Search the KHIT Blog

Tuesday, November 27, 2018

AWS to open its machine learning certification curriculum to all developers

Found this news item interesting:

Amazon’s own ‘Machine Learning University’ now available to all developers

by Dr. Matt Wood | on  | in Amazon Machine Learning, Artificial Intelligence | Permalink |  Comments |  Share
Today, I’m excited to share that, for the first time, the same machine learning courses used to train engineers at Amazon are now available to all developers through AWS.

We’ve been using machine learning across Amazon for more than 20 years. With thousands of engineers focused on machine learning across the company, there are very few Amazon retail pages, products, fulfillment technologies, stores which haven’t been improved through the use of machine learning in one way or another. Many AWS customers share this enthusiasm, and our mission has been to take machine learning from something which had previously been only available to the largest, most well-funded technology companies, and put it in the hands of every developer. Thanks to services such as Amazon SageMaker, Amazon Rekognition, Amazon Comprehend, Amazon Transcribe, Amazon Polly, Amazon Translate, and Amazon Lex, tens of thousands of developers are already on their way to building more intelligent applications through machine learning.

Regardless of where they are in their machine learning journey, one question I hear frequently from customers is: “how can we accelerate the growth of machine learning skills in our teams?” These courses, available as part of a new AWS Training and Certification Machine Learning offering, are now part of my answer.
There are more than 30 self-service, self-paced digital courses with more than 45 hours of courses, videos, and labs for four key groups: developers, data scientists, data platform engineers, and business professionals. Each course starts with the fundamentals, and builds on those through real-world examples and labs, allowing developers to explore machine learning through some fun problems we have had to solve at Amazon. These include predicting gift wrapping eligibility, optimizing delivery routes, or predicting entertainment award nominations using data from IMDb (an Amazon subsidiary). Coursework helps consolidate best practices, and demonstrates how to get started on a range of AWS machine learning services, including Amazon SageMaker, AWS DeepLens, Amazon Rekognition, Amazon Lex, Amazon Polly, and Amazon Comprehend.
New AWS Certification for Machine Learning
To help developers demonstrate their knowledge (and to help employers hire more efficiently), we are also announcing the new “AWS Certified Machine Learning – Specialty” certification. Customers can take the exam now (and at half price for a limited time). Customers at re:Invent can sit for the exam this week at our Training and Certification exam sessions.
The digital courses are now available at no charge at and you only pay for the services you use in labs and exams during your training.
Deserves a close look (notwithstanding the inappropriate marketing hype use of the word "University")..

Recall my prior post "Data Science?"

One AWS track, "Data Scientist":

A clip from a couple of the courses:

Curious omission of "finite / discrete math."

I bring most of that stuff to the table (though my calculus is by now a 3rd of a century old). AWS claims that the courses are available "at no charge," but I'm not seeing where you can get a deeper topical look (e.g., "syllabus") without "enrolling."

e.g., Linear and Logistical Regression:

Overview from enrollment page:
How do you make predictions in machine learning? Naumaan Nayyar, AWS Applied Scientist, will lead you through the key points—specifically, linear models for regression, least squares error, maximum likelihood estimate, regularization, logistic regression, empirical loss minimization, and gradient-based optimization methods.
As an experienced stats/regression modeling analyst (large pdf **), I'll just comment that that's a lot of material to cover to competency in 8.5 hours. "Lead you through the key points..."
** From my 2003 bank white paper: "This paper documents the Custom Risk Score (CRS 2003) development process undertaken by the Risk Management Acquisitions Group Model Development Team, a rigorous analytical and modeling effort leading to the derivation of a revised and improved FNBM CRS metric set with which to effectively evaluate and segment new account applicants for optimal profitability consistent with effective management of credit risk. The study was performed in a manner consistent with statistical and credit industry best practices, and was analytically comprised of a systematic application of Factor Analysis, Cluster Analysis, and Logistic Regression methods applied to a broad array of internal and external (bureau) credit history attributes pertaining to a suitable sample of FNBM accounts.The resulting tool consists of a stratified "suite" of scorecards more closely calibrated to empirically evident incoming applicant clusters, and thus represents a significant improvement over its predecessor CRS 2001, which, while effective and statistically valid in its own right, was based on a portfolio-wide single-score logistic regression underwriting model..."
You are not gonna get functionally up to speed on all of that kind of stuff in 8.5 hours.

