Ep. 235: Christian Schwarz on AI Bubbles, Using AI in Trading and Synthetic Data
Sep 27, 2024
auto_awesome
Christian Schwarz, Managing Director and Head of Data Science at Standard Chartered, shares his insights on AI's impact in finance. He questions whether we are in an AI bubble and discusses why credit markets are slower to adopt AI technologies. The conversation also covers the innovative use of generative AI to create synthetic data, enhancing trading strategies while addressing privacy concerns. Schwarz highlights the need for adaptive models and the benefits of long short-term memory (LSTM) models in the fast-evolving financial landscape.
Christian Schwarz highlights the potential of synthetic data generated by AI to enhance trading strategies and overcome data scarcity in credit markets.
He stresses the importance of collaboration between data scientists and traders in finance to effectively integrate AI and improve decision-making processes.
Deep dives
Christian Schwartz's Career Journey
Christian Schwartz shares his background in financial mathematics, detailing how his academic focus on statistics and stochastic processes led to a career in finance. He wrote his diploma thesis on credit portfolio management, which opened doors for him at Credit Suisse in Zurich. His journey took a significant turn during the global financial crisis (GFC), where he adapted by using tools he had developed for portfolio management, which built his credibility on the trading desk. By leveraging game theory, he gained unique insights into the sovereign crisis that allowed him to engage with major hedge funds and expand his understanding of macro markets.
Evaluating the Hype Around AI
Christian discusses the current hype surrounding artificial intelligence (AI) in the financial sector, asserting that while there are impressive capabilities within AI technologies, there exists a degree of uncertainty and exaggerated expectations. He suggests there’s a 10 to 20 percent chance of an 'AI winter' due to the difficulty of maintaining high commercial impact amid rising expectations. Schwartz points out that traditional applications of AI, such as fraud detection and image recognition, have often been reclassified as standard techniques as they become more commonplace. The shift in the yardstick for what constitutes AI complicates the assessment of its commercial value and adoption, especially in regulated institutions.
Synthetic Market Data Generation
Addressing challenges in the credit market that lag behind others like equities and FX in incorporating AI, Christian proposes a solution of generating synthetic credit market data using a unique AI model. This model captures the probabilistic distributions and stochastic processes of historical data, allowing it to simulate realistic market conditions that have never occurred. By implementing a Generative AI approach similar to systems like ChatGPT, the model generates synthetic time series data that can be utilized for trading strategy development, risk management, and scenario analysis. This advancement is crucial as it enables finance professionals to work with more abundant and diverse datasets, overcoming the issues of data scarcity and ensuring more comprehensive market analysis.
The Importance of Collaboration in Finance
Christian emphasizes the significance of collaboration between data scientists and traders in finance for successfully integrating AI and machine learning into trading practices. Instead of presenting AI as a replacement for traders, he advocates for partnership, whereby data scientists can help traders with existing problems through simple yet effective solutions. By addressing mundane tasks and automating processes, AI can free traders to focus on more strategic decision-making. This collaborative approach not only enhances productivity but also fosters trust and mutual understanding, ultimately benefiting both the trading teams and the clients they serve.
Christian is Managing Director and Head of Data Science in the financial markets division at Standard Chartered. He focuses on credit and previously worked at Mizuho and Credit Suisse. His postgraduate studies were in financial mathematics. In this podcast we discuss whether we are in an AI bubble, why credit markets have lagged other markets in adopting AI, using generative AI to create synthetic data, and much more.