Saturday, October 5, 2013

Recorded Future

I mentioned this link in the prior post and it captured my interest.  Worth an entry of its own.  The home site of the link here.

There is an elegant beauty in the subject line.  It is the first line of the web site.  Excellent web site design.  Added to recorded future is:  "Creating and Insightful World."

You Beauty!

"Intelligence analysts are under increasing pressure to forecast what the future may hold. Recorded Future intelligence analysis tools help analysts understand trends in big data, and foresee what may happen in the future.
Groundbreaking algorithms extract temporal and predictive signals from unstructured text. Recorded Future organizes this information, delineates results over interactive timelines, visualizes past trends, and maps future events -- all while providing traceability back to sources.
From OSINT to classified data, Recorded Future offers innovative, massively scalable solutions."


This conceptual thought must roll around in my head and to get the feel what it sounds like:  "Recorded Future Creates an Insightful World."  What is the information structure design that supports this?

Past is Prologue?  Yes but with a new twist.  Is "Recorded Future" a paradox?  Conundrum or simply a term for a new Problem Domain that is coming into being as a child of the information age.  Predictive analysis is old school but there are new tools.  Especially in the area of Social Science that used to be sneered at by the Hard Science world.  What is the new term that cleanses "Social Science" its old stigmas?  Intelligence?  After all, it is simply knowing what people will do as a function of known "facts" that rest more on belief systems than anything else.

Maybe I should think more about "We Hold these Truths to be Self Evident"?  Is that a pointer to the Recorded Future?

Ahha!  (Definitely a spell check word) A quality hit on the first google try!

This link:

The content is a tell that a nerve has been touched and and a system designed to introduced smoke and mirror diversion has been activated.  The government job of "smoke and mirror comment maker"  Must be high paid.  It has to be subtle and believable.  Maybe recruit from fundamentalist sysadmins.  Amusing as well as insightful discussion.

This link gives a keyword for storage of closed captioned video:  Closed Caption Asset Management or CCAM.  It talks in detail about the technicalities. It is a blog with a home page of a commercial player in the Closed Captioning and subtitling software market called CPC which was acquired by Telestream.  Little fish eaten by bigger fish.

A DARPA research paper here.  Thin Veil for what it is really all about?


  1. A concrete example will serve to illustrate the reading task framework. In an early Phase 1 reading domain, NFL Scoring, the task is to read news articles about National Football League games and answer queries about final scores, scoring events, and periods played. This domain was selected considering the motivating performance task, “Dynamically infer and display status information regard- ing a sports competition from a play-by-play account,” and the notional application, “For the sports junkie, a set- top plug-in reads the closed captioning stream and main- tains more comprehensive status than the mini-scoreboard overlay typically used in broadcast video, such as a sum- mary of the game’s scoring events.”
    Note that this task engenders inference: Suppose the an- nouncer says, “The kick is good, and it’s 6 to 3,” and sup- pose the only earlier score extracted was “Atlanta Falcons 3, New York Jets 0.” A domain-specific reasoning system (DSRS) can infer that the “kick” must have been a field goal, not a point after touchdown; it may need more in- formation (e.g., a future score) to determine which team is ahead at this point. For simplicity early in the program, we have adjusted this performance task so that we can work with news stories rather than play-by-play tran- scripts: Given any score in an NFL game and any partial information about the contributing scoring events, the DSRS will return the most likely scoring event combina- tions and their probabilities. Ultimately, we target reading systems capable of exploiting such probabilistic informa- tion from DSRS results during the reading process (e.g., to determine the most likely grouping of extracted formal statements into consistent contexts).

Interesting statement on a Fox News TV channel web broadcast:

"IMPORTANT NOTE: WGHP is now complying with the FCC-mandated requirement to caption IP-based internet video streams. This requirement, with a compliance deadline of March 30, 2013, means that we must caption all live video programming. Another deadline to caption non-live programming arrives in September 2013. See this link for more information. We are working to enable the option to turn “on” and “off” closed captioning, however until this can be done, closed captioning will be “on” for all users. "

This is a very interesting analysis


ABSTRACT
We perform an automatic analysis of television news pro- grams, based on the closed captions that accompany them. Specifically, we collect all the news broadcasted in over 140 television channels in the US during a period of six months. We start by segmenting, processing, and annotating the closed captions automatically. Next, we focus on the analy- sis of their linguistic style and on mentions of people using NLP methods. We present a series of key insights about news providers, people in the news, and we discuss the biases that can be uncovered by automatic means. These insights are contrasted by looking at the data from multiple points of view, including qualitative assessment.

This site Nightly News provides CC of a story about NSA spying as well as a written transcript which is supposed to be the CC in transcript form.  With CC turned on the written transcript can also be read (pause as necessary).  The CC and the transcript do not match in basic ways.

Closed Captioning is the law here 

Furthermore:

"Pre-recorded programming that is edited for Internet distribution must be captioned if it is shown on television with captions on or after September 30, 2013."


No comments:

Post a Comment