arXiv subjects

Junda Zhu

Publications and source records attributed to Junda Zhu.

7 recordsLinked to original sources

arXiv2026

Impact of mode completeness on the accuracy of the coupling theory of quasinormal modes: a strict numerical demonstration

The coupling theory of quasinormal modes (QNMs) for a coupled system of generally lossy and dispersive optical nanoresonators has been established in a rigorous manner based on the first principle of Maxwell's equations [Phys. Rev. B 102, 045430 (2020)], and can achieve superior computational efficiency and physical intuitiveness compared with full-wave numerical methods if a small set of modes can achieve a high accuracy. The QNMs suffer from an exponential divergence of far field and can form a complete basis inside but not outside the resonator. In the QNM coupling theory (QCT), it is required that the QNMs of each resonator form a complete basis in expanding the scattered field both inside and outside the resonator, which can be achieved by using regularized QNMs (RQNMs). However, a strict numerical demonstration of the impact of the mode completeness of RQNMs on the accuracy of QCT by using a virtually complete basis of RQNMs is still absent. In this paper, we will provide such a numerical demonstration along with an improvement of the QCT and some theoretical demonstrations on a rigorous incorporation of RQNMs into the QCT. The RQNMs are obtained by introducing an equivalent surface current (ESC) encircling the resonator (called ESC-RQNMs) or the perfectly matched layer (PML) surrounding the computational domain (called PML-RQNMs). The numerical example is selected as two one-dimensional resonators of slabs in the extreme coupling case of direct contact, for which a virtually complete basis of RQNMs can be solved either analytically (for ESC-RQNMs) or numerically (for PML-RQNMs). The results show that by using a virtually complete basis of RQNMs, the QCT can achieve a high accuracy in predicting both the source-free eigenmodes and the source-excited scattered field of the coupled system, which is not true if using the incomplete basis of not-regularized QNMs (i.e., physical QNMs).

physics.optics↗

arXiv2025

Reasoning-to-Defend: Safety-Aware Reasoning Can Defend Large Language Models from Jailbreaking

Large Reasoning Models (LRMs) have recently demonstrated impressive performances across diverse domains. However, how the safety of Large Language Models (LLMs) benefits from enhanced reasoning capabilities against jailbreak queries remains unexplored. To bridge this gap, in this paper, we propose Reasoning-to-Defend (R2D), a novel training paradigm that integrates a safety-aware reasoning mechanism into LLMs' generation process. This enables self-evaluation at each step of the reasoning process, forming safety pivot tokens as indicators of the safety status of responses. Furthermore, in order to improve the accuracy of predicting pivot tokens, we propose Contrastive Pivot Optimization (CPO), which enhances the model's perception of the safety status of given dialogues. LLMs dynamically adjust their response strategies during reasoning, significantly enhancing their safety capabilities defending jailbreak attacks. Extensive experiments demonstrate that R2D effectively mitigates various attacks and improves overall safety, while maintaining the original performances. This highlights the substantial potential of safety-aware reasoning in improving robustness of LRMs and LLMs against various jailbreaks.

cs.CL↗

arXiv2025

Local perfect chirality at reflection-zeros away from exceptional points in optical whispering gallery microcavity

Recently, a local and imperfect chirality of the resonant eigenmode at the exceptional point (EP) has been reported in the optical whispering gallery microcavity system perturbed by two strong nanoscatterers [Phys. Rev. A 108, L041501 (2023)]. Here, we discover a local perfect chirality of the resonant eigenmode away from the EP in the parameter space of the strongly perturbed microcavity system. By considering the multiple scattering process of the azimuthally propagating modes (APMs) at the nanoscatterers with a first-principles-based model, the local perfect chirality is predicted to result from the unidirectional reflectionlessness, i.e., the reflection-zero (R-zero) of the APMs at the two nanoscatterers. Numerical results and model predictions consistently show that the structural parameters of the R-zero typically deviate from those of the EP, which means that the pair of split resonant eigenmodes at the R-zero have different complex resonance frequencies and electromagnetic fields. In general, only one of the pair of split eigenmodes exhibits a local perfect chirality within the local azimuthal range divided by the two nanoscatterers. With the decrease of the two nanoscatterers' sizes or their relative azimuthal angle, the R-zero tends to coincide with the EP.

