BEGIN:VCALENDAR VERSION:2.0 PRODID:researchseminars.org CALSCALE:GREGORIAN X-WR-CALNAME:researchseminars.org BEGIN:VEVENT SUMMARY:Ilaria Peri (University of London) DTSTART:20201028T210000Z DTEND:20201028T220000Z DTSTAMP:20260404T110957Z UID:Quantitative_Finance/1 DESCRIPTION:Title: On the properties of Lambda quantiles and financial ap plications\nby Ilaria Peri (University of London) as part of Quantitat ive Finance Seminar\n\n\nAbstract\nThe aim of this talk is to provide an o verview of the theory on Lambda quantiles and present its versatility via financial applications. Lambda quantiles are generalised quantiles\, intro duced by Frittelli\, Maggis\, P. (2014) under the name of Lambda Value at Risk. In particular\, Lambda quantiles differ from usual quantiles in that the constant lambda is replaced with a threshold function Lambda allowing for more flexibility of the confidence level. We discuss alternative defi nitions of Lambda quantiles and derive their fundamental properties. We pr ovide an axiomatic foundation for non-increasing Lambda quantiles based on the well-known locality property of quantiles that here we formalize. As original statistical application\, we introduce the so-called Lambda quant ile regression. We present the estimation of Lambda quantiles in a market risk setting by comparing methods based on classical assumptions on the re turn distribution and the Lambda quantile regression. We conclude with a b acktesting exercise and discuss how this backtesting framework can be exte nded to other risk measures.\n\nIlaria Peri is a lecturer in mathematical finance at the Birkbeck University of London. She earned her doctorate fro m the University of Milan-Bicocca under the supervision of Marco Frittelli . Prior to joining academics\, she worked as a financial consultant gainin g experience in risk management and banking operations. Her research focus es on risk measures' theory and applications. Her major contribution is th e introduction of the generalized quantile called Lambda value at risk on which she has been conducting theoretical studies and empirical applicatio ns. Her research has been published in internationally recognized journals and presented at invited seminars in academic and professional contexts\, including regulatory authorities.\n LOCATION:https://stable.researchseminars.org/talk/Quantitative_Finance/1/ END:VEVENT BEGIN:VEVENT SUMMARY:Lukasz Szpruch (University of Edinburgh and Alan Turing Institute) DTSTART:20201125T220000Z DTEND:20201125T230000Z DTSTAMP:20260404T110957Z UID:Quantitative_Finance/2 DESCRIPTION:Title: Gradient Flows\, Stochastic Control and Robust pricing and hedging via neural SDEs\nby Lukasz Szpruch (University of Edinbur gh and Alan Turing Institute) as part of Quantitative Finance Seminar\n\n\ nAbstract\nThere is overwhelming empirical evidence that deep neural netwo rks trained with stochastic gradient descent perform (extremely) well in t he high dimensional settings. Nonetheless\, a complete mathematical theory that would provide theoretical guarantees why and when these methods work so well has been elusive. In this talk\, I will demonstrate how one may l everage control theory and the theory of statistical sampling to study the convergence of stochastic gradient algorithms used in deep learning. Conv ersely\, I will show that machine learning perspective leads to new algori thms for (stochastic) control problems and offers a fresh perspective on c lassical quantitative finance problems. Indeed\, modern data science techn iques are opening the door to more robust and data-driven model selection mechanisms. Indeed\, deep generative modelling is opening the door to more robust and data-driven model selection mechanisms. By combining neural ne tworks with risk models based on classical stochastic differential equatio ns (SDEs)\, we find robust bounds for prices of derivatives and the corres ponding hedging strategies while incorporating relevant market data. Neura l SDEs allow consistent calibration under both the risk-neutral and the re al-world measures. Thus the model can be used to simulate market scenarios needed for assessing risk profiles and hedging strategies. We develop and analyse novel algorithms needed for efficient use of neural SDEs. We vali date our approach with numerical experiments using both local and stochast ic volatility models. We will also show that neural SDEs can be used to ca librate to SPX/VIX options.\n\nBio: Lukasz is a Reader (Associate Professo r) at the School of Mathematics\, University of Edinburgh. He is also the director of the Finance and Economics programme at The Alan Turing Institu te\, the UK national institute for data science and AI. Previously Lukasz was a Nomura Junior Research Fellow at the Institute of Mathematics\, Univ ersity of Oxford\, and a member of the Oxford-Man Institute for Quantitati ve Finance. Lukasz has a broad research interest in Machine and Reinforcem ent Learning\, Statistics and Game Theory.