arXiv subjects

Phuong Mai Dinh

Publications and source records attributed to Phuong Mai Dinh.

12 recordsLinked to original sources

arXiv2026

GaussianPSL: Soft partitioning for complex PSL problem

Many practical applications of multi-objective optimization (MOO), including engineering design, autonomous systems, and machine learning, often yield complex Pareto frontiers (e.g., discontinuous, degenerate, or non-convex), which pose challenges for traditional scalarization and Pareto Set Learning (PSL) methods that struggle to approximate them accurately. In this paper, we propose GaussianPSL, a novel framework that uses soft partitions of the Pareto decision/objective space to address the challenges posed by complex Pareto frontiers. Our method dynamically partitions the space, enabling simple MLP networks to learn localized features within each region and then aggregate this information for the final prediction. This partition-aware strategy enhances both exploration and convergence, reduces sensitivity to initialization, and improves robustness against local optima. Experimental results demonstrate that the proposed approach consistently outperforms standard PSL models in learning complex Pareto fronts while maintaining model simplicity. Overall, GaussianPSL offers a new direction for effective, scalable MOO in challenging frontier geometries.

cs.LG↗

arXiv2025

Improving Pareto Set Learning for Expensive Multi-objective Optimization via Stein Variational Hypernetworks

Expensive multi-objective optimization problems (EMOPs) are common in real-world scenarios where evaluating objective functions is costly and involves extensive computations or physical experiments. Current Pareto set learning methods for such problems often rely on surrogate models like Gaussian processes to approximate the objective functions. These surrogate models can become fragmented, resulting in numerous small uncertain regions between explored solutions. When using acquisition functions such as the Lower Confidence Bound (LCB), these uncertain regions can turn into pseudo-local optima, complicating the search for globally optimal solutions. To address these challenges, we propose a novel approach called SVH-PSL, which integrates Stein Variational Gradient Descent (SVGD) with Hypernetworks for efficient Pareto set learning. Our method addresses the issues of fragmented surrogate models and pseudo-local optima by collectively moving particles in a manner that smooths out the solution space. The particles interact with each other through a kernel function, which helps maintain diversity and encourages the exploration of underexplored regions. This kernel-based interaction prevents particles from clustering around pseudo-local optima and promotes convergence towards globally optimal solutions. Our approach aims to establish robust relationships between trade-off reference vectors and their corresponding true Pareto solutions, overcoming the limitations of existing methods. Through extensive experiments across both synthetic and real-world MOO benchmarks, we demonstrate that SVH-PSL significantly improves the quality of the learned Pareto set, offering a promising solution for expensive multi-objective optimization problems.

cs.LG↗

arXiv2022

Unexpected dipole instabilities in small molecules after ultrafast XUV irradiation

We investigate the depletion of single-electron states in small molecules under the influence of very short XUV pulses. In N$_2$, for a certain window of XUV energies around 50 eV, we observe a marked occupation inversion, i.e. a situation where depletion of the deepest bound valence electron state is much larger than for any other state. This represents a realistic mechanism which is able to cut, almost instantaneously, a hole into a deep lying state, a situation which is often assumed ad hoc in numerous theoretical studies of energetic ultrafast processes. This occupation inversion furthermore drives a dipole instability, i.e. a spontaneous reappearance of the dipole signal long after the laser pulse is over and the dipole signal has died out. The dipole signal that emerges from this instability can be identified as a particular low-energy structure in photo-electron spectra.

physics.atm-clus↗

arXiv2016

Stochastic TDHF in an exactly solvable model

We apply in a schematic model a theory beyond mean-field, namely Stochastic Time-Dependent Hartree-Fock (STDHF), which includes dynamical electron-electron collisions on top of an incoherent ensemble of mean-field states by occasional 2-particle-2-hole ($2p2h$) jumps. The model considered here is inspired by a Lipkin-Meshkov-Glick model of $Ω$ particles distributed into two bands of energy and coupled by a two-body interaction. Such a model can be exactly solved (numerically though) for small $Ω$. It therefore allows a direct comparison of STDHF and the exact propagation. The systematic impact of the model parameters as the density of states, the excitation energy and the bandwidth is presented and discussed. The time evolution of the STDHF compares fairly well with the exact entropy, as soon as the excitation energy is sufficiently large to allow $2p2h$ transitions. Limitations concerning low energy excitations and memory effects are also discussed.

