Macro-Finance Salon (No. 79) and Fintech Open Classes (No. 13): “AI+” Driving a New Generation of Financial Innovation - Artificial Intelligence Helps Identify the Grey Rhino

2017-12-06 IMI
On December 6, the Fintech Open Classes (No. 13) was held at Renmin University of China. Luo Ping, doctoral supervisor of the Institute of Computing Technology Chinese Academy of Sciences and research fellow of Shanghai Key Laboratory of Intelligent Information Processing, delivered a keynote speech titled “‘AI+’ Driving a New Generation of Financial Innovation - Artificial Intelligence Helps Identify the Grey Rhino”. Gang Jianhua, associate professor in School of Finance at Renmin University chaired the lecture. First, Luo Ping introduced the capital-flow-based financial ecosphere, and pointed out that the financial sector was a document-intensive industry. The ecosphere of capital flows mainly includes three parts--raising funds, allocating funds and generating returns. If the returns are generated, the new funds would circulate in the new fundraising process. And all financial activities are recorded in great amount of various public or non-public financial documents. For example, when the investment banks helped enterprises made initial public offering or issue corporate bonds, they would compose a prospectus of hundreds of pages and submit it for regulatory review. After approved, the prospectus would be publicly disclosed to all primary or secondary market investors. Viewing from the perspective of the consumption chain, investment banks or enterprises would compose the financial document as the seller, financial regulators (PBC, CBRC, CSRC and CIRC) would audit the document as the supervisor, and the investment institutions would read and analyze the document as the buyer. Because the financial document would disclose financial information, it must be true, accurate and complete. Any mistake made in the document, even a slight one hidden in the five or six hundred pages, would lead to great risk. Therefore, we expect the machine to understand some documents and to help financial practitioners to alleviate some of their tedious work. Then Luo Ping introduced AutoDoc--the first big data intelligent financial document product in China. The product can replace labor with artificial intelligence technology and improve efficiency at work. Its core lies in analyzing documents. It processes loads of non-structural documents into structured data, so that the computer can analyze and compare the data. The key problem is to how to make the computer process the natural text language so that the computer can translate it into specific semantic and structured information. There are three characteristics of this transformation process. First, it’s instantaneous. Once a document is published on the Exchange website, it can be structured into the database in 20 seconds. Secondly, it’s precise. The structural language is very precise, yet the human language is quite abundant. The computer needs to master the various expressions in human language, and make them precisely structured. Thirdly, it’s in-depth. We hope that the computer cannot only deal with some simple financial indicators, but also further understand the content of footnotes, in order to develop a thorough understanding of a company.