A new page on Google Classroom is available for this year: classroom code hjq3qdg.

Artificial intelligence for trading:

the impact of artificial intelligence in interdependent financial market networks

Teacher: Novella Bartolini

This is either an Honor Program or a Subsidiary Formative Activity (in Italian AFC, Attività Formativa Complementare). 

It  addresses pioneering research issues in the field of artificial intelligence for financial markets. Students will learn how to design AI based trading strategies and algorithms, and will learn how to assess their performance and impact. They will also learn how to analyze and use historical market data to perform realistic market simulations.

The AFC is highly research based. Prerequisites include familiarity in statistics and probability theory, linear algebra and calculus, strong algorithmic background, and Python programming.

The course will enable unique opportunities for the students to interact with experts and work on real problems and with real data.

Course outline

1. Introduction to Finance and Trading:

a. Financial markets, participants and structure,

b. Exchanges, stocks, orders, and traders.

2. Computing and network architectures for high performance trading.

3. Basic concepts of algorithmic trading, AI based strategies.

4. Network and multi-agent models of financial markets.

5. Interdependencies across financial markets:

a. Dependencies induced by portfolio rebalancing,

b. Material interdependencies among heterogeneous assets,

c. Exchange Traded Funds (ETF) dependencies,

d. Interdependencies caused by financial arbitrage,

e. Behavioral interdependencies (herd behavior, panic, news).

6. High-Frequency Trading (HFT) and dark pools.

7. Information asymmetry under legitimate algorithmic trading.

8. Risk assessment, and perturbation of risk propagation across interdependent market networks.

9. AI bias in algorithmic trading.

10. Fairness in financial markets and related metrics. Propagation of unfairness.

11. Flash crashes in financial market networks.

12. Multi-agent simulation applied to financial markets.

13. Synthetic market data generation. Trace driven market simulations with LOBSTER data (https://lobsterdata.com/) on ABIDES (in Python).

Readings

Books on algorithmic trading

1. Irene Aldridge, High-Frequency Trading: A Practical Guide to Algorithmic Strategies and Trading Systems, Wiley, 2010

2. Alvaro Cartea et al., Algorithmic and High Frequency Trading, Cambridge University Press, 2015

3. Ionut Florescu, Maria Mariani, H. Eugene Stanley, Frederi G. Viens, Handbook of High-Frequency Trading and Modeling in Finance, Wiley, 2016

Research papers of interest to the course

4. Johan Walden, Trading, Profits, and Volatility in a Dynamic Information Network Model, Review of Economic Studies, Vol. 86, 2019

5. David Byrd, Maria Hybinette, Tucker Hybinette Balch, ABIDES: Towards High-Fidelity Multi-Agent Market Simulation, Proceedings of the 2020 ACM SIGSIM Conference on Principles of Advanced Discrete Simulation (ACM SIGSIM-PADS '20)

6. M. Shearer, D. Byrd, T. H. Balch, Towards Explaining Exchange Traded Funds’ Impact on Market Volatility Using an Agent-based Model, Proceedings of the 33rd Conference on Neural Information Processing Systems (NeurIPS 2019)

Background books

7. Yves Hilpisch, Python for Finance, O’Reilly, 2014

8. Tom Mitchell, Machine Learning, McGraw Hill, 1997

9. Kevin P. Murphy, Machine Learning: a Probabilistic Perspective, 2012

Evaluation

The final evaluation is based on a project and a related discussion.