\n LOCATION:https://stable.researchseminars.org/talk/Quantitative_Finance/2/ END:VEVENT BEGIN:VEVENT SUMMARY:Mark Reesor (Wilfrid Laurier University) DTSTART:20210127T220000Z DTEND:20210127T230000Z DTSTAMP:20260404T110957Z UID:Quantitative_Finance/3 DESCRIPTION:Title: Know Your Clients' Behaviours: A Cluster Analysis of F inancial Transactions\nby Mark Reesor (Wilfrid Laurier University) as part of Quantitative Finance Seminar\n\n\nAbstract\nn Canada\, financial a dvisors and dealers are required by provincial securities commissions and self-regulatory organizations---charged with direct regulation over invest ment dealers and mutual fund dealers---to respectively collect and maintai n know your client (KYC) information\, such as their age or risk tolerance \, for investor accounts. With this information\, investors\, under their advisor's guidance\, make decisions on their investments that are presumed to be beneficial to their investment goals. Our unique dataset is provide d by a financial investment dealer with over 50\,000 accounts for over 23\ ,000 clients covering the period from January 1st to August 12th 2019. We use a modified behavioral finance recency\, frequency\, monetary model for engineering features that quantify investor behaviours\, and (unsupervise d) machine learning clustering algorithms to find groups of investors that behave similarly. We show that the KYC information---such as gender\, res idence region\, and marital status---does not explain client behaviours\, whereas eight variables for trade and transaction frequency and volume are most informative. Hence\, our results should encourage financial regulato rs and advisors to use more advanced metrics to better understand and pred ict investor behaviours.\n\nThis is joint work with John Thompson\, Longlo ng Feng\, and Adam Metzler of Wilfrid Laurier University and Chuck Grace o f the Richard Ivey School of Business\, Western University.\n\nBio: Mark R eesor earned his Master's and PhD degree in Statistics from the University of Waterloo. After earning his doctorate Mark worked as an analyst in the Financial Markets Department at the Bank of Canada. From 2002 until 2016\ , Dr. Reesor was a faculty member at Western University in the Departments of Statistics and Actuarial Science and of Applied Mathematics and in the finance area at the Richard Ivey School of Business. Since 2016\, he has been in the Math Department at Wilfrid Laurier University. Dr. Reesor has a varied research program including works in personal finance\, corporate finance\, financial stability\, securities class actions\, risk management \, derivatives\, and Monte Carlo Methods. In addition\, Mark was a foundin g member of the Committee to Establish the National Institute of Finance\, helping to create of the U.S. Office of Financial Research.\n LOCATION:https://stable.researchseminars.org/talk/Quantitative_Finance/3/ END:VEVENT BEGIN:VEVENT SUMMARY:Alexey Rubtsov (Ryerson University) DTSTART:20210224T220000Z DTEND:20210224T230000Z DTSTAMP:20260404T110957Z UID:Quantitative_Finance/4 DESCRIPTION:Title: Systemic Risk Driven Portfolio Selection\nby Alexe y Rubtsov (Ryerson University) as part of Quantitative Finance Seminar\n\n \nAbstract\nWe consider an investor whose objective is to trade off tail r isk and expected growth of the investment. We measure tail risk through po rtfolio's expected losses conditioned on the occurrence of a systemic even t: financial market loss being exactly at\, or at least at\, its VaR level and investor's portfolio losses being above their CoVaR level. We obtain a closed-form solution to the investment problem\, and decompose it in ter ms of the Markowitz mean--variance portfolio and an adjustment for systemi c risk. We show that VaR and CoVaR confidence levels control\, respectivel y\, the relative sensitivity of the investor's objective function to portf olio--market correlation and portfolio variance. Our empirical analysis de monstrates that the investor attains higher risk-adjusted returns\, compar ed to well known benchmark portfolio criteria\, during times of market dow nturn. Portfolios that perform best in adverse market conditions are less diversified and concentrate on few stocks which have low correlation with the market.\n LOCATION:https://stable.researchseminars.org/talk/Quantitative_Finance/4/ END:VEVENT BEGIN:VEVENT SUMMARY:Stan Uryasev (Stony Brook University) DTSTART:20210428T210000Z DTEND:20210428T220000Z DTSTAMP:20260404T110957Z UID:Quantitative_Finance/5 DESCRIPTION:Title: Drawdown Beta and Portfolio Optimization\nby Stan Uryasev (Stony Brook University) as part of Quantitative Finance Seminar\n \n\nAbstract\nWe introduce a new dynamic portfolio performance risk measur e called Expected Regret of Drawdown (ERoD) which is an average of drawdow ns exceeding a specified thresholde(e.g.