physics.atm-clus↗

arXiv2016

Strong-field effects in the photo-emission spectrum of the C$_{60}$ fullerene

Considering C$_{60}$ as a model system for describing field emission from the extremity of a carbon nanotip, we explore electron emission from this fullerene excited by an intense, near-infrared, few-cycle laser pulse ($10^{13}$-$10^{14}~{\rm W/cm}^2$, 912 nm, 8-cycle). To this end, we use time-dependent density functional theory augmented by a self-interaction correction. The ionic background of C$_{60}$ is described by a soft jellium model. Particular attention is paid to the high energy electrons. Comparing the spectra at different emission angles, we find that, as a major result of this study, the photoelectrons are strongly peaked along the laser polarization axis forming a highly collimated electron beam in the forward direction, especially for the high energy electrons. Moreover, the high-energy plateau cut-off found in the simulations agrees well with estimates from the classical three-step model. We also investigate the build-up of the high-energy part of a photoelectron spectrum by a time-resolved analysis. In particular, the modulation on the plateau can be interpreted as contributions from intracycle and intercycle interferences.

physics.atm-clus↗

arXiv2001

Thermalization and elliptic flow at RHIC

We discuss signatures of thermalisation in heavy ion collisions based on elliptic flow. We then propose a new method to analyse elliptic flow, based on multiparticle azimuthal correlations. This method allows one to test quantitatively the collective behaviour of the interacting system.

hep-ph↗

arXiv2001

Flow analysis from cumulants: a practical guide

We have recently proposed a new method of flow analysis, based on a cumulant expansion of multiparticle azimuthal correlations. Here, we describe the practical implementation of the method. The major improvement over traditional methods is that the cumulant expansion eliminates order by order correlations not due to flow, which are often large but usually neglected.

nucl-ex↗

arXiv2001

Flow analysis from multiparticle azimuthal correlations

We present a new method for analyzing directed and elliptic flow in heavy ion collisions. Unlike standard methods, it separates the contribution of flow to azimuthal correlations from contributions due to other effects. The separation relies on a cumulant expansion of multiparticle azimuthal correlations, and includes corrections for detector inefficiencies. This new method allows the measurement of the flow of identified particles in narrow phase-space regions, and can be used in every regime, from intermediate to ultrarelativistic energies.

nucl-th↗

arXiv2001

New method for measuring azimuthal distributions in nucleus-nucleus collisions

The methods currently used to measure azimuthal distributions of particles in heavy ion collisions assume that all azimuthal correlations between particles result from their correlation with the reaction plane. However, other correlations exist, and it is safe to neglect them only if azimuthal anisotropies are much larger than 1/sqrt(N), with N the total number of particles emitted in the collision. This condition is not satisfied at ultrarelativistic energies. We propose a new method, based on a cumulant expansion of multiparticle azimuthal correlations, which allows to measure much smaller values of azimuthal anisotropies, down to 1/N. It is simple to implement and can be used to measure both integrated and differential flow. Furthermore, this method automatically eliminates the major systematic errors, which are due to azimuthal asymmetries in the detector acceptance.

nucl-th↗

arXiv2001

J/Psi suppression in central Pb-Pb collisions

We discuss the recent NA50 J/Psi production data in Pb--Pb collisions, in particular the second drop at high transverse energies which correspond to the most central collisions. Using a model which relates the J/Psi suppression to the local energy density, we show that the data can be explained by taking into account transverse energy fluctuations at a given impact parameter. Predictions of this model for RHIC are briefly discussed.

nucl-th↗

arXiv2000

Collective flow and multiparticle azimuthal correlations

The measurement of azimuthal distributions in nucleus-nucleus collisions relies upon the assumption that azimuthal correlations between particles result solely from their correlation with the reaction plane (i.e. flow). We show that at SPS energies, the ansatz is no longer valid, and two-particle correlations due to momentum conservation, final state interactions or resonance decays become of the same order as those arising from flow. This leads us to introduce new methods to analyse collective flow, based on a cumulant expansion, which enable to extract smaller values than those accessible to the standard analysis.

nucl-th↗

arXiv2000

Transverse energy fluctuations and the pattern of J/Psi suppression in Pb-Pb collisions

The NA50 collaboration has recently observed that the J/psi production rate in Pb-Pb collisions decreases more rapidly as a function of the transverse energy for the most central collisions than for less central ones. We show that this phenomenon can be understood as an effect of transverse energy fluctuations in central collisions. A good fit of the data is obtained using a model which relates J/psi suppression to the local energy density. Our results suggest that the J/psi is completely suppressed at the highest densities achieved in Pb-Pb collisions.

nucl-th↗