arXiv subjects

Meng Han

Publications and source records attributed to Meng Han.

At least 73 records · Page 4Linked to original sources

arXiv2025

Local Distance Query with Differential Privacy

Differential Privacy (DP) is commonly employed to safeguard graph analysis or publishing. Distance, a critical factor in graph analysis, is typically handled using curator DP, where a trusted curator holds the complete neighbor lists of all vertices and answers queries privately. However, in many real-world scenarios, such a curator may not be present, posing a significant challenge for implementing differentially private distance queries under Local Differential Privacy (LDP). This paper proposes two approaches to address this challenge. The first approach generates a synthetic graph by randomizing responses and applies bitwise operations to reduce noise interference. However, like other synthetic graph methods, this approach suffers from low utility. To overcome this limitation, we propose a second approach, the first LDP method specifically designed for distance queries, which captures the global graph structure by continuously aggregating local distance vectors from neighboring vertices. This process enables the accurate updating of global distances. We demonstrate the effectiveness of our method through comprehensive theoretical analysis and experimental evaluations on real-world datasets.

cs.CR↗

arXiv2025

Pushing the Limits of Safety: A Technical Report on the ATLAS Challenge 2025

Multimodal Large Language Models (MLLMs) have enabled transformative advancements across diverse applications but remain susceptible to safety threats, especially jailbreak attacks that induce harmful outputs. To systematically evaluate and improve their safety, we organized the Adversarial Testing & Large-model Alignment Safety Grand Challenge (ATLAS) 2025}. This technical report presents findings from the competition, which involved 86 teams testing MLLM vulnerabilities via adversarial image-text attacks in two phases: white-box and black-box evaluations. The competition results highlight ongoing challenges in securing MLLMs and provide valuable guidance for developing stronger defense mechanisms. The challenge establishes new benchmarks for MLLM safety evaluation and lays groundwork for advancing safer multimodal AI systems. The code and data for this challenge are openly available at https://github.com/NY1024/ATLAS_Challenge_2025.

cs.CR↗

arXiv2025

Attosecond Control and Measurement of Chiral Photoionisation Dynamics

Many chirality-sensitive light-matter interactions are governed by chiral electron dynamics. Therefore, the development of advanced technologies harnessing chiral phenomena would critically benefit from measuring and controlling chiral electron dynamics on their natural attosecond time scales. Such endeavors have so far been hampered by the lack of characterized circularly polarized attosecond pulses, an obstacle that has recently been overcome (Han et al. Optica 10 (2023) 1044-1052, Han et al. Nature Physics 19 (2023) 230-236). In this article, we introduce chiroptical spectroscopy with attosecond pulses and demonstrate attosecond coherent control over photoelectron circular dichroism (PECD) (Goetz et al. Physical Review Letters 122 (2019) 013204, Goetz et al. arXiv:2104.07522), as well as the measurement of chiral asymmetries in the forward-backward and angle-resolved photoionisation delays of chiral molecules. We show that co-rotating attosecond and near-infrared pulses can nearly double the PECD and even change its sign compared to single-photon ionisation. We demonstrate that chiral photoionisation delays depend on both polar and azimuthal angles of photoemission in the light-propagation frame, requiring three-dimensional momentum resolution. We measure forward-backward chiral-sensitive delays of up to 120 as and polar-angle-resolved photoionisation delays up to 240 as, which include an asymmmetry of $\sim$60 as originating from chirality in the continuum-continuum transitions. Attosecond chiroptical spectroscopy opens the door to quantitatively understanding and controlling the dynamics of chiral molecules on the electronic time scale.

