arXiv subjects

Henryk Gzyl

Publications and source records attributed to Henryk Gzyl.

At least 19 recordsLinked to original sources

arXiv2026

An Entropic Factor Model for Robust Portfolio Replication

Portfolio replication, or the construction of a tradable basket of assets to match the risk-return profile of a target benchmark, is fundamentally an ill-posed inverse problem. When restricted to a subset of available assets, classical variance-minimizing models often yield unstable, over-leveraged portfolios highly vulnerable to market shocks. We propose a unified, two-stage methodology rooted in information theory to achieve robust portfolio replication. First, we model the constituent asset returns against target factors, estimating parameters within data-driven empirical bounds via an entropy minimization principle. Second, using the same entropic approach, we determine the optimal weight replication. In both cases we use an entropy function of the Fermi-Dirac type defined directly on sets of constraints of the inverse problem. We validate this Entropic Factor Model (EFM) against standard Ordinary Least Squares (OLS) across five numerical experiments, including standard equity tracking, multi-asset synthesis, and severe stress-test scenarios. Empirical results demonstrate that the EFM consistently outperforms OLS in terms of annualized turnover and net-of-fees returns. Crucially, during the COVID-19 market crash and under severe idiosyncratic data corruption, the entropic framework acts as a probabilistic ``circuit breaker", defensively reducing capital allocation to compromised assets and providing a highly robust, risk-averse solution for generalized portfolio replication.

q-fin.PM↗

arXiv2026

On spooky action at a distance and conditional probabilities

The aim of this exposé is to make explicit the analogy between the classical notion of non-independent probability distribution and the quantum notion of entangled state. To bring that analogy forth, we consider a classical systems with two dependent random variables and a quantum system with two components. In the classical case, afet observing one of the random variables, the underlying sample space and the probability distribution change. In the quantum case, when and event pertaining to one of the components is observed, the post-measurement state captures, both, the change in the state of the system and implicitly the new probability distribution. The predictions after a measurement in the classical case and in the quantum case, have to be computed with the conditional distribution given the value of the observed variable.

quant-ph↗

arXiv2026

A predictive solution of the EPR paradox

In this work, we examine the paradox proposed by Einstein, Podolsky, and Rosen (EPR). They argued that since one may know the exact momentum of a particle without measurement and subsequently measure its position, a contradiction with the Heisenberg uncertainty principle arises. We demonstrate that there is no paradox by two equivalent approaches: first, by computing the quantum conditional expectation to make predictions after a measurement; and second, using the von Neumann post-measurement state. We establish the equivalence between these two methods. In both cases the predictor is an operator valued function of the observables being measured. This ensures that no violation of the Heisenberg uncertainty principle occurs.

quant-ph↗

arXiv2026

On the EPR paradox in systems with finite number of levels (Revised)

In this work we reexamine the EPR paradox for composite systems with a finite number of levels. The analysis emphasizes the connection between measurements and conditional probabilities. This connection implies that when a measurement is performed, the microscopic states compatible with the measurement is different from the class of all possible microscopic states, therefore the new quantum state and the probability distribution change and become a function of the observable being measured. Therefore, the predictions that one can make given the knowledge of the result of a measurement change. Systems with finitely many levels are simpler to describe because the analysis is not encumbered by the mathematical technicalities of the continuous case, the underlying physical interpretations are the same and the experimental setups used to test quantum mechanics with the paradox in mind finitely many levels.e same.

quant-ph↗

arXiv2025

Qubit thermalization by random pulses: Asymptotic state factorization

Here we consider an analytically tractable model of a two level quantum system subject to random shocks and prove that it decays asymptotically to a trivial state, that is, to a state in which the two levels have equal probability of occupation. In a two qubit system, if the shocks affect each qubit independently, the equilibrium density matrix becomes a simple product of the one qubit equilibrium density matrices regardless of the nature of the initial state. This has potential applications to entangles qubits in quantum computers.

quant-ph↗

arXiv2025

Classification by Separating Hypersurfaces: An Entropic Approach

We consider the following classification problem: Given a population of individuals characterized by a set of attributes represented as a vector in ${\mathbb R}^N$, the goal is to find a hyperplane in ${\mathbb R}^N$ that separates two sets of points corresponding to two distinct classes. This problem, with a history dating back to the perceptron model, remains central to machine learning. In this paper we propose a novel approach by searching for a vector of parameters in a bounded $N$-dimensional hypercube centered at the origin and a positive vector in ${\mathbb R}^M$, obtained through the minimization of an entropy-based function defined over the space of unknown variables. The method extends to polynomial surfaces, allowing the separation of data points by more complex decision boundaries. This provides a robust alternative to traditional linear or quadratic optimization techniques, such as support vector machines and gradient descent. Numerical experiments demonstrate the efficiency and versatility of the method in handling diverse classification tasks, including linear and non-linear separability.

cs.LG↗

arXiv2025

Portfolio optimization in incomplete markets and price constraints determined by maximum entropy in the mean

