After defeating a Japanese chess expert, the next task of this computer program is to find out if you can make the payments of a new mortgage.
Heroz Inc., a small Japanese company whose engineers designed computer systems that were the first to defeat an active professional shogi player, the Japanese version of chess, says it is now working on adapting its applications to the financial industry.
While Deep Blue became the first computer to beat a top chess player in 1997, the computer programming required to defeat a shogi master is even more complex than that required for conventional chess. This is mainly due to the fact that the amount of potential movements in Japanese chess is much greater, since the rules of shogi on the reuse of captured pieces are more complicated.
Heroz hopes that the lessons he has learned on how to recreate human judgment to help his computers win in shogi can be applied to the crunching of data for banks when deciding whether clients are creditworthy, according to Heroz CFO Daisuke Asahara . “There are times when computers can see that what humans perceive is wrong,” said Asahara, a former investment banker at Goldman Sachs Group Inc., in an interview last month.
He said the evidence shows that Heroz computers can successfully reduce the huge amount of data on consumer deposit and withdrawal information, and social media, to help banks when they are making decisions to grant personal loans. The company is also working with Nomura Holdings Inc. to investigate whether its technology can be applied to financial market forecasts, Asahara said.
The Deep Blue computer developed by International Business Machines Corp. became the first computer to beat a top chess player, when it defeated then world champion Garry Kasparov in 1997. The added complexity of shogi meant that it took another 16 years before that the computer program Heroz capable of outwitting an active shogi professional, with his 2013 victory over Shinichi Sato.
These advanced computers can apply a kind of judgment on which of the many trillions of possible chess moves they should perform, through an automatic learning process based on programming and past game data, according to Ho-fung Leung, professor of the. Chinese University of Hong Kong that specializes in artificial intelligence. According to Leung, the process avoids the need to painstakingly examine all possible chess moves.
“The approach they are taking is very different from the conventional approach of the tree of possibilities for computer chess,” Leung said.
There are about 10 in power of 120 possible moves in a conventional chess game, but that is overshadowed by 10 and the power of 220 possible moves, the number one followed by 220 zeros, in a shogi game, according to Heroz. website.
Asahara would not disclose the names of the banks Heroz works with when analyzing consumer credit data. But Japan’s Financial Services Agency is trying to make banks more active in the field, saying in December that it will recommend legal changes to streamline the use of advanced technology in finance.
Japan’s largest lenders, including Mizuho Financial Group Inc., are already developing services that use artificial intelligence technology, such as IBM’s Watson computer, mainly in call center automation, according to Mizuho spokesperson Masako Shiono.
Heroz was founded in 2009 by Takahiro Hayashi and Tomohiro Takahashi, both former employees of NEC Corp. of Japan. The company of 70 employees develops and commercializes shogi and chess games for smartphones, which are the source of data for the development of their automatic learning techniques.
Its advance in the financial analysis will be assisted by a sale of 100 million yen ($ 837,000) of Heroz’s equity to the Hifumi Hatchery Fund on December 28, said Asahara. He said that Heroz will use funds from Hifumi, a Japanese startup fund, to invest in facilities such as cloud servers and hire more computer engineers.