physics.chem-ph↗

arXiv2025

Direct Behavior Optimization: Unlocking the Potential of Lightweight LLMs

Lightweight Large Language Models (LwLLMs) are reduced-parameter, optimized models designed to run efficiently on consumer-grade hardware, offering significant advantages in resource efficiency, cost-effectiveness, and data privacy. However, these models often struggle with limited inference and reasoning capabilities, which restrict their performance on complex tasks and limit their practical applicability. Moreover, existing prompt optimization methods typically rely on extensive manual effort or the meta-cognitive abilities of state-of-the-art LLMs, making them less effective for LwLLMs. To address these challenges, we introduce DeBoP, a new Direct Behavior Optimization Paradigm, original from the Chain-of-Thought (CoT) prompting technique. Unlike CoT Prompting, DeBoP is an automatic optimization method, which focuses on the optimization directly on the behavior of LwLLMs. In particular, DeBoP transforms the optimization of complex prompts into the optimization of discrete, quantifiable execution sequences using a gradient-free Monte Carlo Tree Search. We evaluate DeBoP on seven challenging tasks where state-of-the-art LLMs excel but LwLLMs generally underperform. Experimental results demonstrate that DeBoP significantly outperforms recent prompt optimization methods on most tasks. In particular, DeBoP-optimized LwLLMs surpass GPT-3.5 on most tasks while reducing computational time by approximately 60% compared to other automatic prompt optimization methods.

cs.CL↗

arXiv2025

Attosecond X-ray Core-level Chronoscopy of Aromatic Molecules

Attosecond photoemission or photoionization delays are a unique probe of the structure and the electronic dynamics of matter. However, spectral congestion and spatial delocalization of valence electron wave functions set fundamental limits to the complexity of systems that can be studied and the information that can be retrieved, respectively. Using attosecond X-ray pulses from LCLS, we demonstrate the key advantages of measuring core-level delays: the photoelectron spectra remain atom-like, the measurements become element specific and the observed scattering dynamics originate from a point-like source. We exploit these unique features to reveal the effects of electronegativity and symmetry on attosecond scattering dynamics by measuring and calculating the photoionization delays between N-1s and C-1s core shells of a series of aromatic azabenzene molecules. Remarkably, the delays increase with the number of nitrogen atoms in the molecule and reveal multiple resonances. We identify two previously unknown mechanisms regulating the associated attosecond dynamics, namely the enhanced confinement of the trapped wavefunction with increasing electronegativity of the atoms and the decrease of the coupling strength among the photoemitted partial waves with increasing symmetry. This study demonstrates the unique opportunities opened by measurements of core-level photoionization delays for unraveling attosecond electron dynamics in complex matter.

physics.chem-ph↗

arXiv2025

CaMDN: Enhancing Cache Efficiency for Multi-tenant DNNs on Integrated NPUs

With the rapid development of DNN applications, multi-tenant execution, where multiple DNNs are co-located on a single SoC, is becoming a prevailing trend. Although many methods are proposed in prior works to improve multi-tenant performance, the impact of shared cache is not well studied. This paper proposes CaMDN, an architecture-scheduling co-design to enhance cache efficiency for multi-tenant DNNs on integrated NPUs. Specifically, a lightweight architecture is proposed to support model-exclusive, NPU-controlled regions inside shared cache to eliminate unexpected cache contention. Moreover, a cache scheduling method is proposed to improve shared cache utilization. In particular, it includes a cache-aware mapping method for adaptability to the varying available cache capacity and a dynamic allocation algorithm to adjust the usage among co-located DNNs at runtime. Compared to prior works, CaMDN reduces the memory access by 33.4% on average and achieves a model speedup of up to 2.56$\times$ (1.88$\times$ on average).

cs.AR↗

arXiv2025

NeuRel-Attack: Neuron Relearning for Safety Disalignment in Large Language Models

Safety alignment in large language models (LLMs) is achieved through fine-tuning mechanisms that regulate neuron activations to suppress harmful content. In this work, we propose a novel approach to induce disalignment by identifying and modifying the neurons responsible for safety constraints. Our method consists of three key steps: Neuron Activation Analysis, where we examine activation patterns in response to harmful and harmless prompts to detect neurons that are critical for distinguishing between harmful and harmless inputs; Similarity-Based Neuron Identification, which systematically locates the neurons responsible for safe alignment; and Neuron Relearning for Safety Removal, where we fine-tune these selected neurons to restore the model's ability to generate previously restricted responses. Experimental results demonstrate that our method effectively removes safety constraints with minimal fine-tuning, highlighting a critical vulnerability in current alignment techniques. Our findings underscore the need for robust defenses against adversarial fine-tuning attacks on LLMs.

cs.LG↗

arXiv2025

FineQ: Software-Hardware Co-Design for Low-Bit Fine-Grained Mixed-Precision Quantization of LLMs

