Just got an interesting glimpse into the future of "spook shop" media trolling. Everybody worth the powder to blow them to hell claims some sort of social "sentiment analysis" capability. But the next step in analytic tagging of the media fire hose is "intent analysis." After all, clumsy sentiment analysis is a simple thumbs-up/thumbs-down/neutral indication with pretty abysmal accuracy ratings (the algorithms can't distinguish sarcasm from honest opinion). That's why I've relied on Quantum Leap's PBA (Pattern Based Analytics) to give me a more accurate picture of really trending themes. I can do a pretty good job of predicting GROUP INTENT based upon Quantum Leap Buzz, and regard it as a cool (even indispensable) tool to have in myt bag of predictive marketing tricks.
But what about the "spook shops" and their deep dive into a single thread, a single document, a single thought in context? How do you determine the INTENT of a single author? I just got an enlightening monograph today from Israeli/American friend of mine, and he gave me permission to publish it. If I were a Cyber Privateer who wanted a "Dead Man's Switch" (which I describe in Principle #1 of The Perfect Virus, Oversight), I'd implement it with an INTENT ANALYSIS CCS (Command and Control System) and then turn it loose on mankind. This Dead Man's Switch would make my continued health an important consideration to a world at the mercy of my Perfect Virus.
So thank you Michael S. Pincus (firstname.lastname@example.org), founder and president of Mnemotrix Systems, Inc., for an important addition to my Perfect Virus Oversight Module: a "dead man's switch" that can sniff out my enemies long after they've put a bullet into my head. Here's Mike's monograph in its entirety:
There is in the private sector developed and owned by Mnemotrix Systems, Inc., an intelligence sensemaking automaton that can connect-the-dots from Internet scraped data and obtained memos, reports, and conversation transcripts based on metaphoric and semiotic reverse inferencing exposing both expressed but hidden intentions and underlying subconscious thoughts in social media, reports, articles, news and other material from bad guys.
One of the greatest advantages this technology provides for the community is it can take metaphors and apply it to social media. Scraping millions of tweets and blogs through metaphors and being able to distill INTENT is a vital key to locating real threats without bias. The automaton's reasoning is based on a human savant model of the neural linguistic patterns in the human subconscious. Analysts using the automaton to help connect-the-dots have reported that it helps to awaken their subconscious through reflections in the data teaching the analyst to think like the subconscious mind "thinks." It helps the analyst augment those behind-the-scenes thought processes that are always running but frequently buried by pervasive thoughts and bias helping the analyst to locate INTENT in data.
This is not a research project waiting for funding. It is fully complete technology already in place, in places.
The following examples are real from a recent study of the system done by an analyst in the Pentagon.
The analyst sent the following:
The SISMA automaton will augment analysis capabilities by allowing the analyst to type in something like "sudan reaction to action" and it comes up with "South Sudan accuses Sudan of Airstrikes". Bingo. The query "Mali civilian crisis" comes up with "in the face of the deteriorating security situation in the north of Mali, and the looming food crisis in the region, I urge all parties to take care of the civilian population and ensure respect for human life" where the keywords are "Mali" "crisis" and "civilian". While that might not seem impressive, it came from an article entitled "EU suspends development operations in Mali". An analyst might not have found this information if he was looking for EU reactions to the Mali coup.
Thought is not linear, usually. Of course, one must be careful of mirror image bias, but one must always be. The analyst
working in the Pentagon also ran a non-specific query "reaction to a threat" into the 5-day Forecast, and got articles with some of the interesting lines (with the analysts comments in **), but to preface, in the first article (Headline: Insight: Valentine's Day mission gives BOJ new personality, Tokyo Reuters) the only 2 'words' the savant automaton drew out of "reaction to a threat" were 'challenge' and 'action'; an analyst would not find the below concepts with other search engines:
In another query "reaction to a threat" it found from other material: "The governor had just been publicly heckled by politicians over seemingly endless deflation and a yen near record highs" (*this is the threat*). "Bold steps were needed to restore the BOJ's standing, and quickly" (*this sets up the reaction to that threat*) - "...working that weekend at the BOJ knew surprise would be the key to success and so abstained from the usual practice of leaking plans to the media to massage expectations." (*a reaction to a threat*) - "So the BOJ very deliberately sent investors and politicians the signal they have been waiting for -- that it was ready to deploy unconventional weaponry to lead Japan out of deflation" (*a reaction to a threat*) - "Those who have worked with him say he is particularly uncomfortable doing something just for psychological effect (*reaction to a threat*) and in his view that includes feeding more cash to banks that have trouble absorbing the funds already available to them (*the threat, or problem that the reaction might create*) - "Indeed, the central bank has found itself playing down the impact of it's February move to temper expectations of further bombshell steps." (*another reaction to a threat*)
There are many more instances of reactions to threats in this article. Perhaps the article would be found looking for " 'bank*' AND 'japan*' AND 'BOJ' AND 'deflation' but you might get myriad other responses.
Indeed when this search string "reaction to a threat" is put through Google it comes up with a lot of things that are close, and some that aren't. However, I would challenge the analyst to decipher, in simple terms, the overall theme of the reporting. Sure, in time one could come up with it, deflation is high and the BOJ is doing something about it - but to say "the BOJ may react X way to currency deflation" and discover this immediately saves a lot of time, especially if Japanese economy is one's area of interest.
In another Reuters article, under the same search query "reaction to a threat", the automaton found this particularly illuminating hit: "Bahrain's Sunni Muslim minority, fearful of Shi'ite political assertiveness (*this is the threat*), is spawning factions that rail against compromise with the island's sectarian majority (*this is the reaction to the threat!*), while nursing their own grudges against its Sunni ruling family (*a tertiary reaction to the threat!*). Within the same article: "'It's obvious they're trying to harm the economy to pressure the government' Ge said, warning of a possible violent response if Wefaq gained cabinet seats in the current climate" ; "warning" and "response" were the words the Automaton pulled out of the search query, but the article is loaded with reactions to threats, including vigilantism, and different threats/demands.
Another query of "reaction to a threat" found "India is particularly keen to strengthen its maritime capabilities (*reaction*), given China's pursuit of a powerful 'blue water' navy which Delhi sees as a threat to key shipping routes in the Indian Ocean and Indian energy assets in the South China Sea" (*the threat*).
The query "impending action" came up with some very interesting hits as well; including "dangerous times", and "...bond auction on Wednesday added to worries that the impact of the European Central Bank's one trillion euro injection of cheap three-year funds into the banking system may be coming (*impending*) to an abrupt halt (*action*)".What I find eerily interesting is that this post is dated Friday the 13th. Coincidence? Time will be the judge of that.
This kind of augmened Sensemaking IS the future for intelligence and it is here now.