Sunday, March 30, 2014

Encryption and differential privacy discussed on way out of NSA sinkhole

The U.S. government found itself in a very defensive position vis-à-vis data privacy in the wake of Edward Snowden's NSA disclosures. In January, Pres. Obama promised to appoint a group to look more deeply at U.S. Intelligence programs, which acted as if by fiat from 9-11 on. A recent MIT event took a look at encryption and differential privacy technology as part the review effort.

The latest on that is an Administration proposal to turn over the storage of phone records to phone companies, and to tighten the requirements for subpoenas thereof.  One doesn’t necessarily get a warm feeling on that… but some long time NSA watchers see it as a step forward.

When Obama charged John Podesta, long-time Democratic operative and now White House Counselor, to head the study group, he also said to look at big data commerce and its potential to threaten civil liberties.
The White House enlisted academics, including MIT's Computer Science and Artificial Intelligence Lab Big Data Initiative group, as part of that effort.  In March I covered a related workshop on “Big Data and Privacy: Advancing the State of the Art in Technology and Practice” and, together with colleague Ed Burns, reported this on a SearchDataManagement.com Talking Data podcast.

Both Burns and I felt the MIT conference was a bit high on the technology side (encryption and differential privacy being prominent) and bit low on the privacy side. The notion that data is like the "new gold" or the "new oil" seems overblown, until you see a room full of policy and commerce people discussing how much data is going to change the world as we know it. Whether they are right or wrong is less important than the palpable sense something akin to gold or oil ''fever'' is in the air.

Podesta had planned to attend the event, but was hampered by snow in Washington (although one might guess that, this being the weekend of the Russian Crimean Peninsula incursion, staying close to the White House was wise). He spoke with the assembled by teleconference. Below are some riffs from his published remarks. – Jack Vaughan

"…one purpose of this study is to get a more holistic view of the state of the technology and the benefits and challenges that it brings.  This Administration remains committed to an open, interoperable, secure and reliable internet – the fundamentals that have enabled innovation to flourish, drive markets and improve lives.  

"There is a lot of buzz these days about “Big Data” – a lot of marketing-speak and pitch materials for VC funding. 
"(But) the value that can be generated by the use of big data is not hypothetical.  The availability of large data sets, and the computing power to derive value from them, is creating new business models,
 "With the exponential advance of these capabilities, we must make sure that our modes of protecting privacy – whether technological, regulatory or social – also keep pace.

Related
http://cdn.ttgtmedia.com/rms/editorial/sDM-TalkingDataPodcast-March31-BigDataPrivacyWorkshop.mp3


Saturday, March 22, 2014

Through the scanner darkly, darkly; and the future of information

scanner eye by jvaughan
The digitization of everything is an elixir for some people. It spawns visions. If we could only open up all the data…how about taking the college facebook and putting it on line …. why not street-level and satellite-level photos of every home in the U.S. of A. Ok! Build and sell a picture database of all the license plates on all the cars and trucks on the road? Gee, I don't know. The Department of Fatherland Security recently moved to create a national license-plate recognition database to garner data from commercial and law enforcement tag readers. Then, with NSA skulduggery still a little too current, they canceled it a' sudden. Note that commercial tag reader systems remain out there. DRN or Digital Recognition Network provides "data that puts your company in the driver's seat" helping you repo your assets (e.g., cars) and reduce asset charge-offs. Together with Vigilant Solutions of Livermore, Calif., the company is fighting a Utah law that banned the private, commercial use of the license plate scanning technology. DRN was the only speaker at a hearing on the topic at the Mass State House earlier this month. They see it as their first amendment right to make money taking pictures of stuff. When you think of all the big data uses of license plates beyond immigration, repossession, well its boggling. Probably their more big data apps to come, that we cant even think of, but why not collect the data for that big day in the future? The undercurrent is, if I don’t want the NSA or DFS to do it, why would I want some Starbuck's guzzling nerdster to? Re-jiggering of status quo is what massive levels of data can do. Google has met a few people who don't want pictures of their houses in Google's database and, apparently, will remove them if you ask. I don't think First Amendment rights to take pictures are a foundation for massively scaled reproduction, and would not my license plate in some software company data services offering. In "Who owns the Future," Jaron Lanier lays down some framework for a more credible understanding of where we want to go with data and privacy. By asking questions of the future he takes a sharper picture of the present: "... as technology advances i this century, our present intuition about the nature of information will be remembered as narrow and shortsighted." - Jack Vaughan

