Natural Language Processing: Useful Papers Selected by an Informed Human
July 28, 2020
Nope, no artificial intelligence involved in this curated list of papers from a recent natural language conference. Ten papers are available with a mouse click. Quick takeaway: Adversarial methods seem to be a hot ticket. Navigate to “The Ten Must Read NLP/NLU Papers from the ICLR 2020 Conference.” Useful editorial effort and a clear, adult presentation of the bibliographic information. Kudos to jakubczakon.
Stephen E Arnold, July 27, 2020
DarkCyber for July 28, 2020, Now Available
July 28, 2020
The July 28, 2020, DarkCyber is now available. You can view the program on YouTube or on Vimeo.
DarkCyber reports about online, cyber crime, and lesser known Internet services. The July 28, 2020, program includes six stories. First, DarkCyber explains how the miniaturized surveillance device suitable for mounting on an insect moves its camera. With further miniaturization, a new type of drone swarm becomes practical. Second, DarkCyber explains that the value of a stolen personal financial instrument costs little. The vendors guarantee 80 percent success rate on their stolen personally identifiable information or fullz. Third, SIM card limits are in place in South Africa. Will such restrictions on the number of mobile SIM cards spread to other countries or are the limits already in place, just not understood. Fourth, Coinbase bought a bitcoin deanonymization company. Then Coinbase licensed the technology to the US Secret Service. Twitter denizens were not amused. Fifth, Microsoft released a road map to a specific type of malware. Then two years later the story was picked up, further disseminating what amounts to a how to. DarkCyber explains where to download the original document. The final story presents DarkCyber’s view of the management lapses which made the Twitter hack a reality. Adult management is now imperative at the social media company doing its best to create challenges for those who value civil discourse and an intact social fabric.
The delay between our June 9, 2020, video about artificial intelligence composing “real” music and today’s program is easy to explain. Stephen E Arnold, the 76 year old wobbling through life, had the DarkCyber and Beyond Search team working on his three presentations at the US National Cyber Crime Conference. These programs are available via the NCC contact point in the Massachusetts’ Attorney General Office.
The three lectures were:
- Amazon policeware, which we pre-recorded in the DarkCyber format
- A live lecture about investigative software
- A live lecture about Dark Web trends in 2020.
Based on data available to the DarkCyber team, the septuagenarian reached about 500 of the 2000 attendees. Go figure.
Kenny Toth, July 28, 2020
2020: Reactive, Semi-Proactive, and Missing the Next Big Thing
July 27, 2020
I wanted to wrap up my July 28, 2020, DarkCyber this morning. Producing my one hour pre recorded lecture for the US National Cyber Crime Conference sucked up my time.
But I scanned two quite different write ups AFTER I read “Public Asked To Report Receipt of any Unsolicited Packages of Seeds.” Call me suspicious, but I noted this passage in the news release from the Virginia Department of Agriculture and Consumer Services:
The Virginia Department of Agriculture and Consumer Services (VDACS) has been notified that several Virginia residents have received unsolicited packages containing seeds that appear to have originated from China. The types of seeds in the packages are unknown at this time and may be invasive plant species. The packages were sent by mail and may have Chinese writing on them. Please do not plant these seeds.
And why, pray tell. What’s the big deal with seeds possibly from China, America’s favorite place to sell soy beans? Here’s the key passage:
Invasive species wreak havoc on the environment, displace or destroy native plants and insects and severely damage crops. Taking steps to prevent their introduction is the most effective method of reducing both the risk of invasive species infestations and the cost to control and mitigate those infestations.
Call me suspicious, but the US is struggling with the Rona or what I call WuFlu, is it not? Now seeds. My mind suggested from parts unknown that perhaps, just perhaps, the soy bean buyers are testing another bio-vector.
As the other 49 states realize that they too may want to put some “real” scientists to work examining the freebie seeds, I noted two other articles.
I am less concerned with the intricate arguments, the charts, and the factoids and more about how I view each write up in the context of serious thinking about some individuals’ ability to perceive risk.
The first write up is by a former Andreessen Horowitz partner. The title of the essay is “Regulating Technology.” The article explains that technology is now a big deal, particularly online technology. The starting point is 1994, which is about 20 years after the early RECON initiatives. The key point is that regulators have had plenty of time to come to grips with unregulated digital information flows. (I want to point out that those in Mr. Evans’ circle tossed accelerants into the cyberfires which were containable decades ago.) My point is that current analysis makes what is happening so logical, just a half century too late.
The second write up is about TikTok, the Chinese centric app banned in India and accursed of the phone home tricks popular among the Huawai and Xiaomi crowd. “TikTok, the Facebook competitor?’s” point seems to be that TikTok has bought its way into the American market. The same big tech companies that continue to befuddle analysts and regulators took TikTok’s cash and said, “Come on down.” The TikTok prize may be a stream of free flowing data particularized to tasty demographics. My point is that this is a real time, happening event. There’s nothing like a “certain blindness” to ensure a supercharged online service will smash through data collection barriers.
News flash. The online vulnerabilities (lack of regulation, thumb typing clueless users, and lack of meaningful regulatory action) are the old threat vector.
The new threat vector? Seeds. Bio-attacks. Bio-probes. Bio-ignorance. Big, fancy thoughts are great. Charts are wonderful. Reformed Facebookers’ observations are interesting. But the now problem is the bio thing.
Just missing what in front of their faces maybe? Rona masks and seed packets. Probes or attacks? The motto may be a certain foreign power’s willingness to learn the lessons of action oriented people like Generals Curtis LeMay or George Patton. Add some soy sauce and stir in a cup of Sun Tzu. Yummy. Cheap. Maybe brutally effective?
So pundits and predictive analytics experts, analyze but look for the muted glowing of threat vector beyond the screen of one’s mobile phone.
Stephen E Arnold, July 27, 2020
TileDB Developing a Solution to Database Headaches
July 27, 2020
Developers at TileDB are working on a solution to the many problems traditional and NoSQL databases create, and now they have secured more funding to help them complete their platform. The company’s blog reports, “TileDB Closes $15M Series A for Industry’s First Universal Data Engine.” The funding round is led by Two Bear Capital, whose managing partner will be joining TileDB’s board of directors. The company’s CEO, Stavros Papadopoulos, writes:
“The Series A financing comes after TileDB was chosen by customers who experienced two key pains: scalability for complex data and deployment. Whole-genome population data, single-cell gene data, spatio-temporal satellite imagery, and asset-trading data all share multi-dimensional structures that are poorly handled by monolithic databases, tables, and legacy file formats. Newer computational frameworks evolved to offer ‘pluggable storage’ but that forces another part of the stack to deal with data management. As a result, organizations waste resources on managing a sea of files and optimizing storage performance, tasks traditionally done by the database. Moreover, developers and data scientists are spending excessive time in data engineering and deployment, instead of actual analysis and collaboration. …
“We invented a database that focuses on universal storage and data management rather than the compute layer, which we’ve instead made ‘pluggable.’ We cleared the path for analytics professionals and data scientists by taking over the messiest parts of data management, such as optimized storage for all data types on numerous backends, data versioning, metadata, access control within or outside organizational boundaries, and logging.”
So with this tool, developers will be freed from tedious manual steps, leaving more time to innovate and draw conclusions from their complex data. TileDB has also developed APIs to facilitate integration with tools like Spark, Dask, MariaDB and PrestoDB, while TileDB Cloud enables easy, secure sharing and scalability. See the write-up for praise from excited customers-to-be, or check out the company’s website. Readers can also access the open-source TileDB Embedded storage engine on Github. Founded in 2017, TileDB is based in Cambridge, Massachusetts.
