IBM Watson: Now Thinking Critically
May 21, 2020
The Watson AI Lab with a team from Harvard and MIT developed a new AI dubbed Clevrer that reasons and recognizes casual relationships. Venture Beat shares the details in the article, “MIT Researchers Release Clevrer To Advance Visual Reasoning And Neurosymbolic AI.”
Clevrer is built on the data set Clevr developed in 2016 by Stanford and Facebook AI Research. It was designed to analyze visual reasoning abilities of neural networks and Neuro-Symbolic Concept Learner was added in 2019. The data set includes a 20,000 synthetic videos of colliding objects paired with over 30,000 natural language questions and answers about objects in the videos. The data set is important about for building smarter AI:
“MIT-IBM Watson Lab director David Cox told VentureBeat in an interview that he believes the data set can make progress toward creating hybrid AI that combines neural networks and symbolic AI. IBM Research will apply the approach to IT infrastructure management and industrial settings like factories and construction sites, Cox said.
‘I think this is actually going to be important for pretty much every kind of application,’ Cox said. ‘The very simple world that we’re seeing are these balls moving around is really the first step on the journey to look at the world, understand that world, be able to make plans about how to make things happen in that world. So we think that’s probably going to be across many domains, and indeed vision and robotics are great places to start.”
Clevrer consists of two AI that are complimentary. It will build better and reliable models that will require less data to “train” them and they will also be more energy efficient.
AI are being designed to handle huge data streams, but it is better if they are designed to solve problems. AI systems needs to have a logical components, be able to reconfigure themselves, act on environment, interpret information, and define their own mental models. AI systems need to be smarter than a calculator and better than the modern enterprise system.
Wasn’t there an IBM patent for a “Clever” system? Yes, gentle reader, there was. Hence, clevrer. Clever, right?
Whitney Grace, May 21, 2020
Translation: Improvement Attracts Money
May 21, 2020
Foreign languages remain a problem for modern society, even with the bevy of translation software available. Most translation software lack native language fluidity and are unreliable. Lilt makes AI-powered business translation software and Venture Beat says that: “Lilt Raises $25 Million For AI Enterprise Translation Tools.” Lilt plans to use the money for further NLP research and go-to-market strategy acceleration.
Lilt’s clients translate information into seven languages and find the manual translation process slows down business practices. Lilt overcomes translation issues with:
“Lilt tackles this with human translators and CAT, a tool that helps them work more efficiently, using hotkeys, style guides, and a proprietary neural machine translation engine. CAT can be tailored to a company’s content, translation history, and other linguistic assets and configured to automatically add in previously translated segments when it finds matches within documents. The tool’s termbase and lexicon features help translators use the correct terminology in a given context, chiefly by showing them a range of possible translations for a certain word. And the engine taps AI and machine learning to analyze translation data and make predictive suggestions.”
Like most AI technology, Lilt’s systems requires new data, in this case languages, to learn. Translators work with the engine to accept, amend, or reject its translations.
The company has competitors such as Unbabel and the market for AI-based translation software is projected to be worth $983.3 million by 2022.
Didn’t Google “solve” machine translation already too? Obviously not completely.
Whitney Grace, May 21, 2020
JEDI? What JEDI?
May 21, 2020
The battle royale players are Amazon, Microsoft, White House officials, the Department of Defense, former employees of high profile firms, law firms, and consultants. The subject? JEDI. The procurement has been entertainment worthy of a William Proxmire Golden Fleece Award. (Remember Senator William Proxmire?)
DarkCyber spotted “Scoop: Google Lands Cloud Deal with Defense Department.” We know this is a scoop because the word “scoop” appears in the headline. Subtle.
The write up reports as a real news scoop:
Google Cloud has landed a deal to help the Defense Department detect, protect against, and respond to cyber threats, Axios has learned. The deal, with the Defense Innovation Unit (DIU), is in the “seven figures…”
The main point is to make clear that it will be business as usual at the Pentagon. The single vendor idea is not making much headway when it comes to information technology.
