Automatic Text Categorization Goes Mainstream
January 10, 2019
Blogger and scaling consultant Abe Winter declares, “Automatic Categorization of Text Is a Core Tool Now.” Noting that, as of last year, companies are using automatic text categorization regularly, Winter clarifies what he is, and is not, referring to here:
“I’m talking about taking a database with short freeform text fields and automatically tagging them according to a tagged sample corpus. I’m not talking about text synthesis, anything to do with speech, automatic chat, question answering, or Alexa Skills.”
Though Winter observes the trend, he is not sure why 2018 was a tipping point. He writes:
“We’ve had some of the building blocks for this kind of text processing for decades, including the stats tools and the training corpuses. Does deep learning help? I don’t know but at minimum it helps by delivering sexy headlines that keep AI in the news, which in turn convinces business stakeholders this is something they can get behind. It wasn’t magic before and it’s not magic now; the output of these algorithms still requires some amount of quality control and manual inspection. But business leaders are now willing to admit that the old manual way of doing things also had drawbacks….”
The write-up goes on to observe that, while text categorization now works well enough for the mainstream, speech and conversation interfaces still fall short of flawless functionality. He directs our attention to this Google Duplex conversation agent demo as he alludes to some troubling trends in corporate AI deployment. He closes with a word to programmers wondering whether they should add natural language processing to their toolkits:
“The part of the question I can’t answer is how big is the job pool, how long will the bubble last and how much expertise do you need to get more money than you make now? … For myself, I’m learning the basic techniques because they feel core to my industry skill set. I’m staying open to chances to apply them and to work with experts. I’m not even at the midpoint of my career and want to stay ahead of the curve.”
It does seem that natural language processing is not about to go away any time soon.
Cynthia Murrell, January 10, 2019
The First Home Quantum Computer?
January 9, 2019
I read “IBM Unveils Its First Commercial Quantum Computer.” The write up stated with no trace of sarcasm:
we’re not quite there yet, but the company also notes that these systems are upgradable (and easy to maintain).
Gentle reader, do you know how to maintain a cryogenic system? No background in low temperature physics? No experience working with super cooled fluids? Hey, no problemo. IBM offers services too.
The write up points out that IBM wants the quantum computer to be a work of art. How about delivering useful computing capability?
Imagine this in your WeWork space:
It’s a nine-foot-tall and nine-foot-wide airtight box, with the quantum computing chandelier hanging in the middle, with all of the parts neatly hidden away.
What is more interesting is that IBM rolled out this product at the consumer electronics show.
Quick buy IBM stock. This practical device will deliver:
Games? No.
Applications? No.
Visualizations? No.
Er, PR? Yes.
Stephen E Arnold, January 9, 2019
Dedrone Study Analyzes Drone Activity at UK Airports
January 9, 2019
We wonder if would be bad actors are reading about drones?
A very brief write-up at OodaLoop calls our attention to an interesting study—“2018 Results: Dedrone UK Airport Counter-Drone Study.” By quietly installing monitoring devices at four UK airports,
Dedrone was able to track drone incursions, and deploy counter measures, at those locations. The company shares their results in a detailed blog post, complete with charts, conclusions, and a list of sources for further research. We suggest curious readers check it out. Meanwhile, the OodaLoop piece zeroes in on these takeaways:
“With 285 drones detected over a 148-day period (just under 2 per day), the study concluding with three key learnings and next steps: ‘1. the problem of unauthorized drones at airports is real, not anecdotal: Drones have appeared and disrupted UK airports in the past year, causing loss of revenue due to closed runways. Drone pilots fly a broad spectrum of technology from different drone manufacturers, and detection technology must be able to capture all drone activity: Drone detection systems must be able to detect all kinds of drones, regardless of the manufacturer. While DJI is the global market leader in drone technology by sales, they only represent 44% of the incursions at the airports studied. 3. UK drone pilots come out to fly at airports around the same time and days, and airports can strategically prepare for increased incursions during these period: The majority of the incursions occurred on weekend afternoons when drone hobbyists may be flying drones to capture footage for personal use.’ Finally, it is important to recognize that ‘all drones near airports are a threat, regardless of the pilot’s intentions.’”
Yes, I mentioned Dedrone deployed counter measures at its test airports when incursions were discovered. Though they may seem the most obvious, airports are not the only sites at risk from pesky drones. Dedrone has leapt upon on an emerging need—to secure organizations’ airspace from the increasing risk of drone intrusions. Founded in 2014, the company brought its first solution to market the next year. Based in San Francisco, they also happen to be hiring as of this writing.
