Video Type: Conference

Video Description

A Talk from Annual Conference 2024

Market generators are a rapidly evolving class of models that simulate financial market behavior using neural networks. These deep learning models learn the underlying distribution of financial data and generate synthetic market scenarios, offering a powerful alternative to classical stochastic models. The expression “Market Generator” only entered the vocabulary of financial modelling around 2019, but by today it has already grown into an area of its own right, with a rapidly growing number of research contributions appearing on the matter. This talk aims to unravel the evolution of the rapidly evolving landscape of generative models in Machine Learning in the context of more classical modelling techniques in finance for the evolution of stock prices but also in the context of the most recent trends of Generative AI that include LLMs.

More concretely we will present a very powerful generative model for time-series modelling: State-of-the-art performance for irregular time series generation has been previously obtained by training Neural SDEs adversarially as GANs. However typical for GAN architectures suffer from various instabilities in training. Instead, we introduce a novel class of generative models that can be trained non-adversarially, using scoring rules on pathspace based on signature kernels. Most notably this procedure permits to evaluate a corpus of paths against a single observation. But additionally, our procedure permits conditioning on a rich variety of market conditions and significantly outperforms alternative ways of training Neural SDEs on a variety of tasks including the simulation of rough volatility models, the conditional probabilistic forecasts of real-world forex pairs where the conditioning variable is an observed past trajectory. Moreover, our framework can be easily extended to generate spatiotemporal data, including the mesh-free generation of limit order book dynamics.

Speaker Bio

Blanka Horvath

Blanka Horvath is an Associate Professor of the University of Oxford, researcher at the Oxford Man Institute and a Member of the DataSig group affiliated with the Alan Turing Institute and Emmy Noether Fellow of the London Mathematical Society. Prior to her current role, she held faculty positions at the Technical University of Munich and at King’s College London and postdoctoral positions at Imperial College London and ETH Zurich.

Blanka’s research interests in mathematical finance include stochastic volatility modelling with a particular focus on rough volatility models, and more recently, generative models with a special interest for data representation via rough path signatures. Her research has a pronounced emphasis on practical applicability, which was recognised by the 2020 Rising Star Award of Risk Magazine. Several of her research contributions have been implemented by various entities in the financial industry including solutions to pricing and calibration of rough volatility models and to the development of appropriate market generators in the deep hedging paradigm.

Video Description

A Talk from Annual Conference 2024

Career advice. From me?!; Investors = Bosses; My experience; Behavioural Finance 101; The simplest model for Happiness; A few results.

Speaker Bio

Dr. Paul Wilmott

Paul is the founder of the Certificate in Quantitative Finance and Wilmott.com and he is internationally renowned as a leading expert on quantitative finance. His research work is extensive, with more than 100 articles in leading mathematical and finance journals, as well as several internationally acclaimed books on mathematical modeling and derivatives, including the best-selling Paul Wilmott On Quantitative Finance, published by John Wiley & Sons.

Video Description

A Talk from Annual Conference 2024

This talk will explore how certain aspects of the green shift may potentially lead to “green inflation,” driven by money printing and investment in less effective energy production methods, especially in an environment where populism is prevalent. Additionally, we will briefly discuss our views on quantum computers, including quantum gravity computers, and their potential implications for finance and the society.

Speaker Bios

Professor Espen Gaarder Haug

Espen Gaarder Haug has accumulated over 30 years of experience in derivatives trading and research. He served as a proprietary option trader at J.P. Morgan Chase in New York and as a trader for the hedge fund Paloma Partners. Dr. Haug’s prolific publications span journals such as Quantitative Finance, the International Journal of Theoretical and Applied Finance, and Wilmott Magazine.

He currently holds the position of Professor of finance at the Norwegian University of Life Science. Despite being professor in finance he has over the past decade primarily dedicated his efforts to research in the field of physics, particularly focusing on developing a full quantum gravity theory. Professor Haug has also authored numerous articles at the intersection of finance and physics, highlighting the connection between these two disciplines, his talk will be along some of these lines.

Video Description

A Talk from Annual Conference 2024

The panel discusses the importance of real options in the energy ecosystem in the context of energy transition with the focus on battery storage and other critical new technologies. We will debate their potential implications on prices and volatility across different energy markets.

Speaker Bios

Dr. Ilia Bouchouev

Dr. Ilia Bouchouev is the former President of Koch Global Partners where he launched and managed global derivatives trading business for over 20 years. Over the years, he introduced several energy derivatives products and was recognized as one of the pioneers in energy options trading. He is currently a managing partner at Pentathlon Investments and an adjunct professor at New York University, where he teaches energy trading at Courant Institute of Mathematical Sciences. He is also a senior research fellow with Oxford Institute for Energy Studies.

Professor Doyne Farmer

J. Doyne Farmer is Director of the Complexity Economics programme at the Institute for New Economic Thinking and Baillie Gifford Professor of Complex Systems Science at the Smith School of Enterprise and the Environment, University of Oxford. He is also External Professor at the Santa Fe Institute and Chief Scientist at Macrocosm.

His current research is in economics, including agent-based modeling, financial instability and technological progress. He was a founder of Prediction Company, a quantitative automated trading firm that was sold to UBS in 2006. His past research includes complex systems, dynamical systems theory, time series analysis and theoretical biology. His book, Making Sense of Chaos: A Better Economics for a Better World, was published in 2024.

During the 1980s he was an Oppenheimer Fellow and the founder of the Complex Systems Group at Los Alamos National Laboratory. While a graduate student in the 1970s he built the first wearable digital computer, which was successfully used to predict the game of roulette.

Professor Espen Gaarder Haug

Espen Gaarder Haug has accumulated over 30 years of experience in derivatives trading and research. He served as a proprietary option trader at J.P. Morgan Chase in New York and as a trader for the hedge fund Paloma Partners. Dr. Haug’s prolific publications span journals such as Quantitative Finance, the International Journal of Theoretical and Applied Finance, and Wilmott Magazine.

He currently holds the position of Professor of finance at the Norwegian University of Life Science. Despite being professor in finance he has over the past decade primarily dedicated his efforts to research in the field of physics, particularly focusing on developing a full quantum gravity theory. Professor Haug has also authored numerous articles at the intersection of finance and physics, highlighting the connection between these two disciplines, his talk will be along some of these lines.

Video Description

A Talk from Annual Conference 2024

We develop a model based on time changed Lévy processes and study its ability of reproducing the joint S&P500/VIX implied volatility smiles and the VIX futures prices – a problem known in the literature as the ‘joint calibration problem’. The model admits semi-analytical characteristic functions for the key quantities, and therefore efficient Fourier based pricing schemes can be deployed. We focus on a specification of the proposed general setting which uses purely discontinuous processes. Results from the application to market data show satisfactory performances in solving the joint calibration problem, and therefore demonstrate that the class of affine processes can provide a workable fit. (Joint work with Ernst Eberlein and Grégory Rayée).

Speaker Bio

Professor Laura Ballotta

Professor Ballotta works in the areas of quantitative finance and risk management. She has written on topics including stochastic modelling for financial valuation and risk management, numerical methods aimed at supporting financial applications, and the interplay between finance and insurance. Recent major contributions have appeared in Journal of Financial and Quantitative Analysis, European Journal of Operational Research and Quantitative Finance among others. A former pole-vaulter, she is passionate of track and field.

Video Description

A Talk from Annual Conference 2024

Twenty five years ago, several authors proposed to model the forward interest rate curve as an elastic string along which idiosyncratic shocks propagate, accounting for the peculiar structure of the return correlation across different maturities. In this paper, we revisit the specific “stiff” elastic string field theory of Baaquie-Bouchaud 2004 in a way that makes its micro-foundation more transparent. Our model can be interpreted as capturing the effect of market forces that set the rates of nearby tenors in a self-referential fashion. The model is parsimonious and accurately reproduces the whole correlation structure of the FRC over the time period 1994-2023, with an error below 2 \%$. The dependence of correlation on time resolution (also called the Epps effect) is also faithfully reproduced within the model and leads to a cross-tenor information propagation time of a few tens of minutes. Finally, we confirm that the perceived time in interest rate markets is a strongly sub-linear function of real time. In fact, our results are fully compatible with hyperbolic discounting, in line with the recent behavioural Finance literature.

Speaker Bio

Jean-Philippe Bouchaud

Jean-Philippe Bouchaud is a pioneer in Econophysics. He co-founded the company Science & Finance in 1994, which merged with Capital Fund Management (CFM) in 2000. In 2018 he was appointed as an adjunct Professor at Ecole Normale Supérieure, where he teaches a course on complex systems. He is also a member of the French Académie des Sciences. His work includes the physics of disordered and glassy systems, granular materials, the statistics of price formation, stock market fluctuations, and agent based models for financial markets and for macroeconomics. He was awarded the CNRS Silver Medal in 1995, the Risk Quant of the Year Award in 2017, the Lars Onsager prize in 2024. He is the author of several books in physics and finance, including “Trades, Quotes & Prices” (CUP, 2018).

Video Description

A Talk from Annual Conference 2024

Polymodels are a statistical analysis technique for dynamic objects evolving within an environment, which is dynamic as well. Both the object and the environment are described by time series. In a financial context, the object is an asset, a fund, a portfolio, anything that can represent an investment, while the environment is made of variables or “factors” that describe the state of the market. Polymodels provide the individual response of the object to every single variable of the environment, together with a reliability score. These response functions being nonlinear, the fragility/antifragility properties of the object can be evaluated.

We will show how to estimate polymodels and to use them for asset selection, risk assessment and portfolio construction, as well as recent research on the risk of hedge funds. Our focus is the fragilities of those prone to blow up at the time of crisis, despite impeccable track record.

Speaker Bio

Dr. Raphael Douady

Raphael Douady is a French mathematician and economist specializing in data science, financial mathematics and chaos theory at the University of Paris I-Panthéon-Sorbonne. He formerly held the Frey Chair of quantitative finance at Stony Brook University and was academic director of the French Laboratory of Excellence on Financial Regulation. He earned his PhD in Hamiltonian dynamics and has more than 25 years of experience in the financial industry. He has particular interest in researching portfolio risks, for which he has developed especially suited powerful nonlinear statistical and data science models, as well as macroeconomics and systemic risk. He founded fin tech firms Riskdata (risk management for the buy-side) and Datacore Innovations (quantitative portfolio of ETFs). He also advises several quantitative hedge funds and family offices on their risks and trading strategies. With Nassim Taleb and Robert Frey, he founded the Real World Risk Institute, aimed at educating the industry on extreme risks. Douady is a founding member of the Club Praxis, a New York-based think tank advising the French government on its economic policy and sits on the board and, previously on the investment committee, of Friends of IHES, a foundation supporting the Institut des Hautes Etudes Scientifiques (the French brother of Princeton IAS). He is an alumni of Ecole Normale Supérieure in Paris and was awarded a gold medal at the International Mathematical Olympiads.

Video Description

A Talk from Annual Conference 2024

In 1993, John Mulvey co-edited a Special Issue, entitled “Financial Engineering”, in the Annals of Operations Research. In that issue, Guerard, Takano, and Yamane (1992) reported mean-variance efficient portfolios for the Japanese and U.S. equity markets and showed that the use of a regression-weighted composite model of earnings, book value, cash flow, sales, and their relative variables and forecasted earnings, outperformed their respective equity benchmarks by approximately 400 basis points annually. William T. (Bill) Ziemba was the referee of the Guerard et al. (1993) paper. Markowitz and Xu (1994) tested the composite model strategy and found that its excess returns were statistically significant from a variety of models tested, and the composite model strategy was not the result of data mining. Thirty years after the issue, we report factor back testing results and robust regression modeling in creating optimized US and Japanese portfolio results for the 1995-2022 period, a combination of methods and the latest commercially available multi-factor models for portfolio selection. Recent publications by Markowitz, Guerard, and Xu report additional support for the absence of data mining. Furthermore, the weighted latent root regression modeling is still relevant. Our results suggest that stock selection models can be effectively employed to deliver excess returns. The authors believe that financial anomalies exist, persist, and most likely will exist and can be profitably exploited.

Speaker Bio

Dr. John Guerard

John Guerard, PhD, is Co-Chief Investment Engineering and Research, Pacific Spirit Investments, living in Bluffton, South Carolina.

