Every AI investment manager wants to create a fully automated investment trading system that realizes the dream of being able to lie down and make money. AlphaGo's rolling victory in the field of Go has exacerbated this dream. Unfortunately, no such machine has been manufactured so far.
artificial intelligenceThe application in various fields is in the ascendant, and the application in the investment field, especially the secondary market, is attracting people's interest. In the face of the rise of artificial intelligence investment, some people think that the era of artificial intelligence will come, and in the investment field will eventually completely replace people; Then he sneered, thinking that this is just a gimmick. So how do you correctly view the application of artificial intelligence in the investment field? In fact, there is a big difference between Go and the secondary market. In summary, Go is a relatively fixed market with borders, while the secondary market is a dynamic market without boundaries.
First of all, it is meaningless to play chess on the man-machine. The machine is a tool used by people to improve efficiency. It is better to play chess than a man and an airplane. For machines, rich and reliable data is one of the prerequisites for computer programs to make better investment decisions. Test and analyze through massive amounts of data, find out certain rules, write code, and let the machine faithfully execute. The reason why AlphaGo can crush the world of chess is not because of its good technology, but because of its huge server resources and computational efficiency. It can synthesize all the game pieces in the world, all the play of Go, and the way of playing all the characters in the Go game. , play chess by calculation. In the investment field, despite the historical data, there are a lot of noise, some of them.Listed companyIt is also possible to falsify financial data, resulting in data reliability that is far less than the game.
Second, if the machine wants to play a leading role, the rules must be transparent and stable. For example, Go, although there are subtle differences between different countries and regions, it is generally stable, especially in the same game, the rules will not change. . The investment field is different. Whether it is the regulatory policy or the economic environment or the external environment, there will be great changes. For example, in 2016-2018, regardless of the tightening of supervision and mergers and acquisitions, or the recovery of the economic cycle in the economic cycle, In the external environment, the inflow of funds from the north, the market ecology has undergone great changes, and the trading procedures that perform well on historical data are difficult to adapt to this new change.
Finally, artificial intelligence investment decisions rely too much on historical data, while humans' active decision-making advantages are more significant. There is no big unpredictable risk event in the field of Go, but in the investment field, this kind of thing happens from time to time. For these events, there have been no previous incidents. Without historical data to learn, the machine will be helpless, and people can perform systematic analysis. When faced with rare risk events, human decision-making advantages are more significant.
In summary, in the field where the rules are transparent and stable, the historical data is rich and reliable, and no large-scale risk events occur, the machine is not affected by subjective emotions and prejudice, and it is time-sensitive and accurate in the process of receiving, analyzing and making real-time information. Sexuality and consistency are higher than human beings, and relatively more advantageous. In areas where historical data is lacking, rules are opaque, and risk events occur frequently, people's decision-making will be more advantageous. Therefore, in the field of investment, a single machine and a person cannot occupy an absolute advantage. In the future, it should not be a situation in which people and machines oppose each other, but a strong alliance.
At present, artificial intelligence has been applied to research, investment decisions, transaction execution and other aspects.
Research, reading through artificial intelligenceresearch reportAnd company reports can greatly improve research efficiency. For example, the listed company that is currently out of 2018annual reportAnd in the quarterly report of 2019, if every company carefully reads the report and then summarizes it, the artificial coverage is very limited. If you use the machine to read this unstructured text data, you can quickly summarize the key points and extract the core content needed. The researcher interprets the report.
In investment decision-making, artificial intelligence technology is mainly applied to investment decision-making from information processing and knowledge learning. On the one hand, relying on the information processing capability of artificial intelligence, it can efficiently acquire and process unstructured data through artificial intelligence methods, mainly including WeChat data. , search data, Taobao, Jingdong transaction data, etc.; on the other hand, relying on the knowledge learning ability of artificial intelligence, through the artificial intelligence method for asset income forecasting and asset trading.
At the transaction level, the artificial intelligence-built trading strategy is better at finding rules and learning knowledge from complex historical data, and incorporating broader and more complex factors into the analysis of trend predictions to guide future trading decisions; Trading can be significantly improvedInvestment StrategyExecution efficiency, lower impact costs, and to some extent increase the return on the portfolio. Automated trading in the era of artificial intelligence includes automation and intelligence. It emphasizes learning from market data, learning a large amount of historical data, constructing predictive models, optimizing trading algorithms, and obtaining the best trading performance.
In short, machines and people have their own advantages in the field of investment. Machines will not replace people, and people will be inseparable from machines. Human-machine integration may achieve even greater breakthroughs.