Cynthia Murrell, January 9, 2019
Factualities for January 9, 2019
January 9, 2019
The New Year is off to a calculating start. Here are some factualities gathered when some people were watching balls dropped, corks popped, and snacks snacked.
20 percent. Percentage of Americans who think they have a food allergy. 10 percent do. Source: Neuro Science News
Four percent. Percentage of ride hailing users who take the service once a week. Source: Pew Research
75 percent. Percentage of venture capital firms with zero female partners. Source: Recode
1000 times bigger. The size of the virtual reality market compared to the augmented reality market. Evidence presented? None. Source: TechCrunch
5 million. Number of passport records Marriott lost to hackers. Source: Wall Street Journal pay wall
200 million. Number of smartphones Huawei shipped in 2018. Source: Android Central
100 million. Number of Amazon Alexa devices sold. In what interval, The Verge doesn’t include those data.
More than the GDP of Iceland. Amazon’s R&D spending for 2019. Source: Quartz
$22 million. Amount a seven year old made on YouTube in 2018. Source: Next Web
24 frames per hour. The speed of the very slow movie player. Source: Digital Reader
Looks like 2019 has numerical momentum.
Stephen E Arnold, January 9, 2019
Stolen Identities Affordable and Available from a Few Dark Web Vendors
January 8, 2019
This is quite the bargain for bad actors—Kodos Blog reports, “Hackers Charging £10 for Stolen UK Identities.” Writer Ali Raza tells us recent studies show packages of data required to usurp a victim’s identity can be found for as little as £10, or about 12 and a half bucks. We learn:
“On the dark web, these information packages are called fullz (full IDs), and they can be found on numerous black markets. They often contain things such as names, addresses, bank data, online passwords, and more. Researchers believe that a number of high profile hacks that occurred recently are keeping the markets filled with this type of data. Hundreds of millions of internet users have had their data stolen in 2018 alone. Some of the most famous hacks from the last few months include several Facebook incidents, the hack of British Airways, Marriott hotel, and more. Stolen information then gets posted on the hidden part of the web, known as the dark web.”
We presume in order to create more clients, vendors of such data also offer instruction in how to open loans and credit cards in the victims’ names. Not for free, of course; such a guide runs about £6. The article adds:
“[One] seller also offered a sample of stolen information, currently being in their possession. The data includes names, occupation, addresses, and even date of birth and similar information. The sample itself belongs to a Bristol-based Polish-born woman. Researchers have described this type of stolen information as ‘key to online fraud’. As the internet has become a large part of most peoples’ everyday life, demand for this type of info is constantly on the rise.”
And yet, Raza reports, most consumers have no idea how pervasive these data hacks are. In fact, says one expert, most of us already have had our data stolen and sold on the Dark Web—it’s just a question of how often. One can check whether their email address is believed to have been compromised at several websites, the most famous of which may be Have I Been Pwned. To prevent identity theft, users should follow best practices, like using separate, hard-to-guess passwords for different accounts and taking advantage of two-factor authentication where offered.
Cynthia Murrell, January 8, 2019
Microsoft and Kroger: Have These Outfits Actually Shopped at a Kroger Store in Kentucky?
January 8, 2019
The answer is, “Of course not.”
Kroger’s technological capability is modest, even by the low standards which define the Commonwealth of Kentucky. Home of the corrupt sports programs, bourbon, horse racing, river boat gambling, and other intellectual high water marks.
I read “Microsoft and Kroger to Create Data-Driven Connected Grocery Stores.” What this means is that Kroger wants to get rid of humans, keep the lights at mortuary levels, and not have to fool around with pesky customers who spend actual bank notes.
The write up takes a slightly different approach, stating:
The first fruit of the partnership is a digital shelving system, which was actually announced last year and is in the process of rolling out to dozens of Kroger stores across the U.S. Called EDGE (Enhanced Display for Grocery Environment), it bypasses paper price tags for digital shelf displays that can be changed in real time from anywhere, and it also can display promotions, dietary information, and more.
Yep, that’s an idea. But the flaw is that Kroger’s in Kentucky struggle to complete these tasks in an orderly, coherent way:
- Restock. Aisles are choked with people trying to cram products on shelves in aisles clogged with free standing cardboard promotions, mothers wrangling toddlers, and clueless males struggling to locate milk and bread.
- Functioning check outs. At the Louisville Westport Kroger, the store has a dozen next generation self check out machines. At 1225 pm Eastern exactly three of the machines were working. The hapless attendant was clueless and an even more confused “manager” was trying to calm down impatient shoppers. How many human check outs were open at this fine retail outlet? Exactly one. Yeah, Windows 10 will fix this puppy.