John was a member of the McKinley Capital Management Scientific Advisory Board and has served as an Affiliate Instructor in the Department of Applied Mathematics, the Computational Finance and Risk Management Program, The University of Washington, Seattle, WA. He served almost 15 years as Director of Quantitative Research at McKinley Capital Management, in Anchorage, Alaska. John previously worked at Drexel, Burnham Lambert, and Daiwa Securities, where he was co-Portfolio Manager, with Dr. Harry Markowitz, who was awarded The Nobel Prize in Economic Sciences in 1991, on Fund Academy and The Japan Equity Fund. John was awarded the first Moskowitz Prize for outstanding research in socially responsible investing in 1997. John earned his AB in Economics from Duke University, MA in Economics from the University of Virginia, MSIM from the Georgia Institute of Technology, and Ph.D. in Finance from the University of Texas, Austin. John taught at The University of Virginia, in its McIntire School of Commerce, and at Lehigh University.

Video Description

A Talk from Annual Conference 2024

We consider thought under its two indissociable aspects, concept and intuition. You can only expand the concept by analysis, and fail to synthesize objects, in Kant’s sense of objects of possible experience. For that purpose, something must be added to the concept, which is sensibility, or intuition. Concepts without intuition are empty, according to Kant.

Probability theory is a theory of the concrete world, as such indissociable from possible experience. For this reason, there is something more in probability theory than in real analysis. It is the concrete situation, or trial of the world, or random sample ω, without which probability theory would lack the two characteristic notions of independence and repetition. The trial ω is the Kantian sensibility or intuition of probability theory. The event A, which alone is subject to analysis and is assigned probability, thus admits the concrete situation ω as content ( ω ϵ A).

Under the concepts of probability theory, the Black-Scholes-Merton (BSM) analysis leads only to the dynamic replication of contingent payoffs, and to no market thereof. To change constant volatility into stochastic volatility does not help. In order to exist, the derivatives market requires a new form of Kantian sensibility, or intuition, or concrete situation, which we denote by ϖ. The situation ϖ of the market is radically different from the situation ω of probability theory or statistical analysis. Implied volatility is logically (even temporally) incompatible with historical volatility.

The specific sensibility of the market , ϖ, is the “exchange”. As Helyette Geman says: “BSM is risky by definition because it amounts to exchanging volatility.” As a result, BSM must be thought differently. It becomes the first instance of the “market models”, whose conceptual analysis relies on the “trading decision” and the “pricing function”, or the very denial of probability and stochastic structure.

Speaker Bio

Elie Ayache

Elie Ayache is a former option market-maker on MATIF (1987-1990) and LIFFE (1990-1995). He is the co-founder (1999) and currently the CEO of ITO 33, a software company specializing in convertible bond pricing, and more generally equity index and equity-to-credit single name derivative pricing. Elie is the author of The Blank Swan: The End of Probability (Wiley 2010) and The Medium of Contingency: An Inverse View of the Market (Palgrave 2015).

Video Description

A Talk from Machine Learning in Quant Finance Conference 2024

Quantum and Classical machine learning methods are based on fundamentally different mathematical paradigms and have been shown to be capable of efficiently predicting different patterns in data. Early results suggest that these complementary patterns can be combined through ensemble methods to increase detection performance in such use cases as fraud detection, customer analytics, intrusion detection and more.

Speaker Bio

Dr. Noelle Ibrahim

Dr. Noelle Ibrahim is a Technical Sales Executive for IBM Quantum. Previously she was Associate Partner and IBM Quantum Industry Consultant for the Banking and Financial Markets industries. She currently works across sectors and industries, bringing an ability to cross-pollinate ideas to benefit investment banks.

She has worked across verticals within the financial services sector, leading major transformational risk initiatives including stress testing and IFS9. In addition to leading major transformational risk initiatives in the risk space, she has also worked in derivatives pricing, including vanilla and exotic options and modelling of cash flows from structured investment vehicles. She also has experience in the Fintech industry where she worked for a start up applying AI to art as an asset class. She has a depth and breadth of experience in quantitative finance, including quantitative models for Value at Risk (VaR), CVaR, Black-Scholes, Exotic Options pricing and back-testing, credit risk models for PD,EAD,LGD and more.

Noelle has a Ph.D. in Applied Physics from Columbia University, specializing in Quantum Monte Carlo methods for modelling classical and quantum systems. She also holds an M.Sc. in Quantum Optics and Condensed Matter Physics from the University of Toronto and a Bachelor’s degree in Physics from the University of British Columbia.