Large language models (LLMs) have significantly advanced the natural language processing paradigm but impose substantial demands on memory and computational resources. Quantization is one of the most effective ways to reduce memory consumption of LLMs. However, advanced single-precision quantization methods experience significant accuracy degradation when quantizing to ultra-low bits. Existing mixed-precision quantization methods are quantized by groups with coarse granularity. Employing high precision for group data leads to substantial memory overhead, whereas low precision severely impacts model accuracy. To address this issue, we propose FineQ, software-hardware co-design for low-bit fine-grained mixed-precision quantization of LLMs. First, FineQ partitions the weights into finer-grained clusters and considers the distribution of outliers within these clusters, thus achieving a balance between model accuracy and memory overhead. Then, we propose an outlier protection mechanism within clusters that uses 3 bits to represent outliers and introduce an encoding scheme for index and data concatenation to enable aligned memory access. Finally, we introduce an accelerator utilizing temporal coding that effectively supports the quantization algorithm while simplifying the multipliers in the systolic array. FineQ achieves higher model accuracy compared to the SOTA mixed-precision quantization algorithm at a close average bit-width. Meanwhile, the accelerator achieves up to 1.79x energy efficiency and reduces the area of the systolic array by 61.2%.

cs.LG↗

arXiv2025

Hearing carrier-envelope offset frequency and phase in air with a microphone

Attosecond science and frequency metrology rely on the precise measurement and control of the laser pulse waveform, a feat traditionally achieved using optoelectronic techniques. In this study, we conducted a laser-induced acoustic experiment in air ionized by carrier-envelope phase (CEP)-stabilized sub-4 femtosecond pulses. Our results reveal that the acoustic signal exhibits CEP dependence in few-cycle pulses, primarily through amplitude modulation from laser-driven ionization. This novel optoacoustic phenomenon enables not only the measurement of the carrier-envelope offset frequency but also the direct characterization of the waveform of optical pulses through a microphone. Our study highlights the potential of laser-induced acoustic waves for advancing frequency metrology and ultrafast science.

physics.optics↗

arXiv2025

Towards Robust and Secure Embodied AI: A Survey on Vulnerabilities and Attacks

Embodied AI systems, including robots and autonomous vehicles, are increasingly integrated into real-world applications, where they encounter a range of vulnerabilities stemming from both environmental and system-level factors. These vulnerabilities manifest through sensor spoofing, adversarial attacks, and failures in task and motion planning, posing significant challenges to robustness and safety. Despite the growing body of research, existing reviews rarely focus specifically on the unique safety and security challenges of embodied AI systems. Most prior work either addresses general AI vulnerabilities or focuses on isolated aspects, lacking a dedicated and unified framework tailored to embodied AI. This survey fills this critical gap by: (1) categorizing vulnerabilities specific to embodied AI into exogenous (e.g., physical attacks, cybersecurity threats) and endogenous (e.g., sensor failures, software flaws) origins; (2) systematically analyzing adversarial attack paradigms unique to embodied AI, with a focus on their impact on perception, decision-making, and embodied interaction; (3) investigating attack vectors targeting large vision-language models (LVLMs) and large language models (LLMs) within embodied systems, such as jailbreak attacks and instruction misinterpretation; (4) evaluating robustness challenges in algorithms for embodied perception, decision-making, and task planning; and (5) proposing targeted strategies to enhance the safety and reliability of embodied AI systems. By integrating these dimensions, we provide a comprehensive framework for understanding the interplay between vulnerabilities and safety in embodied AI.

cs.CR↗

arXiv2025

Probing the ideal limit of interfacial thermal conductance in two-dimensional van der Waals heterostructures

Probing the ideal limit of interfacial thermal conductance (ITC) in two-dimensional (2D) heterointerfaces is of paramount importance for assessing heat dissipation in 2D-based nanoelectronics. Using graphene/hexagonal boron nitride (Gr/$h$-BN), a structurally isomorphous heterostructure with minimal mass contrast, as a prototype, we develop an accurate yet highly efficient machine-learned potential (MLP) model, which drives nonequilibrium molecular dynamics (NEMD) simulations on a realistically large system with over 300,000 atoms, enabling us to report the ideal limit range of ITC for 2D heterostructures at room temperature. We further unveil an intriguing stacking-sequence-dependent ITC hierarchy in the Gr/$h$-BN heterostructure, which can be connected to moiré patterns and is likely universal in van der Waals layered materials. The underlying atomic-level mechanisms can be succinctly summarized as energy-favorable stacking sequences facilitating out-of-plane phonon energy transmission. This work demonstrates that MLP-driven MD simulations can serve as a new paradigm for probing and understanding thermal transport mechanisms in 2D heterostructures and other layered materials.

physics.comp-ph↗

arXiv2025

Differentially Private Distance Query with Asymmetric Noise

With the growth of online social services, social information graphs are becoming increasingly complex. Privacy issues related to analyzing or publishing on social graphs are also becoming increasingly serious. Since the shortest paths play an important role in graphs, privately publishing the shortest paths or distances has attracted the attention of researchers. Differential privacy (DP) is an excellent standard for preserving privacy. However, existing works to answer the distance query with the guarantee of DP were almost based on the weight private graph assumption, not on the paths themselves. In this paper, we consider edges as privacy and propose distance publishing mechanisms based on edge DP. To address the issue of utility damage caused by large global sensitivities, we revisit studies related to asymmetric neighborhoods in DP with the observation that the distance query is monotonic in asymmetric neighborhoods. We formally give the definition of asymmetric neighborhoods and propose Individual Asymmetric Differential Privacy with higher privacy guarantees in combination with smooth sensitivity. Then, we introduce two methods to efficiently compute the smooth sensitivity of distance queries in asymmetric neighborhoods. Finally, we validate our scheme using both real-world and synthetic datasets, which can reduce the error to $0.0862$.

cs.CR↗

arXiv2024

Artifact Correction in Magnetic Resonance Temperature Imaging for Laser Interstitial Thermotherapy with Multi-echo Acquisitions

In MRI-guided laser interstitial thermotherapy (MRgLITT), a signal void sometimes appears at the heating center of the measured temperature map. In neurosurgical MRgLITT treatments, cerebrospinal fluid pulsation (CSF), which may lead to temperature artifacts, also needs to be carefully managed. We find that signal loss in MR magnitude images can be one distinct contributor to the temperature imaging signal void. Therefore, this study aims to investigate this finding and more importantly. Also, this study intends to improve measurement accuracy by correcting CSF-induced temperature errors and employing a more reliable phase unwrapping algorithm. A gradient echo sequence with certain TE values for temperature imaging is used to quantify T2* variations during MRgLITT and to investigate the development of signal voids throughout the treatment. Informed by these findings, a multi-echo GRE sequence with appropriate TE coverage is employed. A multi-echo-based correction algorithm is developed to address the signal loss-induced temperature errors. A new phase unwrapping method and a new CSF pulsation correction approach are developed for multi-echo signal processing. The temperature imaging method is evaluated by gel phantom, ex-vivo, and in-vivo LITT heating experiments. T2* shortening during heating can be one important cause of the temperate imaging signal voids and this demands the multi-echo acquisition with varied TE values. The proposed multi-echo-based method can effectively correct signal loss-induced temperature errors and raise temperature estimation precision. The multi-echo thermometry in the in-vivo experiments shows smoother hotspot boundaries, fewer artifacts, and improved thermometry reliability. In the in-vivo experiments, the ablation areas estimated from the multi-echo thermometry also show satisfactory agreement with those determined from post-ablation MR imaging.

physics.med-ph↗

arXiv2024

CopyrightMeter: Revisiting Copyright Protection in Text-to-image Models

Text-to-image diffusion models have emerged as powerful tools for generating high-quality images from textual descriptions. However, their increasing popularity has raised significant copyright concerns, as these models can be misused to reproduce copyrighted content without authorization. In response, recent studies have proposed various copyright protection methods, including adversarial perturbation, concept erasure, and watermarking techniques. However, their effectiveness and robustness against advanced attacks remain largely unexplored. Moreover, the lack of unified evaluation frameworks has hindered systematic comparison and fair assessment of different approaches. To bridge this gap, we systematize existing copyright protection methods and attacks, providing a unified taxonomy of their design spaces. We then develop CopyrightMeter, a unified evaluation framework that incorporates 17 state-of-the-art protections and 16 representative attacks. Leveraging CopyrightMeter, we comprehensively evaluate protection methods across multiple dimensions, thereby uncovering how different design choices impact fidelity, efficacy, and resilience under attacks. Our analysis reveals several key findings: (i) most protections (16/17) are not resilient against attacks; (ii) the "best" protection varies depending on the target priority; (iii) more advanced attacks significantly promote the upgrading of protections. These insights provide concrete guidance for developing more robust protection methods, while its unified evaluation protocol establishes a standard benchmark for future copyright protection research in text-to-image generation.

cs.CR↗

arXiv2024

Analytical Expression for Continuum-continuum Transition Amplitude of Hydrogen-like Atoms with Angular-momentum Dependence

Attosecond chronoscopy typically utilises interfering two-photon transitions to access the phase information. Simulating these two-photon transitions is challenging due to the continuum-continuum transition term. The hydrogenic approximation within second-order perturbation theory has been widely used due to the existence of analytical expressions of the wave functions. So far, only (partially) asymptotic results have been derived, which fail to correctly describe the low-kinetic-energy behaviour, especially for high angular-momentum states. Here, we report an analytical expression that overcome these limitations. They are based on the Appell's F1 function and use the confluent hypergeometric function of the second kind as the intermediate states. We show that the derived formula quantitatively agrees with the numerical simulations using the time-dependent Schr{ö}dinger equation for various angular-momentum states, which improves the accuracy compared to the other analytical approaches that were previously reported. Furthermore, we give an angular-momentum-dependent asymptotic form of the outgoing wavefunction and their continuum-continuum dipole transition amplitudes.

physics.atom-ph↗

arXiv2024

GenTel-Safe: A Unified Benchmark and Shielding Framework for Defending Against Prompt Injection Attacks

Large Language Models (LLMs) like GPT-4, LLaMA, and Qwen have demonstrated remarkable success across a wide range of applications. However, these models remain inherently vulnerable to prompt injection attacks, which can bypass existing safety mechanisms, highlighting the urgent need for more robust attack detection methods and comprehensive evaluation benchmarks. To address these challenges, we introduce GenTel-Safe, a unified framework that includes a novel prompt injection attack detection method, GenTel-Shield, along with a comprehensive evaluation benchmark, GenTel-Bench, which compromises 84812 prompt injection attacks, spanning 3 major categories and 28 security scenarios. To prove the effectiveness of GenTel-Shield, we evaluate it together with vanilla safety guardrails against the GenTel-Bench dataset. Empirically, GenTel-Shield can achieve state-of-the-art attack detection success rates, which reveals the critical weakness of existing safeguarding techniques against harmful prompts. For reproducibility, we have made the code and benchmarking dataset available on the project page at https://gentellab.github.io/gentel-safe.github.io/.

cs.CR↗

arXiv2024

Relation Between Photoionisation Cross Sections and Attosecond Time Delays

Determination and interpretation of Wigner-like photoionisation delays is one of the most active fields of attosecond science. Previous results have suggested that large photoionisation delays are associated with structured continua, but a quantitative relation between photoionisation cross sections and time delays has been missing. Here, we derive a Kramers-Kronig-like relation between these quantities and demonstrate its validity for (anti)resonances. This new concept defines a topological analysis, which rationalises the sign of photoionisation delays and thereby sheds new light on a long-standing controversy regarding the sign of the photoionisation delay near the Ar 3s Cooper minimum. Our work bridges traditional photoionisation spectroscopy with attosecond chronoscopy and offers new methods for analysing and interpreting photoionisation delays.

physics.atom-ph↗

arXiv2024

Debatrix: Multi-dimensional Debate Judge with Iterative Chronological Analysis Based on LLM

How can we construct an automated debate judge to evaluate an extensive, vibrant, multi-turn debate? This task is challenging, as judging a debate involves grappling with lengthy texts, intricate argument relationships, and multi-dimensional assessments. At the same time, current research mainly focuses on short dialogues, rarely touching upon the evaluation of an entire debate. In this paper, by leveraging Large Language Models (LLMs), we propose Debatrix, which makes the analysis and assessment of multi-turn debates more aligned with majority preferences. Specifically, Debatrix features a vertical, iterative chronological analysis and a horizontal, multi-dimensional evaluation collaboration. To align with real-world debate scenarios, we introduced the PanelBench benchmark, comparing our system's performance to actual debate outcomes. The findings indicate a notable enhancement over directly using LLMs for debate evaluation. Source code and benchmark data are available online at https://github.com/ljcleo/debatrix .

cs.CL↗