physics.optics↗

arXiv2025

DiffusionAttacker: Diffusion-Driven Prompt Manipulation for LLM Jailbreak

Large Language Models (LLMs) are susceptible to generating harmful content when prompted with carefully crafted inputs, a vulnerability known as LLM jailbreaking. As LLMs become more powerful, studying jailbreak methods is critical to enhancing security and aligning models with human values. Traditionally, jailbreak techniques have relied on suffix addition or prompt templates, but these methods suffer from limited attack diversity. This paper introduces DiffusionAttacker, an end-to-end generative approach for jailbreak rewriting inspired by diffusion models. Our method employs a sequence-to-sequence (seq2seq) text diffusion model as a generator, conditioning on the original prompt and guiding the denoising process with a novel attack loss. Unlike previous approaches that use autoregressive LLMs to generate jailbreak prompts, which limit the modification of already generated tokens and restrict the rewriting space, DiffusionAttacker utilizes a seq2seq diffusion model, allowing more flexible token modifications. This approach preserves the semantic content of the original prompt while producing harmful content. Additionally, we leverage the Gumbel-Softmax technique to make the sampling process from the diffusion model's output distribution differentiable, eliminating the need for iterative token search. Extensive experiments on Advbench and Harmbench demonstrate that DiffusionAttacker outperforms previous methods across various evaluation metrics, including attack success rate (ASR), fluency, and diversity.

cs.CL↗

arXiv2024

ATM: Adversarial Tuning Multi-agent System Makes a Robust Retrieval-Augmented Generator

Large language models (LLMs) are proven to benefit a lot from retrieval-augmented generation (RAG) in alleviating hallucinations confronted with knowledge-intensive questions. RAG adopts information retrieval techniques to inject external knowledge from semantic-relevant documents as input contexts. However, since today's Internet is flooded with numerous noisy and fabricating content, it is inevitable that RAG systems are vulnerable to these noises and prone to respond incorrectly. To this end, we propose to optimize the retrieval-augmented Generator with an Adversarial Tuning Multi-agent system (ATM). The ATM steers the Generator to have a robust perspective of useful documents for question answering with the help of an auxiliary Attacker agent through adversarially tuning the agents for several iterations. After rounds of multi-agent iterative tuning, the Generator can eventually better discriminate useful documents amongst fabrications. The experimental results verify the effectiveness of ATM and we also observe that the Generator can achieve better performance compared to the state-of-the-art baselines.

cs.CL↗

arXiv2024

A Survey of Neural Network Robustness Assessment in Image Recognition

In recent years, there has been significant attention given to the robustness assessment of neural networks. Robustness plays a critical role in ensuring reliable operation of artificial intelligence (AI) systems in complex and uncertain environments. Deep learning's robustness problem is particularly significant, highlighted by the discovery of adversarial attacks on image classification models. Researchers have dedicated efforts to evaluate robustness in diverse perturbation conditions for image recognition tasks. Robustness assessment encompasses two main techniques: robustness verification/ certification for deliberate adversarial attacks and robustness testing for random data corruptions. In this survey, we present a detailed examination of both adversarial robustness (AR) and corruption robustness (CR) in neural network assessment. Analyzing current research papers and standards, we provide an extensive overview of robustness assessment in image recognition. Three essential aspects are analyzed: concepts, metrics, and assessment methods. We investigate the perturbation metrics and range representations used to measure the degree of perturbations on images, as well as the robustness metrics specifically for the robustness conditions of classification models. The strengths and limitations of the existing methods are also discussed, and some potential directions for future research are provided.

cs.CV↗

arXiv2022

Imperfect chirality at exceptional points in optical whispering-gallery microcavities

Non-Hermitian systems have attracted considerable attention for their broad impacts on various physical platforms and peculiar applications. In non-Hermitian systems, both eigenvalues and eigenstates simultaneously coalesce at exceptional points (EPs). As one of the remarkable features of EPs, the field chirality is commonly considered perfect, which is utilized as an intriguing feature to control wave propagation and regarded as a criterion of EP. However, in this work, we discover an imperfect chirality of eigenmodes at the EPs in an optical whispering gallery mode (WGM) microcavity perturbed by two strong nanoscatterers. This counterintuitive phenomenon originates from a strong frequency-dependence of the scattering between the counterpropagating waves at an "effective scatterer", which could be explained by a first-principle-based model considering a dynamic multiple-scattering process of the azimuthally propagating modes. We find that the generally imperfect chirality at the EP tends to be globally perfect with the decrease of the scattering effect induced by the nanoscatterers. Furthermore, the chirality also becomes locally perfect with the decrease of the relative azimuthal angle between the two strong nanoscatterers. This work provides a new understanding of the general properties of chirality at EPs. It will benefit the potential applications enabled by the chirality features of non-Hermitian systems at EPs.

physics.optics↗