\,e=10%). ERoD is similar to Condi tional Drawdown-at-Risk (CDaR) which is the average of osmepercentage of l argest drawdowns. CDaR and ERoD portfolio portfolio optimization are equiv alent and results in the same portfolios. Necessary optimally conditions f or ERoD portfolio optimization lead to Capital Asset Pricing Model (CAPM) equations. ERoD Beta\, similar to the Standard Beta\,relates expected retu rns of securities and market. ERoD Beta equals to [average losses of a sec urityover times intervals when market is in drawdown exceedinge] divided b y [average losses of marketin drawdowns exceedinge]. Therefore\, a negativ e ERoD Beta identifies a security which has positive returns when market i s in drawdown. ERoD Beta accounts for only time intervals when market is i ndrawdown and conceptually differs from Standard Beta which does not disti nguish up and downmovements of the market. However\, ERoD Beta also provid es quite different results compared toDownside Beta which is based on Lowe r Semi-deviation. ERoD Beta is conceptually close to CDaRBeta which is bas ed on a percentage of worst case market drawdowns. We have built a website reporting CDaR and ERoD Betas for stocks and S&P 500 index as an optimal market portfolio. The case study showed that CDaR and ERoD Betas exhibit p ersistence over time intervals and can beused in risk management and portf olio construction. This talk is based on joint work with Rui Ding from Sto ny Brook University.\n\nBio: Stan Uryasev received his M.S. in Applied Mat hematics from the Moscow Institute of Physics and Technology (MIPT)\, Russ ia\, in 1979 and Ph.D. in Applied Mathematics from the Glushkov Institute of Cybernetics\, Kiev\, Ukraine in 1983. From 1979 to 1987 he held a resea rch position at the Glushkov Institute. From 1988 to 1992 he was a Researc h Scholar at the International Institute for Applied System Analysis\, Lux enburg\, Austria. From 1992 to 1998 he held the Scientist position at the Risk and Reliability Group\, Brookhaven National Laboratory\, Upton\, NY. From 1998 to 2019 he was the George and Rolande Willis Endowed Professor a t the University of Florida\, and the director of the Risk Management and Financial Engineering Lab. \n\nHis research and teaching interests include quantitative finance\, risk management\, stochastic optimization\, machin e learning\, and military operations research. See Google Scholar for the list of the most cited publications\, https://scholar.google.com/citations ?hl=en&user=Uwg1zpkAAAAJ. Here is the full list of publications. \n\nHis j oint paper with Prof. Rockafellar on Optimization of Conditional Value-At- Risk in The Journal of Risk\, Vol. 2\, No. 3\, 2000 is among the 100 most cited papers in Finance. Many risk management/optimization packages implem ented the approach suggested in this paper (MATLAB implemented a toolbox). \n\nThe important theoretical contribution presenting a unified scheme fo r portfolio optimization\, statistical estimation\, risk management\, and utility theory: Rockafellar R.T. and S. Uryasev. The Fundamental Risk Quad rangle in Risk Management\, Optimization\, and Statistical Estimation. Sur veys in Operations Research and Management Science\, 18\, 2013. \n\nCollab orative research with industry has been documented in the library of Case Studies containing Portfolio Safeguard (PSG) codes\, data\, and calculatio n results in Text\, MATLAB\, and R environments. See the list of Case Stud ies in Financial Engineering\, Advanced Statistics and other areas.\n LOCATION:https://stable.researchseminars.org/talk/Quantitative_Finance/5/ END:VEVENT BEGIN:VEVENT SUMMARY:Maxime Bergeron (Riskfuel) DTSTART:20210331T210000Z DTEND:20210331T220000Z DTSTAMP:20260404T110957Z UID:Quantitative_Finance/6 DESCRIPTION:Title: Deeply Learning Derivatives: from Hilbert to Riskfuel< /a>\nby Maxime Bergeron (Riskfuel) as part of Quantitative Finance Seminar \n\n\nAbstract\nThe motivation behind Hilbert's 13th problem is often over looked. In his original statement of the problem\, he opens with: "Nomogra phy deals with the problem of solving equations by means of drawing famili es of curves depending on an arbitrary parameter". The question he posed s ought to identify the family of functions amenable to such graphical solve rs that were essential tools of his time. While the question in its origin al (algebraic) form remains open to this day\, in the continuous realm it turns out that there is no such thing as a truly multivariate function. In this talk\, we will explain how these ideas fit into the modern deep lear ning framework and\, ultimately\, allow us to build networks that replicat e the solutions operator of stochastic differential equations governing th e valuation of high dimensional contingent claims.