A solution to a portfolio optimization problem is always conditioned by constraints on the initial capital and the price of the available market assets. If a risk neutral measure is known, then the price of each asset is the discounted expected value of the asset's price under this measure. But if the market is incomplete, the risk neutral measure is not unique, and there is a range of possible prices for each asset, which can be identified with bid-ask ranges. We present in this paper an effective method to determine the current prices of a collection of assets in incomplete markets, and such that these prices comply with the cost constraints for a portfolio optimization problem. Our workhorse is the method of maximum entropy in the mean to adjust a distortion function from bid-ask market data. This distortion function plays the role of a risk neutral measure, which is used to price the assets, and the distorted probability that it determines reproduces bid-ask market values. We carry out numerical examples to study the effect on portfolio returns of the computation of prices of the assets conforming the portfolio with the proposed methodology.

math.OC↗

arXiv2025

Determining a credit transition matrix from cumulative default probabilities

To quantify the changes in the credit rating of a bond is an important mathematical problem for the credit rating industry. To think of the credit rating as the state a Markov chain is an interesting proposal leading to challenges in mathematical modeling. Since cumulative default rates are more readily measurable than credit migrations, a natural question is whether the credit transition matrix (CTM) can be determined from the knowledge of the cumulative default probabilities. Here we use a connection between the CTM and the cumulative default probabilities to setup an ill-posed, linear inverse problem with box constraints, which we solve by an entropy minimization procedure. This approach is interesting on several counts. On the one hand, we may have less data that unknowns, and on the other hand, even when we have as much data as unknowns, the matrix connecting them may not be invertible, which makes the problem ill-posed. Besides developing the tools to solve the problem, we apply it to several test cases to check the performance of the method. The results are quite satisfactory.

q-fin.CP↗

arXiv2024

Lorentz covariant physical Brownian motion: Classical and quantum

In this work, we re-examine the Goldstein-Kaç velocity switching model from two points of view. On the one hand, we prove that the forward and backward Chapman-Kolmogorov equations of the stochastic process are Lorentz covariant when the trajectories are parameterized by their proper time. On the other hand, to recast the model as a quantum random evolution, we consider restating the Goldstein-Kaç model as a Hamiltonian system, which can then be quantized using the standard correspondence rules. It turns out that the density for the random quantum evolution satisfies a Chapman-Kolmogorov equation similar to that of the classical case, and therefore, it is also Lorentz covariant. We compute the average quantum variance. To finish, we verify that the quantum model is also consistent with special relativity and that transitions outside the light cone, that is, transitions between states with disjoint supports in space-time, cannot occur.

quant-ph↗

arXiv2024

A geometry in the set of solutions to ill-posed linear problems with box constraints: Applications to probabilities on discrete sets

When there are no constraints upon the solutions of the equation $\mathbf{A}\mathbfξ= \mathbf{y},$ where $\mathbf{A}$ is a $K\times N-$matrix, $\mathbfξ\in\mathbb{R}^N$ and $\mathbf{y}\in\mathbb{R}^K$ a given vector, the description of the set of solutions as $\mathbf{y}$ varies in $\mathbb{R}^K$ is well known. But this is not so when the solutions are required to satisfy $\mathbfξ \in \mathcal{K}\prod_{i\leq j\leq N} [a_j,b_j],$ for finite $a_j\leq b_j: 1\leq j\leq N.$ Here we provide a description of the set of solutions as a surface in the constraint set, parameterized by the Lagrange multipliers that come up in a related optimization problem in which $\mathbf{A}\mathbfξ = \mathbf{y}$ appears as a constraint. It is the dependence of the Lagrange multipliers on the data vector $\mathbf{y}$ that determines how the solution changes as the datum changes. The geometry on the solutions is inherited from a Riemannian geometry on the set of constraints induced by the Hessian of an entropy of the Fermi-Dirac type which is the objective in the restatement of the optimization problem mentioned above. We prove that the set of solutions is contained in $\ker(\mathbf{A})^\perp$ in the metric defined as the Hessian of the entropy.

math.RA↗

arXiv2023

Quantum systems in Markovian environments

In this work, we develop a mathematical framework to model a quantum system whose Hamiltonian may depend on the state of changing environment, that evolves according to a Markovian process. When the environment changes its state, the quantum system may suffer a shock that produces an instantaneous transition among its states. The model that we propose can be readily adapted to more general settings.\\ To avoid collateral analytical issues, we consider the case of quantum systems with finite dimensional state space, in which case the observables are described by Hermitian matrices. We show how to average over the environment to predict the expected values of observables.

quant-ph↗

arXiv2023

Ill-posed linear inverse problems with box constraints: A new convex optimization approach