Cynthia Murrell, July 27, 2020
Disney and Face-Swapping: Real Actors Days Numbered
July 27, 2020
It took big bucks and a lot of time to insert virtual models of Carrie Fisher and Peter Cushing into 2016’s Rogue One: a Star Wars Story, but Disney may soon pull off similar tricks much more easily. The Verge tells us, “Disney’s Deepfakes Are Getting Closer to a Big-Screen Debut.” The studio presented the technology at the recent Eurographics Symposium on Rendering 2020 in London. The article shares Disney’s video illustrating the studio’s latest developments and contrasting them with earlier face-swapping technologies. It is well worth the investment of four minutes for anyone who is at all curious. Reporter James Vincent writes:
“The deepfakes you’ve probably seen to date may look impressive on your phone, but their flaws would be much more apparent on a larger screen. As an example, Disney’s researchers note that the maximum-resolution videos they could create from popular open-source deepfake model DeepFakeLab were just 256 x 256 pixels in size. By comparison, their model can produce video with a 1024 x 1024 resolution — a sizable increase. Apart from this, the functionality of Disney’s deepfake model is fairly conventional: it’s able to swap the appearances of two individuals while maintaining the target’s facial expressions. If you watch the video, though, note how technically constrained the output seems to be. It only produces deepfakes of well-lit individuals looking more or less straight at the camera. Challenging angles and lighting are still not on the agenda for this tech. As the researchers note, though, we are getting closer to creating deepfakes good enough for commercial projects.”
Left unmentioned are such projects’ darker possibilities—false video evidence of a crime, for example, or faked fodder for political scandal. Still, it is fascinating to watch this technology evolve. It is not surprising Disney’s financial motivations have gotten them this far.
Whitney Grace, July 27, 2020
You Know Times Are Hard When a Blue Chip Firm Stoops to SEO
July 27, 2020
Many years ago I worked at Booz, Allen & Hamilton. After I set up my own consulting firm, I did projects for other outfits which I thought operated in a blue-chip or high-quality mode.
If you read DarkCyber, you may have seen my articles making fun of some consulting firms’ analyses; for example, the outfits producing Gartner-type subjective comparisons enterprise vendors.
I have also been quite clear over the years about search engine optimization. The manipulation of a Web page feeds sales of online advertising and erodes what minimal objective relevance ranking methods remain in use. From my point of view, SEO is a scam. If you want traffic, buy advertising.
Why take time to write again about questionable consulting operations and SEO?
I received this email a day or two ago, and I have informed the sender that I would publish the email, his name, his contact information, and his employer before this item runs in my blog. Now the spam email. Please, note the chatty tone:
We noticed that you featured Boston Consulting Group in four of your articles (Gartner Magic Quadrant in the News: Netscout Matter, Radicati Group: Yet Another Quadrant, Search Engine Optimization: Chasing Semantic Search & Search Companies: Innovative or Not?) and wanted to say thanks so much for the mentions!
We were hoping you could add a link to our homepage [https://www.bcg.com/] in those articles so your readers can easily find the site. Please let me know if you have any questions or if I should direct this email to someone else. Thanks again for your help in advance.
Sincerely,
Connor Hayes
Connor Hayes Hayes.Connor@bcg.com
Global Search Senior Coordinator
T + 1 617 850 3941
Boston, USA
Allow me some observations, and I will offer some comments for Connor Hayes and other SEO “experts”:
1. Connor, and spare me your slathering of Dollar Store taco sauce. I am not into familiarity or hippy dippy “I want a link” pitches.
2. Boston Consulting Group, let’s be classy. SEO spam is something that I associate with outfits less well positioned to sell high-end professional services work.
I asked myself, “Was Connor Hayes influenced by Homer on “The Simpsons”?
I asked myself, “Has BCG lost its sense of professionalism?”
I do recall learning from my father who worked for an entrepreneur R. G. LeTourneau that General Eisenhower and later president of the United States was not impressed that Bruce Henderson, founder of BCG “borrowed” the four square matrix analytic tool. When I heard this anecdote, I suppose the state was set for today’s BCG to embrace search engine optimization. Both the four square star-dog thing and SEO illustrates a similar thought process: Do what needs to be done to become a modern day winner.
I segment the world of professional services consulting into some simple chunks. At the bottom are newly unemployed managers, unemployable college graduates with degrees in home economics, art history, or some similar expertise, and people who just cannot stick with a legitimate company. Many of these individuals become SEO experts.