\n LOCATION:https://stable.researchseminars.org/talk/Quantitative_Finance/6/ END:VEVENT BEGIN:VEVENT SUMMARY:Cody Hyndman (Concordia University) DTSTART:20210929T210000Z DTEND:20210929T220000Z DTSTAMP:20260404T110957Z UID:Quantitative_Finance/7 DESCRIPTION:Title: Arbitrage-free yield curve and bond price forecasting by deep neural networks\nby Cody Hyndman (Concordia University) as par t of Quantitative Finance Seminar\n\nAbstract: TBA\n LOCATION:https://stable.researchseminars.org/talk/Quantitative_Finance/7/ END:VEVENT BEGIN:VEVENT SUMMARY:Samuel Cohen (University of Oxford) DTSTART:20211027T210000Z DTEND:20211027T220000Z DTSTAMP:20260404T110957Z UID:Quantitative_Finance/8 DESCRIPTION:Title: Arbitrage-free neural-SDE market models\nby Samuel Cohen (University of Oxford) as part of Quantitative Finance Seminar\n\n\ nAbstract\nModelling joint dynamics of liquid vanilla options is crucial f or arbitrage-free pricing of illiquid derivatives and managing risks of op tion trade books. This paper develops a nonparametric model for the Europe an options book respecting underlying financial constraints and while bein g practically implementable. In this talk\, we will consider a state space for prices which are free from static (or model-independent) arbitrage an d study the inference problem where a model is learnt from discrete time s eries data of stock and option prices. We use neural networks as function approximators for the drift and diffusion of the modelled SDE system\, and impose constraints on the neural nets such that no-arbitrage conditions a re preserved. In particular\, we give methods to calibrate neural SDE mode ls which are guaranteed to satisfy a set of linear inequalities. We valida te our approach with numerical experiments using data generated from a Hes ton stochastic volatility model\, and with observed market data.\n\nBio: S am Cohen is a mathematician based at the Mathematical Institute in Oxford\ , and at the Alan Turing Institute in London. He obtained his PhD in 2011 at the University of Adelaide\, under the supervision of Robert Elliott. H e is interested in the interaction between statistical learning\, decision making and modelling\, with particular applications in finance and econom ics.\n LOCATION:https://stable.researchseminars.org/talk/Quantitative_Finance/8/ END:VEVENT BEGIN:VEVENT SUMMARY:Beatrice Acciaio (ETH Zürich) DTSTART:20211124T220000Z DTEND:20211124T230000Z DTSTAMP:20260404T110957Z UID:Quantitative_Finance/9 DESCRIPTION:Title: A robust framework for pricing and hedging American op tions\nby Beatrice Acciaio (ETH Zürich) as part of Quantitative Finan ce Seminar\n\n\nAbstract\nIn this talk I will introduce a suitable framewo rk for pricing and hedging American options in a model-independent way. Th is is based on the recently developed concept of adapted Wasserstein dista nces. Beside recovering super-replication duality\, in such a framework we establish existence and a geometric characterization of the extremal pric ing models.\n\nThis is based on a joint work with D. Bartl\, B. Beiglboeck and G. Pammer.\n\nBio: Beatrice Acciaio is Professor of Mathematics at ET H Zurich since 2020. Before joining ETH\, Beatrice was associate professor at the London School of Economics\, and prior to that she has been part o f several research groups\, at the Technical University of Vienna\, the Un iversity of Perugia\, and the University of Vienna. Beatrice completed her PhD in 2006 under the supervision of Walter Schachermayer.\n\nBeatrice's main areas of research are probability\, mathematical finance\, and optima l transport.\n\nBeatrice is member of the Council of the Bachelier Finance Society\, she is Associate Editor for the SIAM Journal on Financial Mathe matics\, for Finance and Stochastics\, for Mathematical Finance\, and for the Bocconi & Springer Series on Mathematics\, Statistics\, Finance and Ec onomics.\n LOCATION:https://stable.researchseminars.org/talk/Quantitative_Finance/9/ END:VEVENT BEGIN:VEVENT SUMMARY:Matthew Lorig (University of Washington) DTSTART:20220126T220000Z DTEND:20220126T230000Z DTSTAMP:20260404T110957Z UID:Quantitative_Finance/10 DESCRIPTION:by Matthew Lorig (University of Washington) as part of Quantit ative Finance Seminar\n\nAbstract: TBA\n LOCATION:https://stable.researchseminars.org/talk/Quantitative_Finance/10/ END:VEVENT BEGIN:VEVENT SUMMARY:Peter Tankov (ENSAE\, Institute Polytechnique de Paris) DTSTART:20220223T220000Z DTEND:20220223T230000Z DTSTAMP:20260404T110957Z UID:Quantitative_Finance/11 DESCRIPTION:by Peter Tankov (ENSAE\, Institute Polytechnique de Paris) as part of Quantitative Finance Seminar\n\nAbstract: TBA\n LOCATION:https://stable.researchseminars.org/talk/Quantitative_Finance/11/ END:VEVENT BEGIN:VEVENT SUMMARY:Lars Stentoft (Western University) DTSTART:20220330T210000Z DTEND:20220330T220000Z DTSTAMP:20260404T110957Z UID:Quantitative_Finance/12 DESCRIPTION:by Lars Stentoft (Western University) as part of Quantitative Finance Seminar\n\nAbstract: TBA\n LOCATION:https://stable.researchseminars.org/talk/Quantitative_Finance/12/ END:VEVENT END:VCALENDAR