Virtual Financial Markets at CBCL

We are studying learning in societies of agents, from the point of view of algorithms that optimize agent performance in particular environments, and also of emergent societal behaviors and properties. We believe that learning is the gateway to understanding and reproducing intelligence. Learning is also a key for systems of interacting agents to work: agents have learning abilities that allow them to adapt to the environment, including other agents. Within our long term goal of understanding the mind we are studying artificial markets and hybrid artificial---real markets (markets with human agents as well as artificial agents).

In the last decade there has been a surge of interest within the finance community in describing equity markets through computational agent models. At the same time, financial markets are an important application area within artificial intelligence for the fields of agent-based modeling and machine learning, since agent objectives and interactions tend to be more clearly defined, both practically and mathematically, in these markets than in other areas. Computational modeling of markets allows for the opportunity to push beyond the restrictions of traditional theoretical models of markets through the use of computational power. At the same time, the artificial markets approach allows a fine-grained level of experimental control that is not available in real markets. Thus, data obtained from artificial market experiments can be compared to the predictions of theoretical models and to data from real-world markets, and the level of control allows one to examine precisely which settings and conditions lead to the deviations from theoretical predictions usually seen in the behavior of real markets.

The virtual financial markets setting provides a rich and dynamic testbed for ideas from machine learning and artificial intelligence and simultaneously allows one to draw insights about the behavior of financial markets. The systems we are developing represent a theoretical/computational infrastructure which is ideal for asking and trying to answer a number of questions.

In summary, our ultimate quest to understand intelligence and the learning processes underlying it is leading us to develop artificial financial markets for conducting experiments and simulations in economics. The markets serve as playgrounds for artificial agents, software traders who have learning capabilities, to interact with each other and with human traders. Through the trading game, the artificial agents can learn from experience and optimize their trading strategies.

We are studying learning in societies of agents, from the point of view of algorithms that optimize agent performance in particular environments, and also of emergent societal behaviors and properties. We believe that learning is the gateway to understanding and reproducing intelligence. Learning is also a key for systems of interacting agents to work: agents have learning abilities that allow them to adapt to the environment, including other agents. Within our long term goal of understanding the mind we are studying artificial markets and hybrid artificial---real markets (markets with human agents as well as artificial agents). In the last decade there has been a surge of interest within the finance community in describing equity markets through computational agent models. At the same time, financial markets are an important application area within artificial intelligence for the fields of agent-based modeling and machine learning, since agent objectives and interactions tend to be more clearly defined, both practically and mathematically, in these markets than in other areas. Computational modeling of markets allows for the opportunity to push beyond the restrictions of traditional theoretical models of markets through the use of computational power. At the same time, the artificial markets approach allows a fine-grained level of experimental control that is not available in real markets. Thus, data obtained from artificial market experiments can be compared to the predictions of theoretical models and to data from real-world markets, and the level of control allows one to examine precisely which settings and conditions lead to the deviations from theoretical predictions usually seen in the behavior of real markets. The virtual financial markets setting provides a rich and dynamic testbed for ideas from machine learning and artificial intelligence and simultaneously allows one to draw insights about the behavior of financial markets. The systems we are developing represent a theoretical/computational infrastructure which is ideal for asking and trying to answer a number of questions. Questions about financial markets: What is the role of bounded rationality (agents with limited capabilities)? How do bounded rationality markets compare with the perfect information, fully rational agents of classical equilibrium theories? What is the role of market rules and mechanisms (e.g. clearing, specialists, continous double auctions vs. sealed double auction markets)? Questions about cognitive behavior: What is the effect of certain cognitive limitations (a la Kahneman) on market behavior (the direct problem)? What can we infer from the overall market about individual agents cognitive abilities (the inverse problem)? What is the simplest realistic model of a financial agent? Questions about learning agents and the evolutionary environment they inhabit Development and comparison of specific learning algorithms Role of selection by competition with other agents and from the environment Role of various mutation mechanisms In summary, our ultimate quest to understand intelligence and the learning processes underlying it is leading us to develop artificial financial markets for conducting experiments and simulations in economics. The markets serve as playgrounds for artificial agents, software traders who have learning capabilities, to interact with each other and with human traders. Through the trading game, the artificial agents can learn from experience and optimize their trading strategies.
Projects	Publications
People: Sanmay Das, Adlar Kim, Tomaso Poggio
Back to CBCL Research Back to CBCL Homepage