Then there are mid-tier consulting firms. These firms capture government contracts, find a niche and generate information and knowledge products, and pontificate on LinkedIn about their organizations’ mastery of knowledge-value in today’s world.
The third group is the top of the professional services pyramid. My perception was that the big leagues attracted the best and the brightest. Examples of these top-tier operations included Arthur D. Little, Bain (formed by unhappy professionals at Boston Consulting Group), BCG itself and its four square star dog thing, Booz, Allen & Hamilton, McKinsey & Company, SRI, and a handful of others.
The names I assign each level are:
- Pigeons, the flocks of consultancies pecking for anything that will sustain them
- Azure-chip consultants, the myriad of good enough firms that pontificate on everything from Amazon AWS to Zulu refugee buying preferences in South Africa
- Blue-chip consultants, the Big Leagues of professional consulting and advisory services.
Some observations are warranted, at least to my way of thinking:
- Blue-chip consulting firms once marketed via word of mouth, repeat business, and sponsoring awards like the original McKinsey payoff for the “best” Harvard Business Review article. Sorry, BCG, McKinsey aced you out there. SEO is definitely a winner for some like Twitch or YouTube luminaries. (Why not retain Dr. Disrespect to build an audience for BCG’s services? He is available for promotional work at this time I believe?)
- The economic downturn appears to require scraping the dregs from the wine barrel for sales leads. Yes, SEO, the better way. Forget the white papers, the speeches, and the thought leadership. It is apparently short cut time.
- The larger issue is that desperation marketing seems to be okay for a once-prestigious firm’s management team. The use of Connor Hayes-type intellects to get me to point to a formerly respected consulting firm is either the sign of a Ted Kaczynski-type thought process or stupid.
Net Net: The fact that BCG appears to endorse and desire SEO backlinks is more evidence of a decline within the ranks of top-tier consulting firms’ marketing and PR methods.
PS.
Connor Hayes, as you progress in your SEO career, why not get ManyVids or TikTok influencers to promote BCG? Let me know when you become a partner, please.
Stephen E Arnold, July 27, 2020
EU Wants Google to Promise It Will Not Use Fitbit Data to Enhance Search
July 27, 2020
We noted “Europe Wants Google to Pledge That Fitbit Data Won’t Further Enhance Search.” Let’s see what “pledge” means:
Your Dictionary says: “The definition of a pledge is something held as security on a contract, a promise, or a person who is in a trial period before joining an organization. An example of a pledge is a cash down payment on a car. An example of a pledge is a promise that you’ll buy a person’s car.”
Dictionary.com says: “A solemn promise or agreement to do or refrain from doing something:a pledge of aid; a pledge not to wage war. Something delivered as security for the payment of a debt or fulfillment of a promise, and subject to forfeiture on failure to pay or fulfill the promise.”
Wordsense.eu says: “From Middle English plege?, from Anglo-Norman plege?, from Old French plege? (Modern French pleige?) from Medieval Latin plevium?, plebium?, from Medieval Latin plebi?? (“I pledge”), from Frankish *plegan? (“to pledge; to support; to guarantee”), from Proto-Germanic *plehan?? (“to care about, be concerned with”). Akin to Old High German pflegan? (“to take care of, be accustomed to”), Old Saxon plegan? (“to vouch for”), Old English pl?on? (“to risk, endanger”).”
The write up says:
EU regulators are asking Google to pledge that Fitbit information will not be used to “further enhance its search advantage.” Another demand involves letting third-parties have “equal” access to that data.
DarkCyber’s comment: Ho, ho, ho. Guarantee? Data are ingested and processed. Ho, ho, ho. No humans involved. Ho, ho, ho. It’s an artificial intelligence system. Ho, ho, ho. Let the lawyers figure it out. Ho, ho, ho. Fitbit users buy products, and Google wants to sell like Amazon. Ho, ho, ho.