Is this AWS offering basically a ML/"Data Science" "boot camp?" Will an AWS "certification" be of significant independent value?

The lense is over here, bro'.

More on "data science" and boot camps:

"Are Data Science Boot Camps Worth It?"
I came to this outfit by way of this author.


On the topic of tech talent, this book (along with several related other titles) is in my queue:

Nice podcast with the author:

50 minutes, well worth your time. Transcript here.I would imagine this would not be White House anti-immigration advisor Stephen Miller's favorite book. He breaks out in a rash upon hearing the word "global."

apropos of all of the foregoing, an interesting snip from the intriguing 2018 Annual Letter promulgated by
"Because of the lack of distribution of talent (most of the top AI experts and new PhD’s unfortunately join Big Tech due to their professional-athlete level compensation packages), deep learning has barely scratched the surface of its potential applications. Over the next few decades, as more engineers are trained in artificial intelligence and as developer tools and frameworks get easier to use, we should see artificial intelligence successfully applied to problems that were previously too difficult using traditional software methods, such as self-driving cars, robotics, and drug design and discovery. AI can truly transform how we work, how we live, and even how we think." -- Chamath Palihapitiya, CEO
'eh? In that regard, we'll have to keep considering Judea Pearl's pioneering AI work:

Some other books pending blog review:

There are more. Always.


From LinkedIn news updates.
Amazon says it’s marketing software that will allow doctors and hospitals to harvest patient medical records to enhance treatment and pare down costs. The move is the internet giant’s latest foray into medical care  and follows recent acquisitions in the healthcare and pharmaceutical industries. Amazon will also develop an app that's embedded in electronic medical records and provides doctors direct links to products to market to patients.  Here’s what people are saying.

Another "holy shit!" book. Another queue-jumper.

Twenty five leading lights of AI/ML etc. Fascinating. Hugely instructive.

UPDATE: I am a bit disappointed by the lack of any reference in the book to "computational biology." to wit,
Computational Biology
Computers, at the end of the day, are machines for turning information into processed information. This is obviously a very powerful and flexible capability, and our use cases for computers have expanded far beyond the “information processing” jobs they were initially tasked with. Yet, they can’t do everything. There are many different kinds of problems we can’t count on computers and software to solve. Many of them have one thing in common, which is that they aren’t problems about information. Instead, they’re problems about the physical world ("atoms vs. bits", as people like to say). Whether we need to build matter up (make materials, design drugs, process fuels), break things down (clean up pollution, treat disease) or identify and interact with things in the physical world (sense lead in water, sequester contaminants), there are a lot of hard problems in the world that computers cannot solve, but that biology can.

For many of these problems, the highest-potential path to fixing them lies in the overlap of computers and biology: computational biology. Computational biology is an emerging discipline that generally refers to two overlapping fields: 1) the practice of taking everything we’ve learned about how to build computers and applying that knowledge to building cells as a programmable, flexible, platform with which we can do useful work, and 2) productizing and automating the tools, processes, and methods we use in the lab to manipulate biology and build living systems. Although we’ve gone through a few waves of “biotech bubbles” over the past twenty years, this time may no longer only be about wildly speculative drug development, but instead about something more concrete and foundational. We'll be able to establish biological systems as engineered, all-purpose platforms that we can put to work the same way we do with computers.

Additionally, within a few years, we’ll reach a convergence point where our recent advances will start to overlap, and eventually blend, into our existing computing frameworks and infrastructure. This will have a profound and disruptive effect on many fields such as drug design and discovery, drug delivery, precision diagnostics and healthcare, engineered materials, ecology, agriculture, and much more. We’ll be able to work with biology in ways that increasingly resemble the way we work with software: as a platform for building tools, applications, and infrastructure. This time, however, we’ll be able to do it using living systems instead of code.
Chamath Palihapitiya, CEO, Social Capital, Oct. 2018 [pdf].
Think about the implications.

More to come...

Friday, November 23, 2018

An important climate change update

From the NY Times:

U.S. Climate Report Warns of Damaged Environment and Shrinking Economy

Recall my prior posts on climate change (and apropos of my recent immediate recent posts on our California wildfire disasters).
WASHINGTON — A major scientific report issued by 13 federal agencies on Friday presents the starkest warnings to date of the consequences of climate change for the United States, predicting that if significant steps are not taken to rein in global warming, the damage will knock as much as 10 percent off the size of the American economy by century’s end.

