Roman Polanski and Sentiments

Polanski in the news

Recently heard that Roman Polanski was captured and there has been lively debate on what to do with him.

I thought this could be an interesting case to see if there is any patterns on sentiments in the digital world. Unlike the Sonia Sotomayor analysis though, this is just a single slice of time look.

In the top 30 terms in the three major web results, most of them seems to be factual with a few negative terms such as "avoided."

(Correction: I originally used "repulsion" as an example of negative term, turns out that it is the title of one of Polanski's films and considered a milestone in the psychological thriller genre. My bad!)

More on the results

A reminder, if you are interested in finding out more about the result details or the inference engine used in this analysis, please let me know at yfjuan (at) parc (dot) com.

Results (top 30 terms ranked in the order of relevance with the most relevant on top)

Google	Yahoo	Microsoft
polanski saleby brach escroqueries extradite 8/18/1933 cleland zebiczny pianiste sexual quitters seigner extradition 2000s eatps.web.aol.com wife rozbijemy filming gets chris letterman arrest murderer hour directing revives diamants swindles began musician	polanski travelled directing extradition filming plotted arrest seigner revives murdered roman pleaded arrested dated credited fleeing letterman filmmaker zebiczny repulsion personals directors celebrated avoided meets zurich dwts favre fired collaborated	polanski extradition obama twitter sotomayor gmail geimer facebook zebiczny bankrupt filmography 8/18/1933 getty looting seigner tmz murderer rozbijemy msnbc extradite vick credited repulsion qaida drama screenwriter afghan feeds ancestor swindlers

===
P@P

Inference Engine and Applications

Inference Engine

One of the things that internet/web enables is to make low-cost data collection and analysis mostly a computational question. Add in the idea of "wisdom of crowd", then you have the making of PARC's inference engine technology - if you would allow for a gross simplification.

What the inference engine is good at is to identify items that may be linked in ways that is not always obvious. My favorite example is how the inference engine easily cracked the heavily redacted story on the location of a CIA agent's first assignment in How to catch a spy.

Web Chatters

But the same technique can also be used to identify emerging trends (chatters) on a specific topic. So, I used the Inference Engine to look at the top terms associated with the Sonia Sotomayor's nomination process to the US Supreme Court at Sotomayor Web Chatter Index.

And, indeed, it is interesting to see what are the top ranking terms associated with that process and how it points to specific concepts that come in and out of favor in the web crowd.

Private Corpus

Finally, it may be worth noting that the same technique can also be applied to data corpus beyond the visible internet. For example, eDiscovery in legal proceeding is a space where this capability would be very helpful by expanding beyond simple keyword search to a more nuanced understanding of how information linked to each other.

===
P@P

Sotomayor Web Chatter Index - Week 12

New items and faded topics

The term "111th" which refers to the fact that Justice Sotomayor is the 111th for the Supreme Court, continued its appearance in the top 20.

For the first time, "Malesko" and "Alito" both showed up in all three rankings. Would this presage the basis for the first analysis of the Supreme Court once she is sworn in and decisions are rendered?

More on the results

A reminder, if you are interested in finding out more about the result details or the inference engine used in this analysis, please let me know at yfjuan (at) parc (dot) com.

Results (top 20 terms ranked in the order of relevance with the most relevant on top)

Google	Yahoo	Microsoft
sonija leval sotomyaor judgeship rulings 111th stating sotomayor coxe gurfein gop malesko judging 68-31 chin ginsburg ruled eight legislator alito	sotomyaor sotomayor sonia souter malesko mcmahon destefano bronxdale alito ginsburg hispanic scotus nominee scotusblog supreme scalia hearings morgenthau 111th allowing	sotomayor alito scotus moynihan obama statehood souter 111th rulings destefano retiring scalia stating legislator ginsburg malesko cabranes scotusblog mahoney rendell

===
P@P

Sotomayor Web Chatter Index - Week 11

Confirmed to the US Supreme Court

By a vote of 68-31, Sonia Sotomayor was confirmed as a Supreme Court justice on August 6th, 2009.

New items and faded topics

All the usual suspects such as "nlj (National Law Journal)", Foxnews, and "wsj.com" dropped off the top 20 list.

The new term that is making a rapid ascend is the term "111th" which, in case you missed the news, Justice Sotomayor is the 111th for the supreme court.

Next week, we will have the final installment on the top 20 analyses.

More on the results

A reminder, if you are interested in finding out more about the result details or the inference engine used in this analysis, please let me know at yfjuan (at) parc (dot) com.

Results (top 20 terms ranked in the order of relevance with the most relevant on top)

Google	Yahoo	Microsoft
sotomyaor stating sotomayor rulings sonia faced sued assail baer gop coxe 111th gurfein meant approval involving judging â nominees murders	sotomyaor sotomayor sonia souter malesko judgepedia mcmahon destefano bronxdale alito scotus ginsburg scotusblog nominee hispanic supreme judgeship hearings morgenthau tasini	alito sotomayor obamaâ scotus moynihan obama statehood 111th souter scotusblog retiring destefano malesko legislator politico cabranes noting mahoney stating judgeship

===
P@P

Sotomayor Web Chatter Index - Week 10

Senate Confirmation Debate

Now that the Senate Judiciary Committee has endorsed the Supreme Court nomination of Judge Sonia Sotomayor, the nomination is now in the full Senate for debate.