Stephen E Arnold, July 27, 2020
Adulting: An Update
July 26, 2020
DarkCyber wants to call attention to examples of adulting. The term refers to a behavior once associated with a responsible, civic-minded firm. High technology companies embrace the principles of high school science club management. Perhaps these examples indicate that high schooling is yielding to adulting, just slowly and with baby steps.
Example 1. “Google will replace Nest thermostats affected by w5 Wi-Fi error” reports a responsible action by the GOOG.
Example 2. “Twitter says it’s looking at subscription options as ad revenue drops sharply” suggests that one of the all-teen, all-the-time darlings of the technically elite may think about an action DarkCyber considers long overdue. We will have to wait and see. Adulting is painful, a bit like giving up one’s dream of running two companies from a far off land.
Example 3. “Amazon, Google and Wish remove neo-Nazi products” reveals that three essentially manic science club operations have reached a simultaneous moment of maturing. Each will remove products identified with potentially destructive concepts. Maturity comes slowly it seems even in the zoom zoom world of high technology.
Stephen E Arnold, July 27, 2020
Honeywell: Yep, Our Sweet Quantum Computer Is the Blue Ribbon Winner
July 25, 2020
Who has the world’s fastest quantum computer? Is it IBM, Microsoft, Apple, or Google? No, none of these companies have that claim to fame. According to The Motley Fool that honor belongs to, “Honeywell Unveils The World’s Fastest Quantum Computer.” Quantum computers are still reserved for companies, universities, and governments with deep pockets, but Honeywell’s newest machine is making them one step closer to commercial use.
IBM used to own the fastest quantum computer, but Honeywell’s device has a process with 64 quantum volume. IBM’s machine only has 32 quantum volume capability. The Honeywell quantum computer processes six cubits. A cubit is a quantum computing unit that stores and processes more than ones and zeros. Most computers are still limited to the famous ones and zeros from binary code. Honeywell’s computer also has a 99.997% fidelity score, meaning it can compute simulations and calculations of high quality.
Quantum computers are still in a state similar to the behemoths that dominated basements last century. Ironically, quantum computers are large themselves:
“The Honeywell system is another step forward in a long and difficult process. Scientists expect quantum computers to handle problems that are essentially unsolvable with current technology in fields such as cryptography, weather forecasting, artificial intelligence, and drug development. However, that future lies many years ahead. These are very early days in the development of usable quantum systems.”
Honeywell does not claim to have the best quantum computer, only the fastest. At doing what exactly?
Whitney Grace, July 25, 2020
Intel: Distracted by Horse Ridge, Engineers Take Another Detour
July 24, 2020
My hunch is that you did not read “Intel Introduces Horse Ridge to Enable Commercially Viable Quantum Computers.” You probably don’t care about some of the hurdles quantum computers face with or without Horse Ridge; for example, cooling, programming, and stability. That’s okay. The magic of the “quantum” horse thing was news only a sparse pasture below the ridge can appreciate. I have in my files one snippet from the PR output, however:
Horse Ridge is a highly integrated, mixed-signal SoC that brings the qubit controls into the quantum refrigerator — as close as possible to the qubits themselves. It effectively reduces the complexity of quantum control engineering from hundreds of cables running into and out of a refrigerator to a single, unified package operating near the quantum device.
Yep, the quantum refrigerator.
Now flash forward to “Intel’s 7nm Is Broken, Company Announces Delay Until 2022, 2023.” The write up explains:
Intel CEO Bob Swan said the company had identified a “defect mode” in its 7nm process that caused yield degradation issues. As a result, Intel has invested in “contingency plans,” which Swan later defined as including using third-party foundries.
Perhaps Intel will consider shifting its R&D focus to refrigeration units. Serving the quantum computing sector seems to be a way to pivot from a business in which Amazon Gravitons, AMD chips, and Apple’s custom designed ARM silicon are making headway.
Is Intel’s future horse features. Ooops. I meant Horse Ridge. Is that a glue factory under construction on a site adjacent Intel’s new fabrication facility?
Stephen E Arnold, July 24, 2020
-
- Subscribe to Beyond Search
Feature archive
News archive
-