The report, which was mandated by Congress and made public by the White House, is notable not only for the precision of its calculations and bluntness of its conclusions, but also because its findings are directly at odds with President Trump’s agenda of environmental deregulation, which he asserts will spur economic growth.

Mr. Trump has taken aggressive steps to allow more planet-warming pollution from vehicle tailpipes and power plant smokestacks, and has vowed to pull the United States out of the Paris Agreement, under which nearly every country in the world pledged to cut carbon emissions. Just this week, he mocked the science of climate change because of a cold snap in the Northeast, tweeting, “Whatever happened to Global Warming?”

But in direct language, the 1,656-page assessment lays out the devastating effects of a changing climate on the economy, health and environment, including record wildfires in California, crop failures in the Midwest and crumbling infrastructure in the South. Going forward, American exports and supply chains could be disrupted, agricultural yields could fall to 1980s levels by midcentury and fire season could spread to the Southeast, the report finds…

 I've just started studying the report. Going to the "Human Health" section first.

From The Atlantic:
"It may seem like a funny report to dump on the public on Black Friday, when most Americans care more about recovering from Thanksgiving dinner than they do about adapting to the grave conclusions of climate science. Indeed, who ordered the report to come out today?"
Let me guess.


From the report "health" summary section:
Impacts from climate change on extreme weather and climate-related events, air quality, and the transmission of disease through insects and pests, food, and water increasingly threaten the health and well-being of the American people, particularly populations that are already vulnerable.

Changes in temperature and precipitation are increasing air quality and health risks from wildfire and ground-level ozone pollution. Rising air and water temperatures and more intense extreme events are expected to increase exposure to waterborne and foodborne diseases, affecting food and water safety. With continued warming, cold-related deaths are projected to decrease and heat-related deaths are projected to increase; in most regions, increases in heat-related deaths are expected to outpace reductions in cold-related deaths. The frequency and severity of allergic illnesses, including asthma and hay fever, are expected to increase as a result of a changing climate. Climate change is also projected to alter the geographic range and distribution of disease-carrying insects and pests, exposing more people to ticks that carry Lyme disease and mosquitoes that transmit viruses such as Zika, West Nile, and dengue, with varying impacts across regions. Communities in the Southeast, for example, are particularly vulnerable to the combined health impacts from vector-borne disease, heat, and flooding. Extreme weather and climate-related events can have lasting mental health consequences in affected communities, particularly if they result in degradation of livelihoods or community relocation. Populations including older adults, children, low-income communities, and some communities of color are often disproportionately affected by, and less resilient to, the health impacts of climate change. Adaptation and mitigation policies and programs that help individuals, communities, and states prepare for the risks of a changing climate reduce the number of injuries, illnesses, and deaths from climate-related health outcomes.
Trump's failure to fight climate change is a crime against humanity
By Jeffrey Sachs

(CNN) - President Donald Trump, Florida Gov. Rick Scott, Florida Sen. Marco Rubio, and others who oppose action to address human-induced climate change should be held accountable for climate crimes against humanity. They are the authors and agents of systematic policies that deny basic human rights to their own citizens and people around the world, including the rights to life, health, and property. These politicians have blood on their hands, and the death toll continues to rise…

As the Earth warms due to the continued burning of coal, oil, and gas, climate-related disasters that include high-intensity hurricanes, floods, droughts, extreme precipitation, forest fires, and heat waves, pose rising dangers to life and property. Hurricanes become more destructive as warmer ocean waters feed more energy to the storms. Warmer air also carries more moisture for devastating rainfalls, while rising sea levels lead to more flooding.

Yet Trump and his minions are the loyal servants of the fossil-fuel industry, which fill Republican party campaign coffers. Trump has also stalled the fight against climate change by pulling out of the Paris Agreement. The politicians thereby deprive the people of their lives and property out of profound cynicism, greed, and willful scientific ignorance…
Jeffrey Sachs is a professor and director of the Center for Sustainable Development at Columbia University. 
 Read all of it.

The World Needs to Quit Coal. Why Is It So Hard?
By Somini Sengupta

HANOI, Vietnam — Coal, the fuel that powered the industrial age, has led the planet to the brink of catastrophic climate change.

Scientists have repeatedly warned of its looming dangers, most recently on Friday, when a major scientific report issued by 13 United States government agencies warned that the damage from climate change could knock as much as 10 percent off the size of the American economy by century’s end if significant steps aren’t taken to rein in warming.

