Valuing corporate bonds in systemic economies is challenging due to intricate webs of inter-institutional exposures. When a bank defaults, cascading losses propagate through the network, with payments determined by a system of fixed-point equations lacking closed-form solutions. Standard Monte Carlo methods cannot capture rare yet critical default events, while existing rare-event simulation techniques fail to account for higher-order network effects and scale poorly with network size. To overcome these challenges, we propose a novel approach -- Bi-Level Importance Sampling with Splitting -- and characterize individual bank defaults by decoupling them from the network's complex fixed-point dynamics. This separation enables a two-stage estimation process that directly generates samples from the banks' default events. We demonstrate theoretically that the method is both scalable and asymptotically optimal, and validate its effectiveness through numerical studies on empirically observed networks.
Traditional stochastic control methods in finance struggle in real world markets due to their reliance on simplifying assumptions and stylized frameworks. Such methods typically perform well in specific, well defined environments but yield suboptimal results in changed, non stationary ones. We introduce FinFlowRL, a novel framework for financial optimal stochastic control. The framework pretrains an adaptive meta policy learning from multiple expert strategies, then finetunes through reinforcement learning in the noise space to optimize the generative process. By employing action chunking generating action sequences rather than single decisions, it addresses the non Markovian nature of markets. FinFlowRL consistently outperforms individually optimized experts across diverse market conditions.
The volatility fitting is one of the core problems in the equity derivatives business. Through a set of deterministic rules, the degrees of freedom in the implied volatility surface encoding (parametrization, density, diffusion) are defined. Whilst very effective, this approach widespread in the industry is not natively tailored to learn from shifts in market regimes and discover unsuspected optimal behaviors. In this paper, we change the classical paradigm and apply the latest advances in Deep Reinforcement Learning(DRL) to solve the fitting problem. In particular, we show that variants of Deep Deterministic Policy Gradient (DDPG) and Soft Actor Critic (SAC) can achieve at least as good as standard fitting algorithms. Furthermore, we explain why the reinforcement learning framework is appropriate to handle complex objective functions and is natively adapted for online learning.
The cytochrome bcl complex is an oligomeric membrane protein complex which is a component of the mitochondrial respiratory chain and of the electron transfer chains of numerous bacteria which use oxygen, nitrogen, and sulfur compounds as terminal electron acceptors. The cytochrome bcl complex also participates in the cyclic transfer of electrons to and from the photosynthetic reaction centers in anoxygenic photosynthetic bacteria. In all of these species the cytochrome bcl complex transfers electrons from ubiquinol to cytochrome c and links this electron transfer to translocation of protons across the membrane in which the bcl complex resides. The mechanism by which the cytochrome bcl complex links electron transfer to proton translocation is the protonmotive Q cycle (1). This protonmotive electron transfer is one of the most important mechanisms of cellular energy transduction, found in a phylogenetically diverse range of organisms (2). The purpose of this review is to explain the protonmotive Q cycle.
Apparue au milieu du xve siècle, la presse à caractères mobiles de Gutenberg fait rapidement de l’imprimé religieux un vecteur essentiel des idées de la Réforme et de la Contre-Réforme. Mais une autre révolution est en marche : on assiste à la naissance d’un véritable marché, dont s’emparent autant les hommes d’Église – qui multiplient les commandes, Luther en tête – que les professionnels de l’imprimerie et de la librairie, à l’instar de la dynastie des Cramoisy. Si les imprimeurs doivent faire face à de véritables défis techniques (imprimer en grande quantité, imprimer des illustrations ou des partitions), ils profitent aussi de la manne financière qui les accompagne. D’autre part, un véritable circuit de distribution voit le jour, parfois très officiellement (avec la mise en place de l’Imprimerie royale en France), parfois de façon clandestine (pour apporter en Angleterre les traductions interdites du Nouveau Testament). Enfin, la production se diversifie, fournissant aux fidèles des textes pour accompagner leur foi, comme les Gesangbücher allemands, et au clergé des ouvrages pour encadrer ses pratiques, comme les rituels français. De la France à l’Espagne en passant par l’Allemagne et l’Angleterre, onze spécialistes du fait religieux s’intéressent aux imprimés en tant qu’objets. Exemples à l’appui, ils présentent les différents acteurs qui ont favorisé la production et la diffusion de ces ouvrages, ainsi que la grande diversité des livres concernés.
We study the capability of arbitrage-free neural-SDE market models to yield effective strategies for hedging options. In particular, we derive sensitivity-based and minimum-variance-based hedging strategies using these models and examine their performance when applied to various option portfolios using real-world data. Through backtesting analysis over typical and stressed market periods, we show that neural-SDE market models achieve lower hedging errors than Black--Scholes delta and delta-vega hedging consistently over time, and are less sensitive to the tenor choice of hedging instruments. In addition, hedging using market models leads to similar performance to hedging using Heston models, while the former tends to be more robust during stressed market periods.
