Saturday, April 22, 2017

Shannon, information and noise

Dr Disruption Signal Sign
This fits with things coming up again. I wrote it for ITWorld when Claude Shannon died. Also has re-run on MoonTraveller.

OBITUARY-APPRECIATION - We live in an age highly influenced by information technology. For many people, it has become the basis for a life's work. For a few, at least, it has meant great fortunes.
Most of the great technologists who set the stage for this era -- for example, Norbert Weiner, Vannevar Bush, Alan Turing, and John von Neumann -- are long dead. But Claude Shannon, the great theorist who formed the most basic tenets of the information age, survived until last weekend. He died at 84 last Saturday in Medford, Mass., after a long fight with Alzheimer's disease.

Shannon's work, like his passing, may not be widely noted among many who have followed him in the information, technology, and e-commerce industries. But there is little question that he is the chief progenitor of information theory and modern digital communications. Shannon's mathematical thinking and writing laid the groundwork for most of today's information technology industry. He is the man who discovered 1's and 0's in electronic communication.

Shannon was born in Petoskey, Mich., and grew up in Gaylord, Mich. He worked as a messenger for Western Union while in Gaylord High School, and attended college at MIT, where he was a member of Tau Beta Pi.

Although the algebra of digital binary bits was first uncovered by mathematician George Boole in the mid-19th century, it was Shannon who saw the value of applying that form of logic to electronic communications. As a student of Vannevar Bush's at MIT in the 1930s, he worked on the differential analyzer, perhaps the greatest mechanical (analog) calculator. His paper, "A Symbolic Analysis of Relay and Switching Circuits," which led to a long association with Bell Laboratories, laid out Shannon's theories on the relationship of symbolic logic and relay circuits.

While at Bell Labs, Shannon wrote the landmark "The Mathematical Theory of Communication." The information content of a message, he theorized, consists simply of the number of 1's and 0's it takes to transmit it. In a real sense, Shannon conceived of the "bit" that is now so widely used to represent data.

Later, he became a professor at MIT. His students included Marvin Minsky and others who became notable in the field of artificial intelligence. While Shannon's thinking could captivate academicians, it was equally appealing to practical engineers.

Shannon's work led to many inventions used by both technology developers and end users. His theories can truly be described as pervasive today.

When I was young, Shannon's work was a tough nut to crack, but it certainly was intriguing. As a high school boy, I was interested in the future -- maybe more so than now, when I live and breathe and work in what that future became. Grappling with Shannon's basic information theories was part of my education about the future.

Growing up in a Wisconsin city across the lake from Shannon's birthplace, I tried to plow through the town library as best I could. I wanted to learn about computers, automation, and the combination of the two that was known in those days (the 1960s) as cybermation. I discovered for myself -- by chance, really -- that the fundamental elements of those ideas were Shannon's inventions.

For the better part of Shannon's life, analog communication ruled. Of course, his greatest achievement was visualizing digital communication. Much of his greatest work revolved around defining information in relation to "noise," the latter phenomenon being quite familiar to anyone who often tried desperately to home in on radio signals before digital communication filters came into being. I came to appreciate that aspect of Shannon's work later on when, as a journalist, I had the opportunity to learn and write about digital signal processing.

Then I found out that Shannon had laid the groundwork for modern error correction coding, an essential element of things like hard disk drive design and digital audio streaming, and probably many things yet to come.

Day and night, data, messages, music, and more swirls around us -- all made possible to some extent by the idea of communicating electronically in 1's and 0's. It is something to think that a Western Union messenger could have conceived of this new world. - Jack Vaughan, originally published in ITWorld.com

Wednesday, April 12, 2017

The road to machine learning

In some spare time last year I worked on reworking a story I'd written for TechTarget on the topic of machine learning and related data infrastructure. I spoke with a few users in the story, so the reworking focuses on their end-use applications, while also covering vendors and technology employed. I did this as a Sway multimedia presentation. I faltered in final completion, in that an audio track was the proverbial bridge too far. So I call it a failure (experience tells us to jump from projects that are "projects of one" but thought it wa worth posting here, for experimental purposes. To read the story it is based on please go to "Machine Learning tools pose educational challenges for users." - Jack Vaughan


Monday, April 10, 2017

But,it worked on my machine . . .

Courtesy: A B. Normal
Functional MRI that shows human brain in action is growing in use within academic and research communities intent on uncovering the inner workings of the brain, says Russell Poldrack of Dept of Psychology, Stanford U. tells a lecture audience at Computer Information Science and Engineering series at NSF. But there are issues. The approach is costly, and it is not clear how reproducible big data analysis of the brain action is. Reproducibility of experiments using the scientific method is very important, but when R programs run on different machine configurations, different analytical results may emerge. Poldrack and others are working with software containers and virtual machines and registries to address this issue of computational reproducibility. - Jack Vaughan

RELATED


Sunday, April 2, 2017

The evening of a playing field?


Hand Of The Buddha
As the Republican congress capitalizes on the friendly Republican White House, it is overturning a lot of rocks, and passing legislation friendly to one or another among various corporate interests - such as the cable ISP and business.

This week, the House voted to upend Obama era FTC regulations that forbade ISPs from selling individuals' browser activity data. The group was to be put somewhat on the outside on the action of what is called big data - required to get formal permission from customers in order to sell browser histories to adtech markets - the ones dominated by Facebook and Google.

What's the difference between an ISP and a Google? Google provides a free service as part of a (admittedly murky) quid pro quo. You get free browser and free search - and you tacitly give them the right to use you as a datum. With the ISP, you pay them - and not with a lot of choice either, as they are more often than not a monopoly in your neighborhood.

The stakes ISPs stuck in the Internet are deep. It can hardly be said this legislation is the evening of a playing field. Several of the companies have pledged to ask for permissions of customers before selling their (anonymized) browser history. It's likely best alled a feel good gesture on the part of the people pulling the puppet strings of government these days. You know, "Monday, geld the EPA." "Tuesday, affirm the right to kill sleeping bears in National Parks." "Wednesday, toss the ISPs a big data bone." - Jack Vaughan

RELATED
http://www.zdnet.com/article/isps-were-not-going-to-sell-your-web-browsing-data/
http://continuations.com/post/158773876945/government-just-gave-your-isp-even-more-power-you
https://www.forbes.com/sites/thomasbrewster/2017/03/30/fcc-privacy-rules-how-isps-will-actually-sell-your-data
https://www.wired.com/2017/03/big-cables-case-selling-data-doesnt-hold/

SOURCES
Cards Against Humanity creator Max Temkin
Matthew Hogan, CEO at DataCoup
ALBERT WENGER, a partner at Union Square Ventures, author "World After Capital”
Dallas Harris, a policy fellow with consumer advocacy group Public Knowledge.