An October report from the United Nations’ scientific panel on global warming found that avoiding the worst devastation would require a radical transformation of the world economy in just a few years.

Central to that transformation: Getting out of coal, and fast.

And yet, three years after the Paris agreement, when world leaders promised action, coal shows no sign of disappearing. While coal use looks certain to eventually wane worldwide, according to the latest assessment by the International Energy Agency, it is not on track to happen anywhere fast enough to avert the worst effects of climate change. Last year, in fact, global production and consumption increased after two years of decline…
In the public imagination, the coal miner has long been a symbol of industrial virility, a throwback to an era when hard labor — particularly men’s labor, rather than robots — fueled economic growth.

That idea has been central to politics. German coal miners have lifted the fortunes of that country’s far-right party. Poland’s right-wing government has promised to open new coal mines. Australia’s prime minister, Scott Morrison, rose to power as a champion of coal.

President Trump has promised, unsuccessfully so far, to revive coal mining jobs and instructed his Environmental Protection Agency to roll back rules to reduce emissions from coal-fired power plants.
That message might be welcome in American coal country, but the industry’s future in the United States is not promising…
Again, read the entire article.


I cut my white collar teeth in a forensic-level environmental radiation / mixed waste lab in Oak Ridge in the 1980s (computer systems development and lab QC - pdf). I have known since those days about this (that hardly ever gets mentioned):
Coal is largely composed of organic matter, but it is the inorganic matter in coal—minerals and trace elements— that have been cited as possible causes of health, environmental, and technological problems associated with the use of coal. Some trace elements in coal are naturally radioactive. These radioactive elements include uranium (U), thorium (Th), and their numerous decay products, including radium (Ra) and radon (Rn). Although these elements are less chemically toxic than other coal constituents such as arsenic, selenium, or mercury, questions have been raised concerning possible risk from radiation. In order to accurately address these questions and to predict the mobility of radioactive elements during the coal fuel-cycle, it is important to determine the concentration, distribution, and form of radioactive elements in coal and fly ash... [USGS, 1997]
Radionuclides in the the coal seams. Lovely.


President Trump on the report he did not and will not read:

POTUS speaks eloquently to the Washington Post, Nov. 27th:
“One of the problems that a lot of people like myself, we have very high levels of intelligence but we’re not necessarily such believers. You look at our air and our water, and it’s right now at a record clean.”
Verbatim. SMH.


From Naked Capitalism:

How Economists Impede Addressing Climate Change

Posted on  by 
On a recent post, some readers recoiled at the idea of putting a monetary value on human life. Yet that happens all the time. Courts come up with damages for injuries and wrongful deaths. Younger people with high earnings are more “valuable” than other people. And remember the Pinto? Companies similarly put a value on how much it is worth to them to spend on safety to prevent deaths and dismemberment.
As this article indicates, this sort of thinking winds up playing a role any time companies or governments look at making financial outlays. And this situation is made worse by the fact that due to reasons of cognitive bias or bad incentives, people and institutions have a predisposition not to do difficult things now. People engage in procrastination and hyperbolic discounting. Politicians find “kick the can down the road” strategies to be the best approach most of the time...
Read it. Read the comments as well. NC rocks.

BTW, see also The Atlantic's "Why the U.S. Can't Solve Big Problems."
_____________ AnthropoceneDenial

More to come...

Wednesday, November 21, 2018

Next up for California, flooding, mudslides, and more misery

It's finally raining in northern California.
Below, a photo from a 2017 CA post-fire mudslide.

Up next: flooding and mudslides and newly homeless refugee miseries -- including severe health problems upticks.
_____________ AnthropoceneDenial

More to come...

Tuesday, November 13, 2018

California is on fire once again

[11/19 update] The first, hellacious photo above is from the Paradise CA "Camp Fire" east of Chico, a good 150 miles from San Francisco proper. This post has accrued across a week as the fires developed and worsened. We expect rain beginning perhaps by the 21st and lasting several days, which will clean the air and douse the fires, but will likely also precipitate mudslides and complicate the ID and recovery of more human remains.

It's just like last year, but horrifically worse.

From The Atlantic:
Smoke From California’s Fires Is Harming the State’s Most Vulnerable
As wildfires burn out of control, they are impacting the state’s other crisis—the growing number of people living on the streets.