Related 
http://www.foxnews.com/politics/2014/02/19/dhs-plan-for-national-license-plate-tracking-system-raises-privacy-concerns/
http://www.googletutor.com/asking-google-to-remove-your-home-from-maps-street-view/
http://betaboston.com/news/2014/03/05/a-vast-hidden-surveillance-network-runs-across-america-powered-by-the-repo-industry/
http://consumerist.com/2014/03/05/the-repo-man-might-be-scanning-your-cars-license-plate-and-location-selling-the-data/
http://www.drndata.com/Content/Docs/DRN%20Vigilant%20Utah%20Press%20Release.pdf
http://www.drndata.com/
http://vigilantsolutions.com/




Saturday, March 15, 2014

I have heard all about Grantland



I'm reading interesting book called Talk Nerdy to Me. This is by the ultra-hip Grantland (as in Grantland Rice) crew whose totally cool cat's sportswriting on the web packaged here bears the blistering subtitle of "Talk Nerdy to Me: Grantland's Guide to the Advanced Analytics Revolution" (sold out says the site today) and it is a more interesting take on big data analytics then many other tomes that you maybe have anted up for. Let's start with "Belichick's Fourth and Reckless" by contributor Bill Simmons. The story centers at times maniacally but on Patriots coach Bill Belichick's famously strange call on fourth-and-2 on November 15, 2009 against Peyton Manning and the Baltimore Colts. It trumps many other coaching failures in Boston sports  fabled history of failures, he writes. It was such a strange call – the Pats were on their own 28, with less than 2 minutes to play and a lead - that people began to look at the statistics trying to see what was in the coaches – the great, mind you - mind. Simmons goes over some of the stats and pretty well proves how at times statistics can lie, or at least outsmart the lazy intellectual (you know the type that works for media!)Bellichick's crazy gambit had backing in stats. "Bellichick did play the percentages if you took those percantages at face value." But Simmons points out for example, that statistics (that going for fourth down had an 80.5 % chance of succeeding) don’t account for the obvious confused funk that had descended on the Pat's in that final quarter.  That there is a big difference between fourth-and-2 on a Sunday in September against a lazed Falcons outfit than there is in November against Petyon Bloody Manning and the Colts. Stop and grok on this:

"I know it's fun to think stats can settle everything, but they can't and they don’t."

If you are playing the statistics card, which one do you choose? Writes Simmons. There are all sorts of statistics to count, but which are the ones to count on? Pulling out all the stop here I am going to recall Mark Twain, or maybe Vin Scully, plenty of argue over who said it:

Statistics are used much like a drunk uses a lamppost: for support, not illumination.

Beware, you would be masters of the big data universe! I said that. - Jack Vaughan

Saturday, March 1, 2014

Duck duck goose

Today's clamor around big data will one day subside. Like the love affair in Cole Porter's Just One of Those Things, it is ''too hot not too cool down''.  It is a sort of process;  vendors and media builds things up and then break things down again. Take as example a recent New York Times story entitled "Big (Bad) Data." The item revolves around the case of A&E's Duck Dynasty star Phil Robertson. His antigay comments in a magazine article went viral on Twitter, and A&E execs, as if in the thrall of big data analytics, suspended him from the show. Then, the Twitter sentiments rebounded, big data was recalibrated, and Robertson was back in. The Times' story suggests the first response was wrong, the second right. But time may prove otherwise. This episode in review is hardly an indictment, although that is how the writer or his editors would have it.  The advent of big data does not obviate the need for exes to have full liberal educations with philosophy, ethology, ethics and economics studies under their belts.  The execs of A&E give vent to the old saw: If you don't know where you're going, any road will take you there.