[Korea Selected financial newspaper reporters] can be utilized to prevent the field of deep learning (Deep Learning) is a non-face-to-face over banking transactions era of artificial intelligence (AI).
According to the financial services industry 16, Shinhan Bank, last February, over financial transactions Detection System (FDS: Fraud Detective System) 'financial first apply deep learning technology operates by introducing a. Deep learning is a technique for processing large amounts of data through artificial neural networks and machine learning new patterns themselves.
Shinhan Bank has collaboration with the Shinhan Financial Futures fins Tech enterprise development programs seuraep 03 companies' InfiniBand trees. Add this deep learning-based detection solutions over financial transactions 'GruDEEP' the company has developed into an existing system implements a detection system that combined the existing trade patterns Model and deep learning model. Deep learning based on the new system introduced at the time of Shinhan Bank has said boyeotdago the pre-test results of more than 50% hit rate than traditional sophisticated fraud detection systems.
Shinhan Bank said, "When operated approximately six months is reported to have also obtained meaningful statistics than hit involvement of the new system," said "through the deep learning can prepare the fraud that increasingly intelligent grasp over financial transactions, new forms of electronic banking is expected to help prevent an accident, "he said.
Deep learning based FDS is expected to gradually spread throughout the financial sector. As a non-face-to-face financial transactions increased significantly because the increasing importance of such processing, detection and analysis of large amounts of data. Existing FDS is because I hold longer deal with the scenarios and rules (rule) input in advance is limited to preventing fraud is exposed to constantly changing. When you attack a roundabout may be vulnerable.
Whereas deep learning based FDS is to capture in real time when it detects abnormal symptoms while learning the unusual financial transactions repeatedly kept in mind the multiple patterns. Finding a specific pattern in the big data is count to reach the foreseeable future. As you have accumulated a pattern fault detection, or may work just past the membrane. The first online payment service created by the American company FDS PayPal (Paypal) is also being utilized to upgrade the FDS based on a deep learning.
In Korea, followed by Shinhan Bank, Agricultural Bank of NH, another local bank BNK jungeseon Busan Bank is considering the introduction of FDS based on deep learning.
Pusan Bank, said, "In the future, will further refine the rules pattern and further develop the capabilities," said "scams evolve to in the legislation permitted range FDS quickly organized themselves and introducing means machine-learning techniques to deliver results that analysis techniques and he said at least we will respond to transaction ".
According to the banks, deep learning based FDS advanced propulsion Trend "of the KFTC, future financial industry costs the development of artificial intelligence techniques including deep learning and machine learning, risk reduction, increased productivity, new business models development the positive effects are expected.
Gimjeongsang KFTC Senior Research Station are important elements of "pin-Tech activated and the Internet are expected to increase transactions non-face-to-face to full-scale launch of such a specialized bank" and "FDS over transactions detection capability of determining the customer confidence in the banking company predicted would be a "
Daily Economic News Finance FNTIMES - commercial reprint, copy, distribute prohibited purposes under the Copyright Act
Copyright ⓒ Korea's Financial News & FNTIMES.com
