Publications

Contribution.Introduces transfer learning for dynamic multiobjective optimization, so search can reuse prior environments.

Why read it.This paper pioneered the integration of transfer learning into evolutionary dynamic multiobjective optimization, helping establish a research direction focused on knowledge transfer across changing environments.

Contribution.Defines evolutionary transfer optimization as a field and gives it a shared agenda.

Why read it.If your research touches transfer learning, surrogate modeling, multi-task optimization, or expensive black-box problems, this IEEE CIM article is the canonical starting point and the source most authors cite to motivate their work.

Contribution.Training-free diffusion that models how the Pareto set of a dynamic multiobjective problem evolves.

Contribution.Recasts dynamic multiobjective optimization as a prompt-guided time-series prediction problem and

Why read it.This paper opens a new line of work at the intersection of LLMs and dynamic multiobjective optimization, moving beyond handcrafted transfer heuristics toward prompt-based prediction.

Contribution.Embeds physics in graph networks for stable long-horizon multiphysics simulation.

Contribution.Uses T-NNGP to produce interpretable multi-physics PDE solutions.

Contribution.Produces interpretable PDE solutions rather than opaque black-box approximations.

Contribution.Makes neural operators respect interface physics in multiphase flow simulation.

Contribution.Infers interpretable intrinsic dynamics from video with bilevel optimization.

Contribution.Geometry-aware, anisotropic boundary correction for aerodynamic simulation.

Contribution.Finds transferable perturbations that break person re-ID across RGB and infrared.

Why read it.By combining gradients from RGB and infrared modalities with a grayscale bridge, the method learns a transferable perturbation that degrades re-ID performance across modalities while remaining visually subtle.

Contribution.Turns image-to-text attacks into a decision-based black-box method with no gradients.

Contribution.Exposes a universal black-box weakness in 3DGS watermarking systems.

Contribution.Uses alignment augmentation and knowledge transfer to bridge sketch/RGB re-ID with few labels.

Why read it.Alignment Augmentation and a Knowledge Transfer Catalyst are used to improve domain alignment and boost cross-modal generalization.

Contribution.Turns fairness diagnostics into actionable explanations for graph neural networks.

Contribution.Stabilizes graph meta-learning by calibrating manifolds with proxy subgraphs.