New items and faded topics

The term "nlj (National Law Journal)" and its sister term "pubarticlenlj.jsp" have kept up its ranking. As a matter of a fact, it shot up to #1 in one of the engine results.

Foxnews has quickly lost its top 20 ranking while "wsj.com" continues to assert its relevance to this discussion. Also appearing for the first time is the term "Barack" and debuting as #1 no less. Similarly, "conservatives" made its top 20 appearance.

Finally, the term "dissented" which appeared in the top 20 last week kept its position this week.

More on the results

A reminder, if you are interested in finding out more about the result details or the inference engine used in this analysis, please let me know at yfjuan (at) parc (dot) com.

Results (top 20 terms ranked in the order of relevance with the most relevant on top)

Google	Yahoo	Microsoft
justices stating cityâ rulings dissented sotomayor sotomayorâ questioning proceedings defenses nominees malesko krimstock â supporting issued ended nearly senator supreme	nlj sotomayor sotomayorâ krimstock sonia malesko souter pubarticlenlj.jsp destefano ginsburg scotusblog nominee hispanic judgeship supreme judgepedia hostednews jury hearings leahy	barack sotomayor moynihan statehood obama leahy souter cuomo destefano cabranes scotusblog online.wsj.com ginsburg conservatives affirmative prosecutor stating nlj rulings politico

===
P@P

Sotomayor Web Chatter Index - Week 9

Senate Judiciary Committee Confirmation

The Senate Judiciary Committee voted, 13 to 6, today to endorse the Supreme Court nomination of Judge Sonia Sotomayor

New items and faded topics

The term "nlj" which refers to "National Law Journal", has kept up its ranking and added a related term "pubarticlenlj.jsp" "judgepedia" - a wiki on judges - kept its place.

Foxnews has come back with a vagence as #3 on Google. On the other hand, Washingtonpost.com disappeared from the top 20 ranking. "wsj.com" continues to do well losing one notch, edged out by "Obama" on Microsoft ranking.

Also interesting is to see the term "dissented" in the top 20. We will see if it would stay next week.

More on the results

A reminder, if you are interested in finding out more about the result details or the inference engine used in this analysis, please let me know at yfjuan (at) parc (dot) com.

Results (top 20 terms ranked in the order of relevance with the most relevant on top)

Google	Yahoo	Microsoft
justices www.foxnews.com sotomayorâ rulings cityâ dissented judiciary defenses sotomayor courts americans souter malesko hearings â senators circumstances krimstock issued reversal	sonia sotomayorâ sotomayor souter destefano krimstock nominee ginsburg nlj hearings morgenthau judgepedia tasini leahy judiciary malesko cuomo judgeship supreme justices	sotomayor statehood moynihan obama leahy online.wsj.com nlj souter destefano cuomo stating rulings ginsburg cabranes republicans prosecutor affirmative pubarticlenlj.jsp politico gained

===
P@P

Sotomayor Web Chatter Index - Week 8

Confirmation Hearing

With the confirmation hearing done, it is interesting to see how the discussion has shifted into some factual and some "speculative" terms.

New items and faded topics

"judgepedia" - a wiki on judges - kept its place. It is also joined by the term "nlj" which refers to "National Law Journal"

Foxnews has completely dropped off from the top 20 rankings after being highly ranked for most weeks. Interesting to see Washingtonpost.com making a come back - I wonder if it is a reflection that now the denizens inside the Washington DC Beltway are doing most of the chattering. As for wsj.com's come back, do you have any thoughts?

Also interesting is to see the term "unanimous" in the top 20 which does seem to reflect the mainstream media's consensus on her confirmation.

More on the results

A reminder, if you are interested in finding out more about the result details or the inference engine used in this analysis, please let me know at yfjuan (at) parc (dot) com.

Results (top 20 terms ranked in the order of relevance with the most relevant on top)

Google	Yahoo	Microsoft
justices repeatedly stating americans sotomayor questioning judiciary dissented reversing helped defenses lawyers nyt â cityâ souter www.washingtonpost.com saying became unconstitutional	sotomayorâ sotomayor sonia souter krimstock ginsburg nominee judgepedia nlj morgenthau tasini cuomo hearings leahy judgeship supreme justices judiciary republicans hispanic	sotomayor moynihan statehood online.wsj.com obama nlj souter banc cuomo leahy rulings politico gop stating reversals ginsburg prosecutor unanimous affirmative cabranes

===
P@P

Sotomayor Web Chatter Index - Week 7

Confirmation Hearing

Now that the confirmation hearing for Ms Sonia Sotomayor's nomination to the US Supreme Court has started, I am curious to see how the chatter profile would change.

New items and faded topics

"Litigator" did not have the staying power to remain in the top 20. "Firefighters" is also waning. The new entrant to the top ranking is "judgepedia" - a wiki on judges, did not know one existed.

