[cryptome] Re: Silent Talker

  • From: Shaun O'Connor <capricorn8159@xxxxxxxxx>
  • To: cryptome@xxxxxxxxxxxxx
  • Date: Fri, 13 Feb 2015 12:43:39 +0000

Fine..until the dataset it relies upon gets hacked. all it needs is for
the use case to be modified and the results can be distorted. hence the
rather shallow news item ( no information on how the system interprets
the micro movements  for example) but then I suppose since a patent is
in the works they can't talk about that can they.
I like the name its been given though. silent talker.
also while 80% accuracy is arguably better than most methods currently
in use the error margin is enough to raise reasonable doubt ( unless the
legal definition of"reasonable doubt" is conveniently watered down to
the point of hearsay standard  or more accurately defined "on the
balance of probability"as is the case in civil litigation)
On 13/02/2015 11:09, doug wrote:
> latest news...see url:http://www.mmu.ac.uk/news/articles/77/
> On 13/02/15 00:07, doug wrote:
>> see url: http://en.wikipedia.org/wiki/Silent_Talker_Lie_Detector
>> Will it replace the polygraph...I wonder...No more Clappers and
>> Edward Snowdens...:-) . That will be the day...The wonders of new
>> technology...
>> ATB
>> Dougie.


Other related posts: