arXiv subjects

Huyen Pham

Publications and source records attributed to Huyen Pham.

At least 19 recordsLinked to original sources

arXiv2026

Evaluating Container Orchestration for Neuromorphic Workloads in Virtual Edge Environments

The growing adoption of edge computing has created an increasing need for workloads capable of operating under strict resource and energy constraints. Neuromorphic computing, and spiking neural networks (SNNs) in particular, offers an energy-efficient alternative to conventional machine learning through event-driven computation. However, how SNN workloads behave when deployed within modern container orchestration frameworks, especially in edge environments, remains largely unexplored. This paper investigates the feasibility of deploying and orchestrating SNN workloads in a virtual edge environment using Kubernetes, focusing on end-to-end latency, throughput, classification accuracy, infrastructure overhead, and runtime behavior under concurrent load. Experiments were conducted on a single-node K3d cluster running on a Windows 11 host with WSL2 and Docker Desktop. The results show that SNN workloads are highly sensitive to resource availability. Restricting CPU to 0.5 cores increased median latency by 47.6x and reduced throughput by 49x, while the most constrained configuration failed due to insufficient memory. Classification accuracy remained stable across all working configurations. From an orchestration perspective, K3d successfully deployed and scaled SNN workloads, though its default round-robin routing policy introduced significant tail latency under replica scaling, highlighting a mismatch between stateless load-balancing assumptions and long-running inference workloads. Overall, this study provides a baseline for deploying neuromorphic workloads in containerized edge environments and highlights the importance of resource provisioning and orchestration configuration. Future work should explore improved routing strategies, memory optimization, and validation on physical edge hardware.

cs.DC↗

arXiv2026

Scattered light noise at LIGO Livingston Observatory during O4

Scattered light is one of the most common sources of noise in the LIGO gravitational wave detectors. Light scattering is a highly non-linear process through which motion at low frequencies gets up-converted and creates noise in a higher frequency band in the detector data. From the beginning of the fourth observation run, many glitches appeared in the data of LIGO Livingston detector in the frequency range 10-40 Hz, and the morphology of these glitches suggested that they were produced by scattered light. From our analysis, we identified two different populations of scattered light glitches, one group having higher SNR than the other. The glitches of the high- SNR group were solely modulated by microseismic ground motion (ground motion in 0.1-1.0 Hz) and in this paper, we present models of possible coupling mechanisms for these glitches. We also present results of a statistical correlation analysis based on our models, which indicates that the microseismic ground motion at the corner station along the X direction is the one most correlated with the noise which create these high SNR glitches. After installing baffles very close to the test mass mirrors, we have noticed a significant reduction in the rate and SNR of these glitches. The low-SNR glitches were primarily modulated by high frequency (10-30 Hz) vertical ground motion at the corner station, and this motion was coupling through a specific vacuum chamber at the corner station. After installing an additional seismic isolation platform in that vacuum chamber, these glitches have disappeared.

astro-ph.IM↗

arXiv2026

Geometric indicators of local plasticity in glasses measured by scanning small-beam diffraction

The notion of defects in crystalline phases of matter has been extremely powerful for understanding crystal growth, deformation and melting. Many of these discontinuities in the periodic order of crystals are well described by the Burgers vector, derived from the particle displacements, which encapsulates the direction and magnitude of slip relative to the undeformed state. Since the reference structure of the crystal is known a priori, the Burgers vector can be determined experimentally using both imaging and diffraction methods to measure the final lattice distortion, and thus infer the particle displacements. Glasses have structures that lack the periodicity of crystals, and thus a well-defined reference state. Yet, measurable structural parameters can still be obtained from diffraction from a glass. Here we examine the usefulness of these parameters to probe deformation in glasses. We find that co-ordinated transformations in the centrosymmetry of local particle arrangements are a strong marker of plastic events. For a glass, determining the local distortions corresponding to these plastic events requires measurements before and after deformation. We investigate two geometric indicators that can be derived from these distortions, namely the continuous Burgers vector and the quadrupolar strain. We find that the Burgers vector again emerges as a robust and sensitive metric for understanding local structural transformations due to mechanical deformation, even in disordered glasses.

cond-mat.soft↗

arXiv2025

Customizing Open Source LLMs for Quantitative Medication Attribute Extraction across Heterogeneous EHR Systems

