows about you
From the video, it is not infallible, but it will rally with you, it will serve the ball, it will give you a ranking, it will speed up or slow down based on how you are doing and feeling, it will note that ‘we are having fun!”
Part of the idea of “arborealization” of technology (a made-up word), this is a term I heard about a decade ago. In short, with technology acceleration powered by Moore’s law and the constant doubling of computing power, tech acceleration is NOT in just a single direction (eg: self-driving cars, faster personal computers), but in ALL directions (eg: ping pong robots, poetry-composing AI, symphony-composing AI, deep-sea diving AI, Google Duplex being able to book an appointment by telephone for you, etc).
Still, weak AI and strong AI are different things. Pointing software at a difficult single problem (Weak AI solution) is very different from building an AI that can tackle ALL problems (Strong AI). I’m reading Life 3.0, a easy-to-read NYT bestseller that is the latest foray into describing the exploding fields of AI and general intelligence.
CMIO’s take? I need one of these robots in my basement ping-pong room. Humans are so disappointing; no one will take me up on my nightly ping-pong challenge.