(Credit:
AFIT)
Here's another reason to get off that antisocial kick and get with the networking.
The Air Force is developing a data-mining technology meant to root out disaffected insiders based on their e-mail activity--or lack thereof, according to an article in this month's International Journal of Security and Networks.
The technology, based on something called Probabilistic Latent Semantic Indexing (PDF), scours an organization's e-mail traffic and constructs a graph of social network interactions illustrating employee activity. If a worker suddenly stops socializing online, abruptly shifts alliances within the organization, or starts developing an unhealthy interest in "sensitive topics," the system detects it and alerts investigators.
Most corporate security efforts focus on electronic threats from the outside, even through insiders with access to sensitive information can pose a greater threat to an organization, according to researchers at the Air Force Institute of Technology at Wright Patterson Air Force Base in Ohio. Alienated individuals who display a secret interest in suspicious topics but never let on by communicating with others are the most likely to be an insider threat, the researchers say. The program could prevent security breaches, sabotage, and even terrorist activity at multinational corporations and military organizations alike, according to the article.
And don't think that just because you're the boss you're off the hook. The team tested Enron's e-mail archive and uncovered several individuals who represented potential insider threats. Granted, none of them were the bosses who had done all the damage, but the researchers were confident that with full access and by turning a "domain on its ear" the software would ferret out potential malefactors and whistleblowers alike.
Growth in the IP lawsuit sector of the economy and the resulting patent-reform movement can leave you thirsting for reliable data on which to base policy judgments and cocktail-party conversation. Initially, I thought the IEEE Spectrum's annual patent stats roundup, Keeping Score in the IP Game would handily rehydrate my chit-chat cortex. After all, the data contains glittering--although frequently unsurprising--infonuggets like:
- Microsoft ranks fifth for total patents awarded in the Computer Systems and Software category (1,469), but first in IEEE's computed measure of the power conferred by those patents. The other four top players in that category are IBM (3,651), HP (5,2115), Toshiba (1,987), and Fujitsu (1,674).
- Hitachi (3,198), Matsushita (2,507), and Sony (1,970) lead the pack for Electronics.
- Samsung (2,474) tops the chart in semiconductor manufacturing, but Intel's portfolio (1,961) has the most power.
- For telecom, Siemens (1,514) has the most patents, but AT&T's (763) pack more punch.
I'm also impressed by the magazine's attempt to quantify the impact of patent portfolios with its Pipeline metric.
That said, however, a couple of aspects of the story bothered me. The tagline reads: "Quantity and concentration of patents strengthen the case for legislative reform." Hmmm. I'll ignore the fact that the story never even tries to draw a causal link between its numbers and patent abuse. The data does show that big companies have a lot of patents. The data also show that some relatively small companies, like Digimarc with its measly 77 patents, can wield considerable power--it ranks fifth in the computer systems and software category, above Sun Microsystems, Fujitsu, NEC and Oracle.
The article does state that "the patents in any given field still go to a few top companies, that there is little change from year to year among the dominant firms, and that big gaps yawn between the leaders and the runners up." But the author fails to point out that by the Pipeline metric, there are far fewer such gaps than between the raw numbers.
I could go on, but I won't, because regardless of my nitpicks, it's an interesting read.
Facing criticism over its privacy practices, the U.S. Department of Homeland Security has opted to abandon a massive new data-mining project intended to flag patterns and pinpoint relationships that indicate terrorist threats.
According to an Associated Press report, the department "quietly" suspended a program known as ADVISE (Analysis, Dissemination, Visualization, Insight, and Semantic Enhancement) after federal government auditors issued a report this spring dogging the agency for failing to assess the privacy risks associated with the scheme. ADVISE was one of a dozen data-mining projects under way at Homeland Security, according to the AP report.
Homeland Security's own inspector general and Privacy Office went on to publish their own reports this summer, and they were no less damning. They concluded the agency had used personally identifiable data about real people--gleaned, for instance, from the no-fly list and from a database of foreign students--in test runs of some of the ADVISE systems without first doing a full assessment of the privacy impact of those activities. The Privacy Office report (PDF) suggested using "synthetic" data in the future instead.
The department made a similar mistake with its so-called Secure Flight program, when it failed to report exactly how and why it collected commercial personal data on nearly 250,000 airline passengers. Under pressure from privacy advocates, the feds recently proposed overhauling that program, which is designed to screen travelers against terrorist watch lists.
Just because Homeland Security is scrapping development of ADVISE doesn't mean it plans to abandon the information-sifting tactic. In fact, it's apparently already contemplating a replacement. A Homeland Security spokesman told the AP reporter that the department's Science and Technology Directorate, which oversaw the project, "determined that new commercial products now offer similar functionality while costing significantly less to maintain than ADVISE."
Since 2003, the department has spent approximately $40 million to develop ADVISE, which it planned to roll out in six phases. Homeland Security assured members of Congress last year that the project was never used in real operations and was still in testing mode. But according to the department's inspector general report, "on at least one occasion, the data was used to produce classified intelligence information."
Information is the new currency. When it comes to social-networking sites and many other online enterprises, your attention is the product that is being sold. So it is not surprising that data mining, particularly efforts to link your online behavior to specific opportunities to market to you, is an exploding trend.
Data mining in itself is not inherently good or bad, but it raises many social issues whose implications we all need to understand and include in our ongoing dialogue. Data mining has benefits, including an opportunity to create a customized online experience that truly serves you better. Misuses can lead to serious breaches of privacy. I encountered several stories on data mining Tuesday that caught my attention.
... Read more- prev
- 1
- next