Harmonizing medication data across Electronic Health Record (EHR) systems is a persistent barrier to monitoring medications for opioid use disorder (MOUD). In heterogeneous EHR systems, key prescription attributes are scattered across differently formatted fields and freetext notes. We present a practical framework that customizes open source large language models (LLMs), including Llama, Qwen, Gemma, and MedGemma, to extract a unified set of MOUD prescription attributes (prescription date, drug name, duration, total quantity, daily quantity, and refills) from heterogeneous, site specific data and compute a standardized metric of medication coverage, \emph{MOUD days}, per patient. Our pipeline processes records directly in a fixed JSON schema, followed by lightweight normalization and cross-field consistency checks. We evaluate the system on prescription level EHR data from five clinics in a national OUD study (25{,}605 records from 1{,}257 patients), using a previously annotated benchmark of 10{,}369 records (776 patients) as the ground truth. Performance is reported as coverage (share of records with a valid, matchable output) and record-level exact-match accuracy. Larger models perform best overall: Qwen2.5-32B achieves \textbf{93.4\%} coverage with \textbf{93.0\%} exact-match accuracy across clinics, and MedGemma-27B attains \textbf{93.1\%}/\textbf{92.2\%}. A brief error review highlights three common issues and fixes: imputing missing dosage fields using within-drug norms, handling monthly/weekly injectables (e.g., Vivitrol) by setting duration from the documented schedule, and adding unit checks to prevent mass units (e.g., ``250 g'') from being misread as daily counts. By removing brittle, site-specific ETL and supporting local, privacy-preserving deployment, this approach enables consistent cross-site analyses of MOUD exposure, adherence, and retention in real-world settings.

cs.AI↗

arXiv2025

Nonlinear Graphon mean-field systems

We address a system of weakly interacting particles where the heterogenous connections among the particles are described by a graph sequence and the number of particles grows to infinity. Our results extend the existing law of large numbers and propagation of chaos results to the case where the interaction between one particle and its neighbors is expressed as a nonlinear function of the local empirical measure. In the limit of the number of particles which tends to infinity, if the graph sequence converges to a graphon, then we show that the limit system is described by an infinite collection of processes and can be seen as a process in a suitable $L^2$ space constructed via a Fubini extension. The proof is built on decoupling techniques and careful estimates of the Wasserstein distance.

math.PR↗

arXiv2023

Policy gradient learning methods for stochastic control with exit time and applications to share repurchase pricing

We develop policy gradients methods for stochastic control with exit time in a model-free setting. We propose two types of algorithms for learning either directly the optimal policy or by learning alternately the value function (critic) and the optimal control (actor). The use of randomized policies is crucial for overcoming notably the issue related to the exit time in the gradient computation. We demonstrate the effectiveness of our approach by implementing our numerical schemes in the application to the problem of share repurchase pricing. Our results show that the proposed policy gradient methods outperform PDE or other neural networks techniques in a model-based setting. Furthermore, our algorithms are flexible enough to incorporate realistic market conditions like e.g. price impact or transaction costs.

q-fin.CP↗

arXiv2022

Optimal Consumption with Reference to Past Spending Maximum

This paper studies the infinite-horizon optimal consumption with a path-dependent reference under exponential utility. The performance is measured by the difference between the nonnegative consumption rate and a fraction of the historical consumption maximum. The consumption running maximum process is chosen as an auxiliary state process, and hence the value function depends on two state variables. The Hamilton-Jacobi-Bellman (HJB) equation can be heuristically expressed in a piecewise manner across different regions to take into account all constraints. By employing the dual transform and smooth-fit principle, some thresholds of the wealth variable are derived such that a classical solution to the HJB equation and the feedback optimal investment and consumption strategies can be obtained in closed form in each region. A complete proof of the verification theorem is provided, and numerical examples are presented to illustrate some financial implications.

q-fin.MF↗

arXiv2021

Portfolio diversification and model uncertainty: a robust dynamic mean-variance approach

This paper focuses on a dynamic multi-asset mean-variance portfolio selection problem under model uncertainty. We develop a continuous time framework for taking into account ambiguity aversion about both expected return rates and correlation matrix of the assets, and for studying the join effects on portfolio diversification. The dynamic setting allows us to consider time varying ambiguity sets, which include the cases where the drift and correlation are estimated on a rolling window of historical data or when the investor takes into account learning on the ambiguity. In this context, we prove a general separation principle for the associated robust control problem, which allows us to reduce the determination of the optimal dynamic strategy to the parametric computation of the minimal risk premium function. Our results provide a justification for under-diversification, as documented in empirical studies and in the static models [16], [34]. Furthermore, we explicitly quantify the degree of under-diversification in termsof correlation bounds and Sharpe ratios proximities, and emphasize the different features induced by drift and correlation ambiguity. In particular, we show that an investor with a poor confidence in the expected return estimation does not hold any risky asset, and on the other hand, trades only one risky asset when the level of ambiguity on correlation matrix is large. We also provide a complete picture of the diversification for the optimal robust portfolio in the three-asset case JEL Classification: G11, C61 MSC Classification: 91G10, 91G80, 60H30

q-fin.PM↗

arXiv2021

Approximation error analysis of some deep backward schemes for nonlinear PDEs

Recently proposed numerical algorithms for solving high-dimensional nonlinear partial differential equations (PDEs) based on neural networks have shown their remarkable performance. We review some of them and study their convergence properties. The methods rely on probabilistic representation of PDEs by backward stochastic differential equations (BSDEs) and their iterated time discretization. Our proposed algorithm, called deep backward multistep scheme (MDBDP), is a machine learning version of the LSMDP scheme of Gobet, Turkedjiev (Math. Comp. 85, 2016). It estimates simultaneously by backward induction the solution and its gradient by neural networks through sequential minimizations of suitable quadratic loss functions that are performed by stochastic gradient descent. Our main theoretical contribution is to provide an approximation error analysis of the MDBDP scheme as well as the deep splitting (DS) scheme for semilinear PDEs designed in Beck, Becker, Cheridito, Jentzen, Neufeld (2019). We also supplement the error analysis of the DBDP scheme of Hur{é}, Pham, Warin (Math. Comp. 89, 2020). This yields notably convergence rate in terms of the number of neurons for a class of deep Lipschitz continuous GroupSort neural networks when the PDE is linear in the gradient of the solution for the MDBDP scheme, and in the semilinear case for the DBDP scheme. We illustrate our results with some numerical tests that are compared with some other machine learning algorithms in the literature.

math.AP↗

arXiv2021

Neural networks-based backward scheme for fully nonlinear PDEs

We propose a numerical method for solving high dimensional fully nonlinear partial differential equations (PDEs). Our algorithm estimates simultaneously by backward time induction the solution and its gradient by multi-layer neural networks, while the Hessian is approximated by automatic differentiation of the gradient at previous step. This methodology extends to the fully nonlinear case the approach recently proposed in \cite{HPW19} for semi-linear PDEs. Numerical tests illustrate the performance and accuracy of our method on several examples in high dimension with nonlinearity on the Hessian term including a linear quadratic control problem with control on the diffusion coefficient, Monge-Amp{è}re equation and Hamilton-Jacobi-Bellman equation in portfolio optimization.

math.OC↗

arXiv2020

Deep neural networks algorithms for stochastic control problems on finite horizon: numerical applications

This paper presents several numerical applications of deep learning-based algorithms that have been introduced in [HPBL18]. Numerical and comparative tests using TensorFlow illustrate the performance of our different algorithms, namely control learning by performance iteration (algorithms NNcontPI and ClassifPI), control learning by hybrid iteration (algorithms Hybrid-Now and Hybrid-LaterQ), on the 100-dimensional nonlinear PDEs examples from [EHJ17] and on quadratic backward stochastic differential equations as in [CR16]. We also performed tests on low-dimension control problems such as an option hedging problem in finance, as well as energy storage problems arising in the valuation of gas storage and in microgrid management. Numerical results and comparisons to quantization-type algorithms Qknn, as an efficient algorithm to numerically solve low-dimensional control problems, are also provided; and some corresponding codes are available on https://github.com/comeh/.

math.OC↗

arXiv2019

Integral operator Riccati equations arising in stochastic Volterra control problems

We establish existence and uniqueness for infinite dimensional Riccati equations taking values in the Banach space L 1 ($μ$ $\otimes$ $μ$) for certain signed matrix measures $μ$ which are not necessarily finite. Such equations can be seen as the infinite dimensional analogue of matrix Riccati equations and they appear in the Linear-Quadratic control theory of stochastic Volterra equations.

