arXiv subjects

Haijiang Li

Publications and source records attributed to Haijiang Li.

5 recordsLinked to original sources

arXiv2026

VERPO: Verified Evidence Regularized Policy Optimization

Verifiable outcome rewards guide language-model post-training, but sequence-level advantages do not identify which token-level decisions should be preserved or revised. Evidence-conditioned Teachers provide denser supervision by replaying sampled trajectories with privileged feedback. Yet indiscriminate imitation risks transferring formatting or reasoning-style shifts that do not support task success. We introduce VERPO, a Verified Evidence Regularized Policy Optimization framework that treats evidence as a proposal for policy correction while retaining the outcome objective. It separates evidence-free reference restoration from signed token-level evidence corrections. Fisher Evidence Contrast attenuates corrections along an estimated evidence-presence direction. A stopped token-wise ZPD controller scales acceptance according to local reward alignment and Fisher movement cost, while the reference channel remains independent of acceptance. Across five scientific-reasoning and tool-use tasks, the best variant on each backbone exceeds the strongest compared baseline in average score. The averages rise from 0.6826 to 0.6857 on Qwen3-4B, from 0.6895 to 0.7058 on Qwen3-8B, and from 0.4751 to 0.5657 on Llama-3.2-1B.

cs.LG↗

arXiv2026

A Framework Integrating the Dynamic Stiffness Matrix with Physics-Informed Neural Networks for Solving Eigenvalue Problems and Analysing Dynamic Response

This paper introduces a framework that integrates the dynamic stiffness matrix (DSM) with physics-informed neural networks (PINN). The DSM-PINN embeds physical constraints within the model and demonstrates robustness, particularly when addressing limited datasets across diverse investigations. In this approach, deep neural network outputs approximate the displacement fields of element nodes. Unlike the finite element method (FEM), the element shape functions are homogeneous solutions to the governing partial differential equation, forming the basis of the exact dynamic stiffness matrix, thereby avoiding high-order derivative terms. This matrix also serves as a frequency-domain spectral element, resulting in a strong-form PINN. The loss function is produced by connecting neural networks with dynamic stiffness matrices. We focus on utilising PINNs to resolve eigenvalue problems by employing the Wittrick-Williams algorithm, which overcomes the challenge of neural networks failing to converge to higher-order eigenvalues. Additionally, the frequency domain-PINN method is used to analyse structural dynamic responses under moving and impulsive loads, addressing the limitation of neural networks in handling complex numbers. Theoretical convergence stability of the suggested approach is also analysed even DSM is an indefinite matrix after implementing the boundary condition. The numerical results validate the practicality and efficacy of the recommended approach.

math.NA↗

arXiv2025

MedGEN-Bench: A Contextually Entangled Benchmark for Open-ended Multimodal Medical Generation

Medical vision-language models (VLMs) are increasingly expected to support clinical workflows through diagnostic text and relevant medical images. However, current medical visual benchmarks have three recurring limitations: query-image misalignment from queries weakly grounded in specific image instances, closed-ended formats that narrow answer space and encourage shortcut-based prediction, and text-centric output paradigms that limit evaluation of image-generation and image-editing capabilities. We introduce MedGEN-Bench, a benchmark for open-ended multimodal medical generation. The evaluation snapshot reported in this manuscript comprises 6,422 image-text pairs reviewed by clinical experts and models, spanning 6 canonical imaging modalities, 15 clinical tasks, and 27 named subtasks. It includes 1,100 Visual Question Answering (VQA) pairs, 3,872 Image Editing pairs, and 1,450 Contextual Multimodal Generation pairs. MedGEN-Bench centers on contextual entanglement: dependence of an instruction's intended output on the particular image instance rather than on task wording alone. The benchmark operationalizes this concept through image-grounded instructions and extends evaluation to open-ended multimodal outputs. Its tiered evaluation protocol combines reproducible reference-based fidelity and similarity measures with a structured, checklist-guided assessment by a medical VLM judge. We evaluate 10 compositional frameworks, 2 dedicated image-editing models, 3 unified models, and 5 VLMs. The results show image-output tasks remain unsaturated. Contextual augmentation increases mean image-instruction similarity from 0.273 to 0.372, while a 1,000-case medical-expert audit shows moderate agreement between judge scores and clinician ratings. Source code and dataset are available at https://yangjj007.github.io/medgen.

cs.CV↗

arXiv2024

Cross-Domain Comparative Analysis of Digital Twins and Universalised Solutions

Digitalisation is one of the main drivers of most economic sectors nowadays and the digital twin, as a reification of digitalisation for complex systems has attracted much attention from both academics and industry. There have been studies focusing on digital twins in a specific sector while there are few exercising insightful comparisons of digital twins from different domains. Considering the digital twinning is a cross-domain transformation, it is beneficial to establish the principles of universality and variation that can explain similarities and differences in any digital twins. This paper first delivers a comparative analysis of digital twins in five domains through a six-dimensional characterisation framework. Then, by departing from the correlations among the domain-specific DT development, a cross-domain Digital Twin Platform-as-a-Service (DT-PaaS) is proposed to universalise the common process, tools and applications, meanwhile being inclusive of variations of every digital twin instance. As a centralised data, modeling and service platform, it is expected to break the barriers between domains by enabling the cross-domain digital twin data sharing, interoperability and development synergy and tackle some complex global challenges such as climate challenge, net zero, pandemics, etc.

eess.SY↗

arXiv2017

Microwave Vortex-Beam Emitter Based on Spoof Surface Plasmon Polaritons

Since the orbital angular momentum (OAM) being investigated intensively in the optical region, there are growing interests in employing OAM to solve the problem in wireless communications as a new method. It is found that the independence between different OAM modes is crucial to wireless communications. Motivated by the tremendous potential of OAM in communication systems, we propose a novel method to generate vortex beams by spoof surface plasmon polaritons (SPPs). A looped double-layer spoof SPP waveguide is applied to realize the transmission of electromagnetic waves. Beam emitting is accomplished through a series of circular patches, whose role is not only the radiation units but also resonators giving rise to the phase shifts required by the vortex beam. The proposed method is validated by both numerical calculation and experiment. The measured results show that the spoof SPPs are radiated by the circular patches and the vortex beam carrying different OAM modes are observed at different frequencies. This is the first time to generate the OAM modes by spoof SPPs. The proposed method possesses smaller size and is much easier to be integrated into integrated circuits. The simple structure and design procedure make the proposed method promising in future wireless communication systems.

physics.class-ph↗