Stanford Study Will Inspire Any IT Pro Intrigued By Machine Learning - InformationWeek
IoT
IoT
Data Management // Big Data Analytics
News
8/23/2016
10:06 AM
Connect Directly
Google+
LinkedIn
Twitter
RSS
E-Mail
50%
50%

Stanford Study Will Inspire Any IT Pro Intrigued By Machine Learning

A study by Stanford researchers finds computers can predict lung cancer patient outcomes better than pathologists. While the study is specific to the medical profession, it illustrates the promise and potential of machine learning for IT professionals in any industry.

13 Ways Machine Learning Can Steer You Wrong
13 Ways Machine Learning Can Steer You Wrong
(Click image for larger view and slideshow.)

For IT organizations, machine learning is looking like an essential capability in the decade ahead. For the past few months, Google CEO Sundar Pichai has been extolling the value of AI and machine learning to his company. Gartner has added machine learning to its 2016 Hype Cycle, putting it at the peak of inflated expectations.

The Hype Cycle, said Gartner research director Mike J. Walker in a statement, lists technologies that show "promise in delivering a high degree of competitive advantage over the next five to 10 years."

Now, researchers at the Stanford University School of Medicine have demonstrated trained computers can outperform doctors when evaluating the slides of lung cancer patients, a finding which underscores the value of machine learning for data analysis tasks involving image recognition.

The Stanford findings underscore the promise of machine learning in enterprises of all stripes.

In a paper published in Nature Communications last week, a group of scientists led by Kun-Hsing Yu described how software-based machine learning is "superior to the current practice utilized by pathologists who assess the images in terms of tumour grade and stage."

Traditionally, pathologists have evaluated patient tissue on glass slides through a microscope to determine the grade or severity of cancerous tumors and the stage to which the disease has progressed. But Michael Snyder, professor and chair of the department of genetics at Stanford, told the Stanford Medicine news service such assessments tend to vary among pathologists, who agree only about 60% of the time.

Computer-driven analysis of medical slides has been a challenge due to the large amount of information contained in whole-slide pathology. "The huge dimension of the original images made it extremely difficult to manipulate, and informatics workflows requiring manual tumour tissue segmentation were not feasible for millions of image tiles," the authors explain in their paper.

(Image: DarkoStojanovic via Pixabay)

(Image: DarkoStojanovic via Pixabay)

The Stanford researchers managed to create an informatics workflow that can be automated and scaled to address the analysis of data-dense imagery. The relevant details of tissue imagery on slides can be difficult to identify by manual inspection, the authors observe, but computers can be trained to identify salient features. The authors believe their approach can quickly and objectively assess the survival odds of lung cancer patients and lead to better therapeutic decision-making.

The researchers based their work on 2,186 images from the Cancer Genome Atlas, a national database of cancer data that includes information about the grade and stage assigned to the cancer imagery. They used those images to train their software to identify 9,879 distinct image characteristics, more than any person would weigh.

In an email, Snyder said the researchers' system performed 15% better on prognosis than pathologists.

[How is Intel approaching AI? Read Intel Buys Deep Learning Startup Nervana to Bolster AI Business.]

The diagnostic capability of computers goes back decades, but it hasn't always been as capable. In 1998, an expert system outperformed human physicians when all parties had the same information. It made accurate diagnoses 65% of the time, compared to 54% among participating doctors, according to a study titled "Comparison Between Diagnoses of Human Experts and a Neurotologic Expert System."

However, when the physicians had access to the complete medical records of patients, they surpassed the expert system, making accurate diagnoses 69% of the time.

Thomas Claburn has been writing about business and technology since 1996, for publications such as New Architect, PC Computing, InformationWeek, Salon, Wired, and Ziff Davis Smart Business. Before that, he worked in film and television, having earned a not particularly useful ... View Full Bio