math.OC↗

arXiv2018

Linear-Quadratic McKean-Vlasov Stochastic Differential Games

We consider a multi-player stochastic differential game with linear McKean-Vlasov dynamics and quadratic cost functional depending on the variance and mean of the state and control actions of the players in open-loop form. Finite and infinite horizon problems with possibly some random coefficients as well as common noise are addressed. We propose a simple direct approach based on weak martingale optimality principle together with a fixed point argument in the space of controls for solving this game problem. The Nash equilibria are characterized in terms of systems of Riccati ordinary differential equations and linear mean-field backward stochastic differential equations: existence and uniqueness conditions are provided for such systems. Finally, we illustrate our results on a toy example.

math.PR↗

arXiv2018

Zero-sum stochastic differential games of generalized McKean-Vlasov type *

We study zero-sum stochastic differential games where the state dynamics of the two players is governed by a generalized McKean-Vlasov (or mean-field) stochastic differential equation in which the distribution of both state and controls of each player appears in the drift and diffusion coefficients, as well as in the running and terminal payoff functions. We prove the dynamic programming principle (DPP) in this general setting, which also includes the control case with only one player, where it is the first time that DPP is proved for open-loop controls. We also show that the upper and lower value functions are viscosity solutions to a corresponding upper and lower Master Bellman-Isaacs equation. Our results extend the seminal work of Fleming and Souganidis [15] to the McKean-Vlasov setting.

math.PR↗

arXiv2016

Robust feedback switching control: dynamic programming and viscosity solutions

We consider a robust switching control problem. The controller only observes the evolution of the state process, and thus uses feedback (closed-loop) switching strategies, a non standard class of switching controls introduced in this paper. The adverse player (nature) chooses open-loop controls that represent the so-called Knightian uncertainty, i.e., misspecifications of the model. The (half) game switcher versus nature is then formulated as a two-step (robust) optimization problem. We develop the stochastic Perron method in this framework, and prove that it produces a viscosity sub and supersolution to a system of Hamilton-Jacobi-Bellman (HJB) variational inequalities, which envelope the value function. Together with a comparison principle, this characterizes the value function of the game as the unique viscosity solution to the HJB equation, and shows as a byproduct the dynamic programming principle for robust feedback switching control problem.

math.PR↗

arXiv2014

Optimal switching for pairs trading rule: a viscosity solutions approach

This paper studies the problem of determining the optimal cut-off for pairs trading rules. We consider two correlated assets whose spread is modelled by a mean-reverting process with stochastic volatility, and the optimal pair trading rule is formulated as an optimal switching problem between three regimes: flat position (no holding stocks), long one short the other and short one long the other. A fixed commission cost is charged with each transaction. We use a viscosity solutions approach to prove the existence and the explicit characterization of cut-off points via the resolution of quasi-algebraic equations. We illustrate our results by numerical simulations.

q-fin.MF↗

arXiv2014

Long time asymptotics for fully nonlinear Bellman equations: a Backward SDE approach

We study the large time behavior of solutions to fully nonlinear parabolic equations of Hamilton-Jacobi-Bellman type arising typically in stochastic control theory with control both on drift and diffusion coefficients. We prove that, as time horizon goes to infinity, the long run average solution is characterized by a nonlinear ergodic equation. Our results hold under dissipativity conditions, and without any nondegeneracy assumption on the diffusion term. Our approach uses mainly probabilistic arguments relying on new backward SDE representation for nonlinear parabolic, elliptic and ergodic equations.

math.PR↗

arXiv2014

Feynman-Kac representation of fully nonlinear PDEs and applications

The classical Feynman-Kac formula states the connection between linear parabolic partial differential equations (PDEs), like the heat equation, and expectation of stochastic processes driven by Brownian motion. It gives then a method for solving linear PDEs by Monte Carlo simulations of random processes. The extension to (fully)nonlinear PDEs led in the recent years to important developments in stochastic analysis and the emergence of the theory of backward stochastic differential equations (BSDEs), which can be viewed as nonlinear Feynman-Kac formulas. We review in this paper the main ideas and results in this area, and present implications of these probabilistic representations for the numerical resolution of nonlinear PDEs, together with some applications to stochastic control problems and model uncertainty in finance.

math.PR↗