The deadliest fire in California’s history continues to burn, and San Francisco is filled with smoke and ash. On Tuesday, for the fifth day in a row, air throughout Northern California contained high amounts of fine-particulate-matter pollution, and the Bay Area Air Quality Management District warned that the air was unhealthy for everyone. “The public should limit outdoor activity as much as possible,” the agency said Monday, urging residents to stay inside with their windows and doors closed.

But for San Francisco’s thousands of homeless people, this warning is impossible to follow. San Francisco, like many California cities, has seen homelessness rise in recent years, as the cost of housing has gone up and zoning laws have limited the construction of new housing units. Despite an initiative passed on November 6 to tax large businesses to fund homeless services and news that the CEO of Twilio had donated $1 million to fund homeless services until the tax kicks in, thousands of people still have nowhere to go in San Francisco on any given night. As the number and deadliness of fires grows in California, the population of people negatively impacted by the air quality is growing, too…
In addition to our long-time chronic homeless population issue, we now have thousands of people newly made homeless by our fires. Same thing last year with the wine country fires. I can't imagine. 

My eyes have been burning since last Friday, and I've had a heavy feeling in my chest, notwithstanding being 140 miles to the southwest of the closest fire near Chico. Our area in Antioch smells like BBQ, the sky a chalky grey-yellow, with the sun a dull orange-reddish orb. We were all advised to close windows, bring the dogs in, and stay inside. On Sunday the manager of my Muir cardiac rehab PT clinic called to say that PT was cancelled for Monday and perhaps today as well. As I post this the death toll has risen to at least 44, with hundreds of people yet missing.

Numerous vehicles have literally melted along the roads. Thousands of structures have been reduced to ashes (some with human remains cremated inside).


Before-and-after pics from one street corner.


BTW: look under your sinks, and out on your garage shelves. Ponder all of those toxic household cleaning products, paints, solvents, landscaping, herbicides, pesticides, and automotive chemicals, etc. The smoke cloud now enveloping us is full of them in addition to burned wood and grass particulates (along with what used to be plastics).

I first came to the Bay Area in 1967 (subsequently lived in Seattle, Birmingham and Tuscaloosa AL, Knoxville TN, and Las Vegas, prior to returning here in 2013). The California population has since more than doubled, from ~19 million to more  than 39 million people. Along with population pressure, persistent drought, exacerbated by climate change has contributed significantly to the frequency and severity of western wildfires.

Tangentially, time for deployment of "exposomics" monitoring tech in fire-affected areas?
No rain in sight for us yet.


Cardiac rehab was open today, btw. Good workout.

The Atlantic has another good one up on widfires:

The Simple Reason That Humans Can’t Control Wildfires



High Stakes, Entrenched Interests And The Trump Rollback Of Environmental Regs

Since his days on the campaign trail, President Donald Trump has promised to roll back environmental regulations, boost the use of coal and pull out of the Paris climate agreement — and he’s moving toward doing all those things.

He has pushed ahead with such action even as a report by the United Nations’ Intergovernmental Panel on Climate Change released in October concluded that without much stronger measures to reduce the use of fossil fuels, a warming planet will witness the spread of tropical diseases, water shortages and crop die-offs affecting millions of people.

Supporters of the administration’s changes — some of whom are skeptical of accepted science — say the administration’s moves will save money, produce jobs and give more power to states.

But critics say new strictures on scientific research and efforts to overturn standards for protecting air, water and worker safety could have long-term, widespread effects that would upend hard-won gains in environmental and public health.

The Trump administration’s many environmental proposals vary widely in target and reach.
For example, the administration has delayed the implementation and enforcement of many Obama-era rules, saying they need time to draw up new rules or study some that are already on the books. Industry generally agrees, arguing these rules are an overreach with negative financial consequences. 

Critics fear that the delays will undermine hard-fought public health protections.