- Accurate data. I routinely locate products on shelves with prices different from what the Kroger check out systems display. The error rate seems to chug along at somewhere between 10 and 15 percent. The solution? Hide the prices so the hapless shopper will not be able to compare what one tag says with what the invisible database says. I suppose one could ask Cortana.
But the kicker is the idea that a shelf will illuminate only when a person is interfacing. It is pretty tough to buy a frozen burrito when the automatic illumination systems does not function. That assumes, of course, that one can actually locate frozen burritos which are in the frozen snack freezer two aisles away from frozen Mexican food.
Should I talk about the crazy Kroger app for wireless shopping and payment. Nope, I am heading to Whole Foods.
Stephen E Arnold, January 8, 2019
DarkCyber for January 8, 2019, Now Available
January 8, 2019
DarkCyber for January 8, 2019, is now available at www.arnoldit.com/wordpress and on Vimeo at https://www.vimeo.com/309717457 . The program is a production of Stephen E Arnold. It is the only weekly video news shows focusing on the Dark Web and lesser known Internet services.
The lead story is a profile of Sintelix, an Australian company developing software for law enforcement and intelligence professionals. The system can acquire content from the hidden Internet, the Surface Web, third-party sources, and content repositories in an organization; for example, arrest records. Sintelix provides IBM Analyst’s Notebook user with a streamlined, modern interface without giving up the unique features of the IBM Analyst’s Notebook. The three key features of the Sintelix technology are its speed of document and content processing. Hundreds of thousands of documents can be analyzed and indexed on a standard office desktop computer in a few hours. Sintelix also includes an application programming interface. This API makes it possible to use Sintelix with a wide range of third party solutions. Also, the system incorporates robust timeline features. Ana analyst can examine events over a month and then zoom into look at activities in an hour on a specific day.
The second story addresses a way to reduce the complexity of the Tor software bundle, which is required to access Dark Web sites, Many Tor users find the bundle confusing, which can lead to careless errors. . A number of user-induced errors can lead to the user’s loss of the privacy which the Tor software appears to offer. The fix is to use a hardware device which can run the Tor software. DarkCyber reports on an older system called PORTAL as well as a new Raspberry Pi approach. Will these devices provide a way to surf the Web in anonymity. Unlikely, but if properly configured, the devices may prevent some types of operator errors.
The third story discusses India’s legislation which mandates that technology companies provide access to encrypted content. Like Australia, India’s action is helpful to law enforcement and intelligence professionals. However, the mandatory decryption may increase the likelihood that bad actors will find a way to exploit the backdoor. The regulations require that a technology company like Apple or Facebook would have to respond to the government request within a day or two. Even with automated decryption technology, the time limit may prove difficult for some companies.
The final story describes a novel type of punishment for child abuse. The UK has begun deporting abusers to their country of origin and stripping the individual of his or her UK citizenship. So far one Indian who amassed 23 counts of child abuse have been flagged for deportation. Three abusers from Pakistan are likely to be deported as well. Once in their home country, authorities may take punitive action against the abusers.
A new blog Dark Cyber Annex will be available at www.arnoldit.com/wordpress. Cyber crime, Dark Web, and company profiles are now appearing on a daily basis.
Kenny Toth, January 8, 2019
Harvard Business Review on Moral Ethical AI
January 7, 2019
I love it when institutions which have a fine track record of generating highly ethical admission, investment, and curricula embrace AI big thoughts. I love references that everyone in Harrod’s Creek, Kentucky, understands and resonates. (Note the categorical affirmative because rural Kentucky is a bubbling pool of intellectual magma.)
Smart software seems to be one of the big topics in 2019 and we are only seven days into the new year. The main point of the HBR write up is to cheerlead for:
take[ing] the standards by which artificial intelligences will operate just as seriously as those that govern how our political systems operate and how are children are educated. It is a responsibility that we cannot shirk.
How will this work in a Harvard MBA fueled environment?
How are those ethics in business courses going? Wasn’t Mr. Zuckerberg admitted to Harvard where he hatched or appropriated for the Facebook service? How about professors who teach one thing and consult another and still manage to find the time to hold office hours—once in a while.
The problem with smart software is the problem of folks who want to get rich quick, break things, and invent the future.
Barn burned. Horses gone. New Amazon campus built on the land once occupied by an institution of higher education. Maybe today that should be “hire” education?
Yep, Aristotle’s son is a consultant to Google too. He may have been working on the Google AI which banned Winston Churchill card game because the entertainment depicts a cigar.