Foxnews has dropped in rankings after being highly ranked for most weeks. Would it come a come back now that the confirmation hearing is now?

More on the results

A reminder, if you are interested in finding out more about the result details or the inference engine used in this analysis, please let me know at yfjuan (at) parc (dot) com.

Results (top 20 terms ranked in the order of relevance with the most relevant on top)

Google	Yahoo	Microsoft
senators esquire judiciary www.foxnews.com jury sotomayorâ reversing senator ruled dissented courts sotomayor reversal sued rulings nominated attorney married belief securities	sotomayorâ sotomayor sonia souter gillibrand krimstock ginsburg schumer judgepedia nominee cuomo judgeship justices morgenthau supreme leahy cabranes firefighters www.foxnews.com republicans	sotomayor gillibrand obama schumer moynihan souter cuomo rulings centrist reversals politico gop republicans leahy ginsburg firefighter prosecutor mahoney cabranes affirmative

===
P@P

Sotomayor Web Chatter Index - Week 7

New items and faded topics

Firefighters made the top 20 list this week. Litigator also made the top 20. Two timely entries on the current discussion. We will see how long they last.

As noted last week, "racist" made an appearance. It disappeared this week and "centrist" actually made its appearance. Gingrich completely dropped off in all three results. Also interesting to note that Foxnews dropped significantly in one of the rankings.

More on the results

A reminder, if you are interested in finding out more about the result details or the inference engine used in this analysis, please let me know at yfjuan (at) parc (dot) com.

Results (top 20 terms ranked in the order of relevance with the most relevant on top)

Google	Yahoo	Microsoft
rulings www.foxnews.com dissented sotomayorâ sotomayor souter litigator questioning nominees krimstock proceedings republicans defenses â cityâ ended res judges reversing efforts	sotomayor sotomayorâ sonia souter krimstock schumer ginsburg nominee cuomo judgeship justices supreme cabranes judgepedia biden firefighters hispanic republicans www.foxnews.com leahy	sotomayor schumer obama moynihan souter cuomo rulings gop centrist republicans politico reversals cabranes ginsburg firefighters prosecutor leahy affirmative litigator nominees

===
P@P

Sotomayor Web Chatter Index - Week 6

New items and faded topics

Jarring but interesting to see that "racist" made the top 20 this week. Palin also vaulted to into the top 20 this week. Letterman made a come back to the top 20.

As noted last week, CNN indeed dropped off the list as suspected. Nightline probably did a program on this topic so it made the top 20, we will see if it stays next week. And, finally, instead of making a run for the top spot that Foxnews occupies, Gingrich lost ranking in all three results. It may be time for him to say a few new things.

More on the results

A reminder, if you are interested in finding out more about the result details or the inference engine used in this analysis, please let me know at yfjuan (at) parc (dot) com.

Results (top 20 terms ranked in the order of relevance with the most relevant on top)

Google	Yahoo	Microsoft
sotomayorhtml www.foxnews.com justices sotomayorâ courts nominees facility concerns serving issued cityâ ended ricans served moved rican hearings filed krimstock sotomayor	www.foxnews.com sotomayor sotomayorâ sonia souter online.wsj.com scotus scotusblog gillibrand gingrich nominee judgeship schumer ginsburg supreme justices krimstock nomination hispanic racist	scotus palin gillibrand gingrich schumer obama scotusblog justices gop souter rulings politico republicans nominee cuomo blotter sotomayor nightline letterman unanimous

===
P@P

How to catch a spy

Source: Plame vs. Whitehouse

For those of you who do not remember, Valerie Plame was working for CIA as an undercover agent and the Whitehouse leaked her CIA identify in 2003. With her cover identify blown, she left CIA in 2005.

In 2007, she published a memoir "Fair Game: My Life as a Spy, My Betrayal by the White House". CIA intervened and redacted (blacked out) "sensitive" information in the published book.

A page of the redacted Fair Game


How to catch a spy, the PARC way

A PARC team has developed a machine learning engine that is able to use contextual information that may not be sensitive by itself but in aggregate provides strong inference on what the missing information should be.

The Plame book is a perfect test case because, although the book has been redacted, the actual information is available in other public sources. In other words, we can run the book through the engine and see what kind of inference the engine can tell us and check it against the known answers.

Test case: where was her first assignment?

So, we fed the available and seemingly innocuous description on the location (redacted) of her first assignment such as "Europe, chaotic, outdoor café, traffic, summer heat" into the software.

Lo and behold, the engine comes back with Greece as the most probable answer which was indeed the case.

--
How would you use this software engine beyond figuring out if your censors are good enough? Conversely, how would you use the output of this engine? How about removing sensitive medical information in unstructured format? Or, finding that smoking gun in the mountain of data and emails in a legal case? This is an instance where tireless software with perfect memory to a large corpus of information is a better solution than the best trained/paid human attention any day.

Let me know how you would use this capability. For the most interesting idea(s), maybe I can get you a copy of the software engine to play with.

Look forward to hearing from you.

===
P@P