Among such efforts:
The Environmental Protection Agency recently argued it needs until 2020 to decide on a controversial Obama-era directive expanding to smaller streams and waterways the types of wetlands protected by the federal Clean Water Act. That directive might mean fewer pollutants released into tributaries of larger waterways, from which millions of people get their drinking water. But the controversial rule has been fought by farming, mining and other industry groups that say it is too restrictive.
The EPA also sought to delay by nearly two years standards to protect workers and emergency responders at chemical plants, part of an Obama-era rule in response to a 2013 fire at a Texas fertilizer plant that killed 15 people. Industry says that the rule is costly and that providing information about chemical storage at plants could raise security concerns.
In March 2017, then-EPA chief Scott Pruitt rejected a petition filed in 2007 by environmental groups seeking to ban a commonly used pesticide, chlorpyrifos, which the groups say harms health, particularly citing developmental damage to children and fetuses. The agency said it needed more time to study the chemical. 
All three of those delays were blocked by federal court judges, although the administration may decide to appeal, so final outcomes are unclear.

But one thing is clear: Everyone is likely to spend a lot of time in court.

“Folks are already lining up to challenge the Affordable Clean Energy rule, and that’s probably true for just about anything this administration does when it comes to environmental reform,” said Nicolas Loris, a research fellow at the Heritage Foundation, a conservative think tank.

The clean energy rule, introduced in August, would replace a more stringent Obama-era rule for coal-burning power plants.

An EPA analysis said the proposed rule would reduce industry costs and create jobs.

The same analysis concluded, though, that the looser standards, which would supersede the never-implemented Obama-era regulation, would cause as many as 1,400 premature deaths and 15,000 new cases of upper respiratory problems annually by 2030.

On another front, scientists are protesting new Trump administration policies they say would effectively curtail their ability to study the health effects of environmental exposures.

This spring, the EPA proposed a rule dubbed Strengthening Transparency in Regulatory Science, which would restrict the use of studies as the basis for advancing environmental regulations if researchers have not released all their raw data, potentially including medical records.

The Trump administration said this step would ensure that data and methods can be checked for accuracy, echoing a long-running argument from industry and some in Congress.

From scientists, though, reaction was immediate, widespread and negative. Hundreds of researchers and dozens of public health organizations said the proposal would quash important research into the effects of pollution and chemicals on health.

No longer would they be able to promise confidentiality of medical records to people who take part in research studies, which would have a chilling effect on their willingness to participate.

Many of the submitted comments noted that such a rule would undermine key studies that led to pollution laws and prevailing attitudes about the interaction of environmental and human health.
Case in point: the seminal 1993 “Six Cities” research by Harvard scientists linking air pollution to premature death.

That study did not disclose the identities of its 22,000 participants or their medical information.
Its findings led in 1997 to new restrictions under the Clean Air Act for fine particles, tiny pieces of soot, dust, carbon and other pollutants that get inhaled deep into the lungs, potentially causing asthma, lung cancer and other health conditions. By 2020, those rules are expected to have prevented more than 230,000 early deaths.

Scientists say the administration is handicapping their ability to do important research. The plan comes amid other efforts critics see as attacking science, such as removing information from government websites about climate change, restrictions on who can sit on EPA advisory boards and a proposal to more narrowly target safety reviews of chemicals.

“By attacking the science that talks about adverse effects on health,” the administration hopes to allow deregulation yet claim “they are not harming people,” said Francesca Dominici, a professor of biostatistics at the Harvard’s T.H. Chan School of Public Health.

The range and scope of the proposed changes has brought praise from some in industry and agriculture for loosening restrictions and giving states more flexibility. But the changes frustrate public health and environmental health advocates.

“We would like to be moving forward rather than fighting these kind of rollbacks,” said Janice Nolen, assistant vice president for national policy at the American Lung Association.

Kaiser Health News (KHN) is a national health policy news service. It is an editorially independent program of the Henry J. Kaiser Family Foundation which is not affiliated with Kaiser Permanente.
Is it too early to start drinking?


The air here remains "unhealthy." It was chokingly smoky today, with a sharp burnt odor. The wildfires body count continues to rise (63 as I write, hundreds still missing). More than 10,000 structures destroyed or damaged (mostly destroyed). Thousand are now homeless, many near Paradise sleeping in cars or tents in a Walmart parking lot.

I had cardiac rehab PT today. Pushed it hard. Felt good, but now I can feel it in my chest. Staying in the house.


More than 630 people reported as missing in the #CampFire area. More than 15,000 structures destroyed. The Bay Area air is so bad that UC Berkeley has closed for the day (as well as Contra Costa County schools). And, coming soon:

No, again, I don't like him.

From Vox:


In The New Yorker

Spot-on CNN OpEd by Tess Taylor. "In California, the apocalypse keeps getting worse."


Finished two books.