We focus on extending existing short-rate models, enabling control of the generated implied volatility while preserving analyticity. We achieve this goal by applying the Randomized Affine Diffusion (RAnD) method to the class of short-rate processes under the Heath-Jarrow-Morton framework. Under arbitrage-free conditions, the model parameters can be exogenously stochastic, thus facilitating additional degrees of freedom that enhance the calibration procedure. We show that with the randomized short-rate models, the shapes of implied volatility can be controlled and significantly improve the quality of the model calibration, even for standard 1D variants. In particular, we illustrate that randomization applied to the Hull-White model leads to dynamics of the local volatility type, with the prices for standard volatility-sensitive derivatives explicitly available. The randomized Hull-White (rHW) model offers an almost perfect calibration fit to the swaption implied volatilities.
In this paper, we model the rating process of an entity as a piecewise homogeneous continuous time Markov chain. We focus specifically on calibrating the model to both historical data (rating transition matrices) and market data (CDS quotes), relying on a simple change of measure to switch from the historical probability to the risk-neutral one. We overcome some of the imperfections of the data by proposing a novel calibration procedure, which leads to an improvement of the entire scheme. We apply our model to compute bilateral credit and debit valuation adjustments of a netting set under a CSA with thresholds depending on ratings of the two parties.
We propose a new approach for trading VIX futures. We assume that the term structure of VIX futures follows a Markov model. Our trading strategy selects a position in VIX futures by maximizing the expected utility for a day-ahead horizon given the current shape and level of the term structure. Computationally, we model the functional dependence between the VIX futures curve, the VIX futures positions, and the expected utility as a deep neural network with five hidden layers. Out-of-sample backtests of the VIX futures trading strategy suggest that this approach gives rise to reasonable portfolio performance, and to positions in which the investor will be either long or short VIX futures contracts depending on the market environment.
We investigate the suitability of toroidal microcavities for strong-coupling cavity quantum electrodynamics (QED). Numerical modeling of the optical modes demonstrate a significant reduction of the modal volume with respect to the whispering gallery modes of dielectric spheres, while retaining the high-quality factors representative of spherical cavities. The extra degree of freedom of toroid microcavities can be used to achieve improved cavity QED characteristics. Numerical results for atom-cavity coupling strength g, critical atom number No, and critical photon number no for cesium are calculated and shown to exceed values currently possible using Fabry-Perot cavities. Modeling predicts coupling rates g/2π exceeding 700 MHz and critical atom numbers approaching 10^(-7) in optimized structures. Furthermore, preliminary experimental measurements of toroidal cavities at a wavelength of 852 nm indicate that quality factors in excess of 108 can be obtained in a 50-µm principal diameter cavity, which would result in strong-coupling values of (g/(2π),n(0),N-0) = (86 MHz, 4.6 x 10^(-4), 1.0 x 10^(-3)).
A photonic nanocavity with a high Q factor of 100,000 and a modal volume V of 0.71 cubic wavelengths, is demonstrated. According to the cavity design rule that we discovered recently, we further improve a point-defect cavity in a two-dimensional (2D) photonic crystal (PC) slab, where the arrangement of six air holes near the cavity edges is fine-tuned. We demonstrate that the measured Q factor for the designed cavity increases by a factor of 20 relative to that for a cavity without displaced air holes, while the calculated modal volume remains almost constant.
Aurélien Alfonsi, Adel Cherchali, Jose Arturo Infante Acevedo
This paper studies the multilevel Monte-Carlo estimator for the expectation of a maximum of conditional expectations. This problem arises naturally when considering many stress tests and appears in the calculation of the interest rate module of the standard formula for the SCR. We obtain theoretical convergence results that complements the recent work of Giles and Goda and gives some additional tractability through a parameter that somehow describes regularity properties around the maximum. We then apply the MLMC estimator to the calculation of the SCR at future dates with the standard formula for an ALM savings business on life insurance. We compare it with estimators obtained with Least Square Monte-Carlo or Neural Networks. We find that the MLMC estimator is computationally more efficient and has the main advantage to avoid regression issues, which is particularly significant in the context of projection of a balance sheet by an insurer due to the path dependency. Last, we discuss the potentiality of this numerical method and analyze in particular the effect of the portfolio allocation on the SCR at future~dates.
Takaaki Koike, Yuri F. Saporito, Rodrigo S. Targino
This paper is concerned with the process of risk allocation for a generic multivariate model when the risk measure is chosen as the Value-at-Risk (VaR). We recast the traditional Euler contributions from an expectation conditional on an event of zero probability to a ratio involving conditional expectations whose conditioning events have strictly positive probability. We derive an analytical form of the proposed representation of VaR contributions for various parametric models. Our numerical experiments show that the estimator using this novel representation outperforms the standard Monte Carlo estimator in terms of bias and variance. Moreover, unlike the existing estimators, the proposed estimator is free from hyperparameters under a parametric setting.
We have demonstrated all-optical bistable switching operation of resonant-tunnelling devices with ultra-small high-Q Si photonic-crystal nanocavities. Due to their high Q/V ratio, the switching energy is extremely small in comparison with that of conventional devices using the same optical nonlinear mechanism. We also show that they exhibit all-opticaltransistor action by using two resonant modes. These ultrasmall unique nonlinear bistable devices have potentials to function as various signal processing functions in photonic-crystal-based optical-circuits.