The FBI Uses Its Hacking Powers for Good
March 4, 2016
In a victory for basic human decency, Engadget informs us, the “FBI Hacked the Dark Web to Bust 1,500 Pedophiles.” Citing an article at Vice Motherboard, writer Jessica Conditt describes how the feds identified their suspects through a site called (brace yourself) “Playpen,” which was launched in August 2014. We learn:
Motherboard broke down the FBI’s hacking process as follows: The bureau seized the server running Playpen in February 2015, but didn’t shut it down immediately. Instead, the FBI took “unprecedented” measures and ran the site via its own servers from February 20th to March 4th, at the same time deploying a hacking tool known internally as a network investigative technique. The NIT identified at least 1,300 IP addresses belonging to visitors of the site.
“Basically, if you visited the homepage and started to sign up for a membership, or started to log in, the warrant authorized deployment of the NIT,” a public defender for one of the accused told Motherboard. He said he expected at least 1,500 court cases to stem from this one investigation, and called the operation an “extraordinary expansion of government surveillance and its use of illegal search methods on a massive scale,” Motherboard reported.
Check out this article at Wired to learn more about the “network investigative technique” (NIT). This is more evidence that, if motivated, the FBI is perfectly capable of leveraging the Dark Web to its advantage. Good to know.
Cynthia Murrell, March 4, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
If You See Something, Say Something Adopts New Cybersecurity Meaning
March 4, 2016
A post-9/11 campaign for increasing security awareness will inform a similar public service announcement campaign to bring cybersecurity top of mind. See something suspicious online? Homeland Security wants to know about it published by NextGov reports on this 2016 Department of Homeland Security initiative. The decision to launch this campaign comes from an IDC recommendation; the US lacks a culture of cybersecurity concern, unlike Israel, according to the article. While $1 million is allotted for this campaign, the article describes bigger future plans,
“Last week, the Obama administration rolled out a new Cybersecurity National Action Plan, which establishes a new public commission on cybersecurity and proposes billions in new funding to upgrade hard-to-secure legacy IT systems in use at federal agencies, among several other steps.”
This year’s cybersecurity public and private sector awareness campaign was modeled after the “If You See Something, Say Something” campaign rolled out after September 11. However, this is not Homeland Security’s first attempt at educating the public about cybersecurity. The department has sponsored October as National Cybersecurity Awareness Month since 2004. As the article mentions, previous educational efforts have not appeared to influence culture. It would be interesting to know what metrics they are using to make that claim.
Megan Feil, March 4, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Google Is Taxing
March 3, 2016
I read “Google Lowered Taxes by $2.4 Billion Using European Subsidiaries.” Interesting stuff now that tax season in the US. My reaction to the headline is that Google is probably conforming to the applicable laws.
The write up points out:
Google saved $2.4 billion in worldwide taxes in 2014 by shifting 10.7 billion euros ($12 billion) in international revenues to a Bermuda shell company, regulatory filings show.
I like Bermuda. The write up referenced the method which I have noted before:
Google’s Dutch subsidiary is the heart of tax structures known as a “Double Irish” and a “Dutch Sandwich” because it involves moving money from one Google subsidiary in Ireland to a Google subsidiary in the Netherlands before moving it out again to a different Irish subsidiary, physically based in Bermuda, where there is no corporate income tax. This movement of cash enables Google parent Alphabet to keep the effective tax rate on its international income in the single digits.
The more coverage the Google sandwiches get, the more regulators will think about the Alphabet Google thing.
I am not sure that publicity with regard to tax methods is particularly helpful.
Stephen E Arnold, March 3, 2016
Why Google Compare Was Terminated with Extreme Prejudice
March 3, 2016
I read “Google Will generate More Revenue from a Fourth Ad Than from Compare.” The answer is in the headline. The Alphabet Google thing is worrying about lawsuits, costs and revenue. Focus is often a good thing for a giant company with Loon balloons, researchers working on solving death, and mechanics building robots unsuitable for use in a pre-school.
The write up reports:
Since Google now shows four ads on “highly commercial queries” instead of three, the search engine clearly believes that the fourth listing will generate more revenue from one of the price comparison websites than it has done from searchers using its own comparison tools. Morling [a Google watcher it seems] told me that “developing a financial comparison service along with continual innovation takes time, resources and expertise”. He added: “Consumers are unlikely to lose out because those sites dedicated to financial comparison are better placed to provide added value such as supporting information and richer functionality.”