Both excellent, both broadly germane to KHIT topics.

Starting two more (notwithstanding that I still have a number of others in progress. Amazon "Buy with 1-click" is gonna BK me).

So much to learn. Love it. Given that we pretty much still have to stay inside, I'm grinding away with my studies,

Topically apropos, I read a Naked Capitalism post, which led me to this (and subsequent interesting stuff):

Startup Boom a “Dangerous, High-Stakes Ponzi Scheme”: Silicon Valley Investor

And, oh, yeah, I finished this one:

The less said about it, the better. I can't recall ever before having a book make me angry.
_____________ AnthropoceneDenial

More to come...

Friday, November 9, 2018

"Epic fail?" Atul Gawande on the EHR

From an excellent New Yorker long read.

The article has an accompanying audio transcript that runs 59.07. Read/listen to all of it.
“Something’s gone terribly wrong. Doctors are among the most technology-avid people in society; computerization has simplified tasks in many industries. Yet somehow we’ve reached a point where people in the medical profession actively, viscerally, volubly hate their computers.”
Take it from the top. Atul Gawande, MD:
On a sunny afternoon in May, 2015, I joined a dozen other surgeons at a downtown Boston office building to begin sixteen hours of mandatory computer training. We sat in three rows, each of us parked behind a desktop computer. In one month, our daily routines would come to depend upon mastery of Epic, the new medical software system on the screens in front of us. The upgrade from our home-built software would cost the hospital system where we worked, Partners HealthCare, a staggering $1.6 billion, but it aimed to keep us technologically up to date.

More than ninety per cent of American hospitals have been computerized during the past decade, and more than half of Americans have their health information in the Epic system. Seventy thousand employees of Partners HealthCare—spread across twelve hospitals and hundreds of clinics in New England—were going to have to adopt the new software. I was in the first wave of implementation, along with eighteen thousand other doctors, nurses, pharmacists, lab techs, administrators, and the like.

The surgeons at the training session ranged in age from thirty to seventy, I estimated—about sixty per cent male, and one hundred per cent irritated at having to be there instead of seeing patients. Our trainer looked younger than any of us, maybe a few years out of college, with an early-Justin Bieber wave cut, a blue button-down shirt, and chinos. Gazing out at his sullen audience, he seemed unperturbed. I learned during the next few sessions that each instructor had developed his or her own way of dealing with the hostile rabble. One was encouraging and parental, another unsmiling and efficient. Justin Bieber took the driver’s-ed approach: You don’t want to be here; I don’t want to be here; let’s just make the best of it.

I did fine with the initial exercises, like looking up patients’ names and emergency contacts. When it came to viewing test results, though, things got complicated. There was a column of thirteen tabs on the left side of my screen, crowded with nearly identical terms: “chart review,” “results review,” “review flowsheet.” We hadn’t even started learning how to enter information, and the fields revealed by each tab came with their own tools and nuances.

But I wasn’t worried. I’d spent my life absorbing changes in computer technology, and I knew that if I pushed through the learning curve I’d eventually be doing some pretty cool things. In 1978, when I was an eighth grader in Ohio, I built my own one-kilobyte computer from a mail-order kit, learned to program in basic, and was soon playing the arcade game Pong on our black-and-white television set. The next year, I got a Commodore 64 from RadioShack and became the first kid in my school to turn in a computer-printed essay (and, shortly thereafter, the first to ask for an extension “because the computer ate my homework”). As my Epic training began, I expected my patience to be rewarded in the same way.

My hospital had, over the years, computerized many records and processes, but the new system would give us one platform for doing almost everything health professionals needed—recording and communicating our medical observations, sending prescriptions to a patient’s pharmacy, ordering tests and scans, viewing results, scheduling surgery, sending insurance bills. With Epic, paper lab-order slips, vital-signs charts, and hospital-ward records would disappear. We’d be greener, faster, better.

But three years later I’ve come to feel that a system that promised to increase my mastery over my work has, instead, increased my work’s mastery over me. I’m not the only one…
Fascinating. While I'm now 5 years out of the daily EHR trenches professionally, my world has been "all Epic all the time" ever since, in my roles as a patient (2015 prostate cancer dx & tx, 2018 SAVR px) and caregiver to my now-late younger daughter (2017-2018 pancreatic cancer dx & tx). All the major players here in the Bay Area -- Kaiser, Sutter, Muir, Stanford, UCSF -- are on Epic. I continue to be an acute observer of the EHR workflows I witness at every encounter, and I frequently query my clinicians about their experiences using Epic.