We welcome your comments on this topic on our social media channels, or [contact us directly] with questions about the site.
Comment  | 
Print  | 
More Insights
Comments
Newest First  |  Oldest First  |  Threaded View
jries921
50%
50%
jries921,
User Rank: Ninja
8/31/2016 | 12:27:24 PM
Re: Data overload?
Got it. The sad thing is that people in jobs and organizations that stress obedience and permit little or no discretion are the ones most likely to be replaced with machines; though that may be management's own punishment, since humans have been known to quietly improvise when official procedures are obviously insane given the circumstances, or when there are no rules applicable to the case. "The Peter Principle" discusses "computerized incompetence" at length, and I think the concept applies here.
SaneIT
100%
0%
SaneIT,
User Rank: Ninja
8/31/2016 | 8:29:57 AM
Re: Data overload?
I think we're on the same page here.  My point was that when it comes to task based jobs humans tend to act like machines that are programmed to do the tasks in a specific manner.  I do a lot of working with processes in various departments that reflect this kind of programming.  "Why are you doing it this way?" is usually met with "That's the way we've always done it."  They are programmed not that it's the right way or the best way but it's THE way.  Many people don't feel like they can change processes like this so I get one of two reactions when they see me coming, either fear that I'm about to change things and they will be uncomfortable with the change or excited because they know there has to be a better way and they'll get to voice that and help move away from something they have felt is broken.
jries921
50%
50%
jries921,
User Rank: Ninja
8/30/2016 | 11:37:04 AM
Re: Data overload?
I didn't say humans couldn't be programmed, but unlike computers, we have wills of our own (though sometimes suppressed by abuse) and thoughts of our own (sometimes to the annoyance of those in authority over us) and are thus in a position to resist our programming, or even to reprogram ourselves (at least in part).  Computers can only follow orders.
SaneIT
50%
50%
SaneIT,
User Rank: Ninja
8/30/2016 | 8:21:35 AM
Re: Data overload?
Since grammar counts, I think the right way for me to have typed that would have been "humans are programmed, just like software."  Humans are indeed programmed.  We can choose what we want to learn and we can adjust how we react and how we use the information we have but watch an assembly line worker some time.  They are programmed to put part B onto part A and repeat for an 8 hour shift.  Someone else holds the knowledge of what parts A and B do and why they are assembled in this manner. When a job such as viewing hundreds of images a day to identify anomalies is performed by humans they fall into a routine the same way software routines work.  It becomes automatic and if they are having an off day and can't process what they are seeing as quickly as they did the day before then errors pile up.  I'm not referring to creative tasks or hobbies but looking at the task based jobs that many humans have settled into.

 
jries921
50%
50%
jries921,
User Rank: Ninja
8/29/2016 | 9:38:05 AM
Re: Data overload?
No, humans are not programmed in the same manner as computers.  That's really the point.  In the end, computers are nothing more than general purpose programmable calculators, completely unable to comprehend the world around them and able to do only what they are told (assuming they are correctly instructed, which is frequently not the case).  They have zero imagination, zero creativity, and only the judgment programmed into them by humans.  They have zero ability to think for themselves.  They are tools, nothing more and nothing less.  We may someday have something like HAL or Data (which would open up a whole new can or worms); but right now, we're not even close.
SaneIT
50%
50%
SaneIT,
User Rank: Ninja
8/29/2016 | 8:30:41 AM
Re: Data overload?
That is a very common issue, humans are programmed just like software.  They go through years of training to spot specific things.  We go through life watching things happen and predicting outcomes based on personal history.  We tend to get a little data blind when we think we know what we're looking at or looking for.  An AI that is designed to look at thousands of factors vs a human that can handle about a dozen at a time will increase accuracy over time.  I look forward to faster more accurate testing in cases like this.
Whoopty
100%
0%
Whoopty,
User Rank: Ninja
8/25/2016 | 7:33:51 AM
Re: Data overload?
I know what you mean, but I'm less concerned about AI of the future than I once was. While I think a lot of people like to paint it as one filled with AI making decisions and doing jobs for us, in reality I think they're more likely to augment us, so they will do the busy work, which lets doctors and the like perform the more complicated, human parts of their profession.

If we all stopped having to do paperwork it would save a lot of messing around. 
jries921
50%
50%
jries921,
User Rank: Ninja
8/24/2016 | 12:43:15 PM
Data overload?
It strikes me that the more complex the data, the harder it becomes for humans to "eyeball it".  That said, I still don't think I want an AI to perform my next medical exam.
Register for InformationWeek Newsletters
Video
Current Issue
The Next Generation of IT Support
The workforce is changing as businesses become global and technology erodes geographical and physical barriers.IT organizations are critical to enabling this transition and can utilize next-generation tools and strategies to provide world-class support regardless of location, platform or device
White Papers
Slideshows
Twitter Feed
Sponsored Live Streaming Video
Everything You've Been Told About Mobility Is Wrong
Attend this video symposium with Sean Wisdom, Global Director of Mobility Solutions, and learn about how you can harness powerful new products to mobilize your business potential.
Flash Poll