What’s next? Awards to Hewlett Packard, IBM, and Oracle?
Good question. We thought we heard cheers from the buildings near the old SeaWorld in Silicon Valley. Maybe that was a party held at IBM Federal off Quince Orchard Road in Gaithersburg?
Probably our team’s collective tinnitus.
Stephen E Arnold, May 21, 2020
German Intelligence Handcuffed
May 20, 2020
DarkCyber noted an interesting news story published on DW.com. The article? “German Intelligence Can’t Spy on Foreigners Outside Germany.” The DarkCyber research team talked about this and formed a collective response: “What the heck?”
The write up reports as actual factual news:
The German government must come up with a new law regulating its secret services, after the country’s highest court ruled that the current practice of monitoring telecommunications of foreign citizens at will violates constitutionally-enshrined press freedoms and the privacy of communications.
The article continued:
The ruling said that non-Germans were also protected by Germany’s constitutional rights, and that the current law lacked special protection for the work of lawyers and journalists. This applied both to the collection and processing of data as well as passing on that data to other intelligence agencies.
This is an interesting development if it is indeed accurate. Some countries’ intelligence agencies do conduct activities outside of their home countries’ borders. Furthermore, there are specialized service and device vendors headquartered in Germany which facilitate extra border data collection, monitoring, tracking, and sense making. These range from Siemens to software and hacking companies.
Restricting the activities of an intelligence unit to a particular geographic “space” sounds like a difficult task. There are “need to know” operations which may not be disclosed to an elected body except under quite specific circumstances. Electronic monitoring and intercepting ranges freely in the datasphere. Telecommunications hardware and service providers like T-Mobile have a number connections with certain German government entities.
Plus DarkCyber surmises that there are current operations underway in certain parts of the world which operate in a way that is hostile to the German state, its citizens, and its commercial enterprises.
Will these operations be stopped? Turning off a covert operation is not like flicking a button on a remote control to kill a Netflix program.
What if the German intelligence community, known to be one of the best in the European Community, goes dark?
The related question is, “What if secret agencies operate in secret?” Who will know? Who will talk? Who will prosecute? Who decides what’s important to protect citizens?
Stephen E Arnold, May 20, 2020
Googler Word Choice: A Delight for Rhetoricians
May 20, 2020
DarkCyber readers may find the “we are excited to talk with you” interview transcript interesting. The star of the interview is the Google boss Sundar Pichai. The expert real news people grilling the replacement for the duo of Larry Page and Sergey Brin are survivors from the downsizing at the Verge. Please, read “Sundar Pichai on Managing Google through the Pandemic: The CEO of Google and Alphabet Joins the Vergecast.”
We noted these Googley words and phrases used by the top Googler, presented as they appear in the interview transcript. The flow is as interesting as words.
Really
Big
Diversity
Foundational
Deeply as in “deeply committed”
Transparency as in “transparency reports”
Shared
Progress as in “modest progress”
Important as in “really important”
Scale as in “scale up better”
Definitely, quite a few definitely
Viewpoints as in “political viewpoints”
Progress as in “made a lot of progress”
hiccups as in “definitely going to have hiccups”
Deeper as in “deeper efforts” and “deeper investments”
Business as in “sustainable business”
Go to as in “go to market investments”
Financial as in “financial sustainability goal”
Ecosystem as in “guide our ecosystem”
Pivotal
Committed as in “pretty committed”
Demonstrated as in “we clearly have demonstrated”
Care as in “we care all the way”
Deeply as in “deeply passionate”
Well as in “really well”
Stuff as in “all that stuff”
Search as in “highly ROI driven”
For sure
Reverted back
Hard as in “hard for me to say”
Clearly
Conservative as in “conservative on the return back for the broad company”
Prioritizing as in “prioritizing people”
Actually as in “actually kind of need to be there”
Buckets as in “have people in two different buckets”
Play as in “to make that play out”
Understand as in “understand what works”
Data as in “driven by data”
Phase as in “brainstorming phrase” and “next phase”
Lines as in “blurred the lines”
Herd as in “herd a bunch of people”
Momentum
Place as in “get to the right place”
RCS
WebRTC
Common as in “common work” and “common teams”
Iterate
Flexibility
Answer as in “user answer” and “technical answer”
Align
Platform
Integration
Behind as in “Android has been behind”
Fragmentation as in “fragmentation in Android”
pain as in “real pain”
Simplifying
Efficiency
Productivity
Touch as in “in touch”
Constraints
Everything as in “we want to do everything”
Onus
Communicating
Breakthrough as in “AI breakthrough”
Separation as in “structural separation”
Think as in “think through that breadth”
Bets as in “different bets”
Play as in “technology play”
Commonality as in “underlying commonality”
Space as in “Internet space”
Convergence
Folks
Focus as in “a lot of focus”
Terrific as in “terrific effort”
Easy as in “make it easy”
Obviously as in “obviously with user consent and privacy protection”
Consistently
Toolkit as in “one more toolkit”
Public as in “just go public”
Conversation as in “responsible conversation”
Basically as in “we basically made that decision”
Extraordinarily as in “extraordinarily public moment”
Information as in “high quality information”
Trumps as in “trumps everything”
Deep as in “deep technological underpinnings”
Need as in “need to step up”
Big as in “big value chain”
Surface as in “surface the highest quality information”
Evolved as in “evolved our approaches”
Inputs as in “we took inputs”
Expertise
Learnings
Flexible
Something as in “something like that”
Compartmentalize
Normalcy as in “real sense of normalcy”
Disruptions as in “disruptions are kind of concerning”
Transitions
Space as in “space to think quietly”
Progress as in “progress better”
Reiterative as in “reiterative process”
Horror
Force block as in “force block times”
Boundaries
Pizza
YouTube as in “YouTube cooking video”
Pattern shifts
Revert as in “people revert back”
Long run
Shifts
Moments as in “moments of opportunity”
Worth
Global as in “global movement”
Humanity
Whole
Trends
Common
Abstractions, colloquialisms, and platitudes would make a 15th century rhetoric teacher chuckle.
Stephen E Arnold, May 20, 2020
Microsoft and Jack Dorsey: Chasing Zen Land?
May 20, 2020
We spotted “Permanent Work from Home Damaging for Workers Well-Being: Nadella.” The write up reports:
Microsoft CEO Satya Nadella has warned that making remote work permanent could have serious consequences for social interaction and mental health for workers as virtual video calls cannot replace in-person meetings.
Nadella asked these questions:
What does burnout look like? What does mental health look like? What does that connectivity and the community building look like? One of the things I feel is, hey, maybe we are burning some of the social capital we built up in this phase where we are all working remote. What is the measure for that?”
What company acts in a manner that will destroy well being? What company pushed employees to burnout? What company tears down communities?
DarkCyber’s research team believes that Jack Dorsey, the multi tasking CEO, fosters these negatives. As you may know, the capitalist tool, shared “Twitter, Square Announce Work From Home Forever Option- What Are The Risks?”
Microsofties can enjoy commutes, chasing around Microsoft’s facilities looking for guests, and sitting in meetings where Windows 10 updates, Amazon’s encroachment on Windows World, and waiting for the US government to award JEDI to the firm which developed Bob, the user interface of the future.
Disagreements about work are interesting. However, concerns and opinions from commercial real estate companies might be more helpful.
Amusing nevertheless. Nadella and Dorsey arguing about worker welfare. When monopolies collide, the well being of workers becomes ascendant.
Stephen E Arnold, May 20, 2020
Harvard Channels MIT: Academic Funding Magnetism
May 20, 2020
The study of mathematical principles that guide evolution is a fascinating field, and Harvard’s Program for Evolutionary Dynamics is a worthy research and teaching program. Its goals include, among others, finding cures for cancer and for infectious diseases. Unfortunately, like many poised in an ivory tower, its director seems to have been afflicted with greed. The Harvard Crimson Reveals, “FAS Places Prof. Nowak on Leave after Report Finds Epstein Used His Program to Rehabilitate Image.” Reporter James S. Bikales writes:
“A University report found Epstein attempted to use Harvard and the Program for Evolutionary Dynamics, which Nowak directs, as a tool to rehabilitate his image following his 2008 conviction for solicitation of minors for prostitution. Epstein likely made more than 40 visits to PED’s offices at One Brattle Square between 2010 and 2018, according to the report, which also states that Nowak approved the posting of flattering and false descriptions of Epstein’s philanthropy and support of Harvard on the PED website.”
Though no evidence was found that donations from the (alleged) underage-sex-ring facilitator and serial abuser were accepted after his conviction, he had donated millions to the PED in the recent past. Epstein also helped facilitate a John Templeton Foundation grant to the program in 2015, which was accepted. Certain pre-conviction perks were also supplied to the convict-to-be, including a fellowship he was unqualified for and an office complete with keycode access to the PED building. There is no evidence Epstein interacted with students during his approximately 40 visits, aside from sitting in on one undergrad math class.
While awaiting trial on federal charges of trafficking and sexually assaulting at least 80 underage girls, Epstein died in August 2019 in his prison cell. Though likely to be less dramatic, Nowak’s fate is still to be decided pending an investigation.
Cynthia Murrell, May 20, 2020
How to Be an Expert: A Very Troubling Method
May 20, 2020
DarkCyber found itself worrying about knowledge. Epistemology is not exactly the hot topic in the midst of the collapse of the unicorns, blue skies in Wuhan, and tiny animals in major US cities.
The write up “4 Unexpected Methods for Becoming an Authority on Nearly Any Subject” contains some troubling advice. DarkCyber fears that many search engine optimization, content marketing, and MBA carpetbaggers will be off to the races after taking the advice in the Copyblogger article.
Make Stuff Up and Tell People Who Don’t Know the Subject
We noted with some discomfort:
If you’re the best in the place where your customers hang out, you’re the best. Don’t turn your nose up at being a big fish in a small pond. There’s a lot of success, satisfaction, and wealth to be found in small ponds.
We think this means make up stuff and pitch it to drop outs from the Baraboo clown school. Is that right?
Another point caught our attention:
Make complicated topics easy to understand.
Yes, but… and the but is important. What if the person explaining quantum mechanics or RNA replication does not understand these subjects. Isn’t this similar to the previous idea of pitching incorrect or misleading information to hungry minds in Thurmon, West Virginia.
Teach the subject.
Okay, a person who does not know the subject teaches the methods of nuclear fuel management and also Hopf fibration calculations. That sounds like a winning approach. How does one get this teaching job, pray tell?
Commit to a sincere desire to help.
What? So a person without the requisite expertise is going to supervise drone operations in a war zone? Is a person with zero knowledge of financial markets is going to engage in currency trading at a quant firm on Wall Street?
The write up illustrates the disdain the marketing profession has for those with concrete, high value information. This is not intellectual dishonesty. The use of any of these methods is far worse.
The enshrinement of duplicity. Super.
Stephen E Arnold, May 20, 2020
Lucidworks: Buzzwording in the Pandemic
May 19, 2020
Lucid Imagination (the outfit which contributed some Lucene/Solr talent to Amazon search) renamed itself Lucidworks. The company then embarked on becoming a West Coast version of Fast Search & Transfer, a Splunk like outfit, and now a customer support provider.
That’s a remarkable trajectory for a company built on open source software with more than $200 million in funding since 2007.