Nearly all I have seen during our many patient encounters across the past few years has been that of clinicians at all license levels whipping around the Epic EHR at lightning speed. Yes, they also all grouse about what they see as nuisance diversionary billing and compliance documentation, but the clinical care aspects of the EHR are about as efficient as you could hope for. That there are hundreds to thousands of clinical variables to be recorded and tracked is simply a core reality of medical care -- not the fault of the EHR.

Paper is not better.

Dr. Gawande:
“… the computer, by virtue of its brittle nature, seems to require that it come first. Brittleness is the inability of a system to cope with surprises, and, as we apply computers to situations that are ever more interconnected and layered, our systems are confounded by ever more surprises. By contrast, the systems theorist David Woods notes, human beings are designed to handle surprises. We’re resilient; we evolved to handle the shifting variety of a world where events routinely fall outside the boundaries of expectation. As a result, it’s the people inside organizations, not the machines, who must improvise in the face of unanticipated events.”
I am reminded of my prior post "Are structured data the enemy of health care quality?" Also, see my "Update on our favorite whipping boy, the EHR."
I might note that the bulk of the litany of complaints set forth in the Gawande article (and those of many others) are hardly news to those of us who have been involved in the EHR wars. I've been listening to these gripes since I came to the health IT space in 2005 with the onset of the QIO 8SOW "DOQ-IT" program.
Responding to the immediately foregoing Gawande observation inescapably leads me to, among other resources, this glorious book I recently finished.

"AI" to the rescue? Skeptics remain legion (including eminent AI pioneer Judea Pearl).
Big Data and causal inference together play a crucial role in the emerging area of personalized medicine. Here, we seek to make inferences from the past behavior of a set of individuals who are similar in as many characteristics as possible to the individual in question. Causal inference permits us to screen off the irrelevant characteristics and to recruit these individuals from diverse studies, while Big Data allows us to gather enough information about them.

It’s easy to understand why some people would see data mining as the finish rather than the first step. It promises a solution using available technology. It saves us, as well as future machines, the work of having to consider and articulate substantive assumptions about how the world operates. In some fields our knowledge may be in such an embryonic state that we have no clue how to begin drawing a model of the world. But Big Data will not solve this problem. The most important part of the answer must come from such a model, whether sketched by us or hypothesized and fine-tuned by machines…

Pearl, Judea. The Book of Why: The New Science of Cause and Effect (p. 352). Basic Books. Kindle Edition.
Now, while none of that speaks to the chronic, clinically enervating "productivity treadmill" concerns so adroitly addressed by Dr. Gawande, it is nonetheless relevant more broadly.

Regarding Dr. Gawande's "brittleness" characterization of computers, Dr. Pearl:
The goal of strong AI is to produce machines with humanlike intelligence, able to converse with and guide humans. Deep learning has instead given us machines with truly impressive abilities but no intelligence. The difference is profound and lies in the absence of a model of reality.

Just as they did thirty years ago, machine learning programs (including those with deep neural networks) operate almost entirely in an associational mode. They are driven by a stream of observations to which they attempt to fit a function, in much the same way that a statistician tries to fit a line to a collection of points. Deep neural networks have added many more layers to the complexity of the fitted function, but raw data still drives the fitting process. They continue to improve in accuracy as more data are fitted, but they do not benefit from the “super-evolutionary speedup.” If, for example, the programmers of a driverless car want it to react differently to new situations, they have to add those new reactions explicitly. The machine will not figure out for itself that a pedestrian with a bottle of whiskey in hand is likely to respond differently to a honking horn. This lack of flexibility and adaptability
[emphasis mine -BG] is inevitable in any system that works at the first level of the Ladder of Causation… [Pearl, pp. 30-31]
"Lack of flexibility and adaptability" -- i.e., "brittleness."

You'd have to study the entire Pearl book to fully appreciate that AI has quite a way to go before it significantly enables digital workflow "adaptability" borne of "causal reasoning" capacity, particularly in the complex data health IT space. Nonetheless, there remains a lot that can be accomplished below the "strong AI" level to get us closer to more digitally "frictionless" clinical workflow usability. 

BTW, apropos, see also my prior posts concerning "The Digital Doctor."

Stay tuned.

More to come...