Stephen E Arnold, January k7, 2019
Amazonia, January 7, 2019
January 7, 2019
The Bezos bulldozer keeps on pushing through the virgin forest. Crunch, crunch—That’s the sound of the power of the machine creating new revenue streets and highways. Consider these bits of Amazonia:
One of the Five Eyes Is Smiling
One branch of the British government has inked a deal with Amazon to build the “Crown Marketplace.” Think in terms of the British version of GSA/DSA running on Amazon’s AWS infrastructure, buying goodies from Amazon’s warehouses, and getting some of the stuff delivered in nifty Amazon trucks. When will GHCQ follow the CIA’s approach and use Amazon for plumbing? Source: The Telegraph which dearly wants your email address.
GovCloud West: EC2 High Memory Arrives
Most commercial outfits won’t care or understand the steady expansion of the breadth and depth of the GovCloud. Mark your calendar, while some folks were guzzling Champaign, Amazon Amazon EC2 High Memory instances with up to 12 TB of memory to the US GovCloud West region. Source: Amazon itself. Want to know more about “high memory”? Click this link.
FBI Uses Amazon Facial Recognition Service
The policeware landscape is being reshaped by the Bezos bulldozer. Navigate to “FBI Pilot Programme Uses Amazon’s Controversial Facial Recognition Software.” Keep in mind that this write up comes from the ever friendly, always objective Sputnik News. The write up reports:
Sputnik reported that the artificial intelligence behind Rekognition, which can identify, track, and analyze people and recognize up to 100 faces in a single image, was being marketed by Amazon to US police departments for as little as $6 a month. That tiny fee gave law enforcement agencies access to Amazon Web Services (AWS). In turn, Amazon requested that those agencies recommend the brand to their partners, including body camera manufacturers, according to documents obtained by the American Civil Liberties Union (ACLU).
That’s a compelling price point for many law enforcement entities. True or false. Well, the secret region is a thing.
Perception Health Embraces the AWS Marketplace
The Amazon watchers at ArnoldIT.com noted this statement:
Perception Health, a leading provider of healthcare market prediction software, announced today their inclusion on the new machine learning (ML) and artificial intelligence (AI) discovery page on AWS Marketplace.
Why? Bezos’ bulldozer is turning to health. Perception Health wants to dabble in the machine learning marketplace Amazon has built along side its streaming data marketplace. Perception likes the strokes Amazon doles out to its partners. Good partner, the Bezos bulldozer rumbles softly. Source: PRNewswire
Where’s That Blog Belong?
The answer is on AWS. WordPress is a popular blogging platform. WPEngine stated:
WP Engine leverages a modern technology stack to make sure our customers have the resources they need to scale their WordPress environments. It’s why we give our customers access to a suite of developer tools they can use to build great websites, and it’s why we utilize best-in-class technologies like Amazon Web Services (AWS) to add resiliency and speed to our digital experience platform.
Different cheer, same enthusiasm. Source: WPEngine
PHP and Amazon
You know PHP. You want zero hardware to drag down your nights and weekends. You will embrace AWS Lamda. Details are in “Severless PHP on AWS Lambda.” If you want to know more about AWS Lambda, click here. Source: PHPDeveloper
Microservices on Amazon
Screw up one part of a microservice based app and you can have an exciting time of it. But what if one wants to combine the goodness of microservices with the Bezos bulldozer? No problem. Details plus code appear in “How to Deploy a Microservice Application to AWS.” Now about those microservices which don’t “service”? Sparse info, gentle reader.
H2O Analytics Run Better on AWS
Hard to believe that an Amazon partner helps market itself and Amazon with such enthusiasm. Here’s an example of nerd cheerleading:
If you haven’t started migrating your analytics to the cloud, then hopefully this will convince you to start reconsidering. The opportunity to have access to a 64, 96 or even 128 core machines with 2TB of RAM rarely crosses the path of most Data Scientists. This can mostly be accredited to the fact that most of us don’t really need such a large machine for what we need to achieve, see Szilard’s twitter posts if you need convincing. Another reason that we don’t use these big machines are purely because we just don’t have access to such machines within our working environments. Luckily for us, access to cloud computing have become more accessible and well, lets be honest, cheap as chips.
Yep, rah rah. Source: Digital Age Economist (aren’t all economists now alive “digital age economists”?)
Google and Its Software Engineering
January 7, 2019
Do you want to know how Google makes the Silicon Valley stew? Navigate to “Software Engineering at Google” and stick in your spoon. This technical paper will help some understand that when Google engineers “relaxing” of relevance algorithms or implements special user tracking features in Google Maps, the inclusions are not accidents. This may be useful information when someone considers a “surprising” behavior of a Google system. When I hear, “wow, that’s a surprise,” I will be a bit skeptical.
Stephen E Arnold, January 7, 2019
-
- Subscribe to Beyond Search
Feature archive
News archive
-
