There is nothing new about facial recognition, it is used by law enforcement and casinos to weed out black-listed players and criminals by matching in-the-flesh humans with data-banked images. But recently, Faception, a Israeli start-up, claims to have developed a program that “understand our personalities” better than ourselves.
The start-up revealed recently signing a contract with a “homeland security agency”. Faception also claims that with this new technology it is possible to detect 15 categories of personalities, including poker players, bingo players and terrorists.
Best of Tells?
According to Shai Gilboa, the company’s chief executive, it is pre-determined by our DNA and through our facial expressions these secrets can be unlocked.
To poker players, this statement doesn’t come as a surprise, but the chief executive believes Faception does a better job detecting personality traits than humans ever could, which might annoy poker players masters, like Daniel Negreanu and Phil Hellmuth.
Both players are entitled a little bragging rights in the arena of tell-reading, Hellmuth won the WSOP bracelet 14 times and Negreanu won it six times, earned around $32.5 million in live tournaments and was introduced to the Poker Hall of Fame in 2014.
However, Faception promises 80% accuracy rate when detecting personality traits using their 15-built-in “classifier”, a daunting technology at a poker table.
At a tournament Faception had organized, the faces of 50 amateur poker players were analysed to prove the software reading abilities.
The software picked 4 players out of the 50 players it thought would achieve the best results at the tournaments. The players were chosen based on how well their facial features matched up facial features of professional poker players in their database. From the four players they picked, two made it to the final three. Gilboa sees this as a successful test.
Even though the results might have been impressive, Pedro Domingos, Professor of the University of Washington and Alexander Todorov, Professor of Princeton University, raised a number of issues regarding Faception data.
According the Professor Domingos, the system could contain outdated information that could distort results. In addition, he believes there are also ethnical issues that could categorize someone as a “murder” (murderer’s personality) erroneously.
Professor Todorov, believes physiognomy is outdated: “Just when we thought that physiognomy ended 100 years ago. Oh well…” he said
Regardless, when it comes to picking out poker talent, Faception seems to have skills.