Consider the linear equation $\mathbf{A}\mathbf{x}=\mathbf{y}$, where $\mathbf{A}$ is a $k\times N$-matrix, $\mathbf{x}\in\mathcal{K}\subset \mathbb{R}^N$ and $\mathbf{y}\in\mathbb{R}^M$ a given vector. When $\mathcal{K}$ is a convex set and $M\not= N$ this is a typical ill-posed, linear inverse problem with convex constraints. Here we propose a new way to solve this problem when $\mathcal{K} = \prod_j[a_j,b_j]$. It consists of regarding $\mathbf{A}\mathbf{x}=\mathbf{y}$ as the constraint of a convex minimization problem, in which the objective (cost) function is the dual of a moment generating function. This leads to a nice minimization problem and some interesting comparison results. More importantly, the method provides a solution that lies in the interior of the constraint set $\mathcal{K}$. We also analyze the dependence of the solution on the data and relate it to the Le Chatellier principle.

math.OC↗

arXiv2023

Canonical equivalence of a charge in a time dependent, spatially-homogeneous electromagnetic field to a time-dependent perturbed oscillator

Here we prove that the classical (respectively, quantum) system, consisting of a particle moving in a static electromagnetic field, is canonically (respectively, unitarily) equivalent to a harmonic oscillator perturbed by a spatially homogeneous force field. This system is canonically and unitarily equivalent to a standard oscillator. Therefore, by composing the two transformations we can integrate the initial problem. Actually, the eigenstates of the initial problem turn out to be entangled states of the harmonic oscillator. When the magnetic field is spatially homogeneous but time-dependent, the equivalent harmonic oscillator has a time-varying frequency. This system can be exactly integrated only for some particular cases of the time dependence of the magnetic field. The unitary transformations between the quantum systems are a representation of the canonical transformations by unitary transformations of the corresponding Hilbert spaces.

quant-ph↗

arXiv2023

Canonical equivalence of a particle in a magnetic field to a simple oscillator

It is proved that a classical (respec. quantum) system consisting of a particle in a constant magnetic field is canonically (respec. unitarily) equivalent to a 2-dimensional harmonic oscillator plus a free particle. It is also shown that the eigenvectors of the discrete spectrum are entangled states of the 2-dimensional harmonic oscillator.

quant-ph↗

arXiv2023

Prediction and estimation of random variables with infinite mean or variance

In this paper we propose an optimal predictor of a random variable that has either an infinite mean or an infinite variance. The method consists of transforming the random variable such that the transformed variable has a finite mean and finite variance. The proposed predictor is a generalized arithmetic mean which is similar to the notion of certainty price in utility theory. Typically, the transformation consists of a parametric family of bijections, in which case the parameter might be chosen to minimize the prediction error in the transformed coordinates. The statistical properties of the estimator of the proposed predictor are studied, and confidence intervals are provided. The performance of the procedure is illustrated using simulated and real data.

math.ST↗

arXiv2022

Which portfolio is better? A discussion of several possible comparison criteria

During the last few years, there has been an interest in comparing simple or heuristic procedures for portfolio selection, such as the naive, equal weights, portfolio choice, against more "sophisticated" portfolio choices, and in explaining why, in some cases, the heuristic choice seems to outperform the sophisticated choice. We believe that some of these results may be due to the comparison criterion used. It is the purpose of this note to analyze some ways of comparing the performance of portfolios. We begin by analyzing each criterion proposed on the market line, in which there is only one random return. Several possible comparisons between optimal portfolios and the naive portfolio are possible and easy to establish. Afterwards, we study the case in which there is no risk free asset. In this way, we believe some basic theoretical questions regarding why some portfolios may seem to outperform others can be clarified.

q-fin.PM↗

arXiv2022

Classical and quantum harmonic oscillators subject to a time dependent force

In this work we address the problem of the quantization of a simple harmonic oscillator that is perturbed by a time dependent force. The approach consists of removing the perturbation by a canonical change of coordinates. Since the quantization procedure uses the classical Hamiltonian formalism as staring point, the change of variables is carried out using canonical transformations, and to transform between the quantized systems the canonical transformation is implemented as a unitary transformation mapping the states of the perturbed and unperturbed system onto each other.

quant-ph↗

arXiv2021

Joint probabilities under expected value constraints, transportation problems, maximum entropy in the mean, and geometry in the space of probabilities

There are interesting extensions of the problem of determining a joint probability with known marginals. On the one hand, one may impose size constraints on the joint probabilities. On the other, one may impose additional constraints like the expected values of known random variables. If we think of the marginal probabilities as demands or supplies, and of the joint probability as the fraction of the supplies to be shipped from the production sites to the demand sites, instead of joint probabilities we can think of transportation policies. Clearly, fixing the cost of a transportation policy is equivalent to an integral constraints upon the joint probability. We will show how to solve the cost constrained transportation problem by means of the method of maximum entropy in the mean. We shall also show how this approach leads to an interior point like method to solve the associated linear programming problem. We shall also investigate some geometric structure the space of transportation policies, or joint probabilities or pixel space, using a Riemannian structure associated with the dual of the entropy used to determine bounds between probabilities or between transportation policies.

math.PR↗