My conclusion is that if a Google service is a liability, that service may be given the Orkut treatment.
Stephen E Arnold, March 3, 2016
IBM Watson, Google DeepMind Is Slicing into Health Care
March 3, 2016
Gentle reader, you may have seen out write ups about IBM Watson and its work to cure cancer and develop innovative recipes for barbeque sauce with tamarind.
I read “Smart Care: How Google DeeepMind Is Working with NHS Hospitals.” The write up points out:
A smartphone app piloted by the NHS could improve communication between hospital staff and help patients get vital care faster.
Yikes, Watson, a phone. Come here I need you will echo in the corridors of these paragons of efficiency throughout Britain.
I learned:
Their research, published in the journal Surgery, showed that half of hospital patients do not get the care they need fast enough, usually because of poor communication, particularly when one team of doctors or nurses hands over to another. In early pilots at St Mary’s Hospital, part of Imperial College Healthcare NHS Trust, where Darzi [former health minister in the Blair government and director of the Institute of Global Health Innovation at Imperial College London] is a consultant surgeon, they found medical staff responded 37% faster when alerted by the Hark app than when they used pagers.
Will an app work cooperatively with IBM Watson? Will DeepMind, the app, keep IBM Watson in the lounge area?
Painful questions for an app to answer or notify in this case of technological innovation.
Stephen E Arnold, March 3, 2016
Artificial Intelligence Competition Reveals Need for More Learning
March 3, 2016
The capabilities of robots are growing but, on the whole, have not surpassed a middle school education quite yet. The article Why AI can still hardly pass an eighth grade science test from Motherboard shares insights into the current state of artificial intelligence as revealed in a recent artificial intelligence competition. Chaim Linhart, a researcher from an Israel startup, TaKaDu, received the first place prize of $50,000. However, the winner only scored a 59.3 percent on this series of tasks tougher than the conventionally used Turing Test. The article describes how the winners utilized machine learning models,
“Tafjord explained that all three top teams relied on search-style machine learning models: they essentially found ways to search massive test corpora for the answers. Popular text sources included dumps of Wikipedia, open-source textbooks, and online flashcards intended for studying purposes. These models have anywhere between 50 to 1,000 different “features” to help solve the problem—a simple feature could look at something like how often a question and answer appear together in the text corpus, or how close words from the question and answer appear.”
The second and third place winners scored just around one percent behind Linhart’s robot. This may suggest a competitive market when the time comes. Or, perhaps, as the article suggests, nothing very groundbreaking has been developed quite yet. Will search-based machine learning models continue to be expanded and built upon or will another paradigm be necessary for AI to get grade A?
Megan Feil, March 3, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Delve Is No Jarvis
March 3, 2016
A podcast at SearchContentManagement, “Is Microsoft Delve Iron Man’s Edwin Jarvis? No Way,” examines the ways Delve has yet to live up to its hype. Microsoft extolled the product when it was released as part of the Office 365 suite last year. As any developer can tell you, though, it is far easier to market than deliver polished software. Editor Lauren Horwitz explains:
“While it was designed to be a business intelligence (BI), enterprise search and collaboration tool wrapped into one, it has yet to make good on that vision. Delve was intended to be able to search users’ documents, email messages, meetings and more, then serve up relevant content and messages to them based on their content and activities. At one level, Delve has failed because it hasn’t been as comprehensive a search tool as it was billed. At another level, users have significant concerns about their privacy, given the scope of documents and activities Delve is designed to scour. As BI and SharePoint expert Scott Robinson notes in this podcast, Delve was intended to be much like Edwin Jarvis, butler and human search tool for Iron Man’s Tony Stark. But Delve ain’t no Jarvis, Robinson said.”
So, Delve was intended to learn enough about a user to offer them just what they need when they need it, but the tool did not tap deeply enough into the user’s files to effectively anticipate their needs. On top of that, it’s process is so opaque that most users don’t appreciate what it is doing, Robinson indicated. For more on Delve’s underwhelming debut, check out the ten-minute podcast.
Cynthia Murrell, March 3, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
Drone2Map: Smart Software
March 2, 2016
If you are interested in mapping and geospatial analyses, you will want to read “ESRI Introduces Drone2Map to Process Aerial Images.” The write up reports:
Drone2Map incorporates Pix4D’s powerful image-processing engine to analyze images taken from drones and convert them into a variety of 2-D and 3-D maps.
What’s interesting to me is that the software is available for public download. You will need to know about ArcGIS and some other tools.
You can find the software at this link. You will have to jump through a couple of hoops. Don’t forget to register your drone.
Stephen E Arnold, March 2, 2016
Yahoo Has AI Advantage Maybe?
March 2, 2016
I read “Don’t Laugh: Yahoo’s Open Source AI Has a Secret Weapon.” Sorry, I did laugh. I find the Yahooligans’ periodic “we’re really good at technology” messages amusing. More interesting is the willingness of with it magazines to cover these breakthroughs.
I learned:
Yahoo published the source code to its CaffeOnSpark AI engine so that anyone from academic researchers to big corporations can use or modify it.
Good. Open source software is useful, very useful.
I noted this passage:
Yahoo, for example, uses it to improve search results on Flickr by determining the contents of different photos. Instead of relying on the descriptions and keywords entered by the people who upload photos to the site, Yahoo teaches its computers to recognize certain characteristics of a photo, such as specific colors or even objects and animals.
Interesting, but other outfits do image recognition reasonably well. Check out Yandex’s image search or look at the wonky similar images feature that makes it oh, so easy for me to lose my train of thought when looking for examples of Palantir’s interface via Google’s image search service.
I learned:
CaffeOnSpark, as the name suggests, combines two existing technologies: the popular deep learning framework Caffe and the up-and-coming data-crunching system Spark that can run on top of the even more popular big data platform Hadoop. What Yahoo did was simply create a way to run Caffee atop Spark clusters. It can be run either on Spark alone or atop Hadoop. Besides making it easy for AI developers to use familiar tools and avoid moving data around… CaffeOnSpark also makes it relatively easy to distribute deep learning processes across multiple servers, something that the open source version of Google’s TensorFlow can’t yet do.
The challenge for Yahoo is to deal with its here and now problems. The outfit is for sale and many of the researchers of yesteryear have ridden off into the sunrise to find companies able to generate revenue from innovations.
When you are for sale, publicity is a definite plus. By the way, companies with technology to distribute deep learning across multiple servers are chugging along and closing some deals based on their know how. When does open source become a source of revenue and when is it a PR play?
Stephen E Arnold, March 2, 2016
Stolen Online Account Info Now More Valuable than Stolen Credit Card Details
March 2, 2016
You should be aware that criminals are now less interested in your credit cards and other “personally identifiable information” and more keen on exploiting your online accounts. As security firm Tripwire informs us in their State of Security blog, “Stolen Uber, PayPal Accounts More Coveted than Credit Cards on the Dark Web.” Writer Maritza Santillan explains:
“The price of these stolen identifiers on the underground marketplace, or ‘the Dark Web,’ shows the value of credit cards has declined in the last year, according to security firm Trend Micro. Last week, stolen Uber account information could be found on underground marketplaces for an average of $3.78 per account, while personally identifiable information, such as Social Security Numbers or dates of birth, ranged from $1 to $3.30 on average – down from $4 per record in 2014, reported CNBC. Furthermore, PayPal accounts – with a guaranteed balance of $500 –were found to have an average selling price of $6.43. Facebook logins sold for an average of $3.02, while Netflix credentials sold for about 76 cents. By contrast, U.S.-issued credit card information, which is sold in bundles, was listed for no more than 22 cents each, said CNBC.”
The article goes on to describe a few ways criminals can leverage these accounts, like booking Uber “ghost rides,” or assembling personal details for a very thorough identity theft. Pros say the trend means service providers to pay closer attention to usage patterns, and to beef up their authentication processes. Specifically, says Forrester’s Andras Cser, it is time to move beyond passwords; instead, he proposes, companies should look for changes in biometric data, like phone position and finger pressure, which would be communicated back to them by our mobile devices. So we’re about to be even more closely monitored by the companies we give our money to. All for our own good, of course.
Cynthia Murrell, March 2, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
-
- Subscribe to Beyond Search
Feature archive
News archive
-