One of the DarkCyber researchers spotted “Lucidworks Develops Deep Learning Solution to Make Chatbots Smarter.” The story appeared in a New Zealand online publication. That’s interesting, but more intriguing is that Lucidworks is following in the marketing footsteps of Attivio, Coveo, and other vendors of search and retrieval. The destination customer service. Who doesn’t love automated customer support chat robots, self serve Web sites with smart software, and the general extinction of individuals who actually know a company’s software or hardware products?
The write up states:
Deep learning is essential for automated chatbots to understand natural language questions and to provide the right answers, which is something that AI-powered search firm Lucidworks has taken on board.
And why?
According to Lucidworks, companies rely on digital portals to provide information to users, whether digital commerce customers looking for product information before purchase, employees hunting for an HR document, or someone looking for an airline’s updated cancellation policies. Information is often scattered across disparate silos and is impossible for a user to locate using natural language questions.
But smart software is available from Amazon with a credit card and some free training courses. Outfits from Algolia to Voyager Search offer the service.
What is interesting is the buzzword salad tossed into this reheated plastic container of mapo tofu:
- AI (artificial intelligence)
- Automated
- Chatbots
- Conversational
- Deep learning
- Digital portals
- Engagement
- Experiences
- Fusion
- Natural language
- Satisfaction
- User intent
- Virtual assistants
Quite vocabulary and what seems an exercise in content marketing. Plus, eager customers in New Zealand will have an opportunity to help the company repay its investors the $200 million plus interest. That works out to 13 years in the enterprise search wilderness before arriving at chatbots.
Options abound and many of them are open source and well documented.
Stephen E Arnold, May 19, 2020
The Cloud Winner? It Is Definitely the Google Opines InfoWorld
May 19, 2020
I find “real” news interesting. Consider “13 ways Google Cloud beats AWS.” Clouds are becoming more alike. Microsoft is making friends at Oracle. IBM (yep, the Watson wizards) are reaching out to Amazon.
But IBM did not reach out to Google. Maybe HP’s cloud division will be Googley. After reading the objective article, the Googley cloud is the big dog.
Let’s look at some of the 13 reasons, excluding the Firebase reference and the use of predatory pricing to win business. (If it worked for Oracle when fighting for some Zoom love, deep and steep discounts may work again.) DarkCyber calls this the Walmart way.
On to the objective, totally factual statements. Not 12 or 14, exactly 13. Must be a lucky number? There was neither an illustration of a Googler walking under a ladder or a black cat risking death wandering Shoreline Drive.
Google has a health care API. Google’s Deepmind also found itself squabbling publicly about its use of certain health care in the UK. Has Google sanitized its act with soap or ultraviolet light? Some evidence about the value of Google’s health care API would be helpful.
Embedded machine learning. The hyperbole about smart software reaches Harrod’s Creek, Kentucky. The idea that one can use the cloud and plug into pre-crafted, ever flexible, smart machine learning modules is an interesting one. Are Google’s systems and methods “better than” SageMaker and the dozens of other AWS doodads littering the company’s oddly disjointed, sometimes bizarro documentation? No, but at least AWS has documentation.
Custom cloud machines. Are most enterprise cloud vendors touting the thrills and excitement of non-standard “machines”? Is Google “better than” Amazon in this aspect of enterprise cloud computing. A standard engineering practice is not unique without some checklists, benchmarks, and technical feature comparison. I can say that Kentucky bourbon is good for “real” journalists, but that statement requires some “proof” beyond a factless article.
There are 10 more of these Google PRish gems in the original article.
But let’s come back to deep discounts. Buying business is going to be a go to strategy for most cloud vendors. What will people buy? Smart software or security? Price or performance? Commodities or cheerleading?
Yikes, Google is the winner.
Stephen E Arnold, May 19, 2020
-
- Subscribe to Beyond Search
Feature archive
News archive
-
