arXiv subjects

Peng Li

Publications and source records attributed to Peng Li.

At least 181 records · Page 10Linked to original sources

arXiv2025

ActiView: Evaluating Active Perception Ability for Multimodal Large Language Models

Active perception, a crucial human capability, involves setting a goal based on the current understanding of the environment and performing actions to achieve that goal. Despite significant efforts in evaluating Multimodal Large Language Models (MLLMs), active perception has been largely overlooked. To address this gap, we propose a novel benchmark named ActiView to evaluate active perception in MLLMs. We focus on a specialized form of Visual Question Answering (VQA) that eases and quantifies the evaluation yet challenging for existing MLLMs. Meanwhile, intermediate reasoning behaviors of models are also discussed. Given an image, we restrict the perceptual field of a model, requiring it to actively zoom or shift its perceptual field based on reasoning to answer the question successfully. We conduct extensive evaluation over 30 models, including proprietary and open-source models, and observe that restricted perceptual fields play a significant role in enabling active perception. Results reveal a significant gap in the active perception capability of MLLMs, indicating that this area deserves more attention. We hope that ActiView could help develop methods for MLLMs to understand multimodal inputs in more natural and holistic ways.

cs.CV↗

arXiv2025

ADO-LLM: Analog Design Bayesian Optimization with In-Context Learning of Large Language Models

Analog circuit design requires substantial human expertise and involvement, which is a significant roadblock to design productivity. Bayesian Optimization (BO), a popular machine learning based optimization strategy, has been leveraged to automate analog design given its applicability across various circuit topologies and technologies. Traditional BO methods employ black box Gaussian Process surrogate models and optimized labeled data queries to find optimization solutions by trading off between exploration and exploitation. However, the search for the optimal design solution in BO can be expensive from both a computational and data usage point of view, particularly for high dimensional optimization problems. This paper presents ADO-LLM, the first work integrating large language models (LLMs) with Bayesian Optimization for analog design optimization. ADO-LLM leverages the LLM's ability to infuse domain knowledge to rapidly generate viable design points to remedy BO's inefficiency in finding high value design areas specifically under the limited design space coverage of the BO's probabilistic surrogate model. In the meantime, sampling of design points evaluated in the iterative BO process provides quality demonstrations for the LLM to generate high quality design points while leveraging infused broad design knowledge. Furthermore, the diversity brought by BO's exploration enriches the contextual understanding of the LLM and allows it to more broadly search in the design space and prevent repetitive and redundant suggestions. We evaluate the proposed framework on two different types of analog circuits and demonstrate notable improvements in design efficiency and effectiveness.

cs.LG↗

arXiv2025

Bench4Merge: A Comprehensive Benchmark for Merging in Realistic Dense Traffic with Micro-Interactive Vehicles

While the capabilities of autonomous driving have advanced rapidly, merging into dense traffic remains a significant challenge, many motion planning methods for this scenario have been proposed but it is hard to evaluate them. Most existing closed-loop simulators rely on rule-based controls for other vehicles, which results in a lack of diversity and randomness, thus failing to accurately assess the motion planning capabilities in highly interactive scenarios. Moreover, traditional evaluation metrics are insufficient for comprehensively evaluating the performance of merging in dense traffic. In response, we proposed a closed-loop evaluation benchmark for assessing motion planning capabilities in merging scenarios. Our approach involves other vehicles trained in large scale datasets with micro-behavioral characteristics that significantly enhance the complexity and diversity. Additionally, we have restructured the evaluation mechanism by leveraging Large Language Models (LLMs) to assess each autonomous vehicle merging onto the main lane. Extensive experiments and test-vehicle deployment have demonstrated the progressiveness of this benchmark. Through this benchmark, we have obtained an evaluation of existing methods and identified common issues. The simulation environment and evaluation process can be accessed at https://github.com/WZM5853/Bench4Merge.

cs.RO↗

arXiv2025

1-Tb/s/λ Transmission over Record 10714-km AR-HCF

We present the first single-channel 1.001-Tb/s DP-36QAM-PCS recirculating transmission over 73 loops of 146.77-km ultra-low-loss & low-IMI DNANF-5 fiber, achieving a record transmission distance of 10,714.28 km.

physics.optics↗

arXiv2025

AdaMMS: Model Merging for Heterogeneous Multimodal Large Language Models with Unsupervised Coefficient Optimization

Recently, model merging methods have demonstrated powerful strengths in combining abilities on various tasks from multiple Large Language Models (LLMs). While previous model merging methods mainly focus on merging homogeneous models with identical architecture, they meet challenges when dealing with Multimodal Large Language Models (MLLMs) with inherent heterogeneous property, including differences in model architecture and the asymmetry in the parameter space. In this work, we propose AdaMMS, a novel model merging method tailored for heterogeneous MLLMs. Our method tackles the challenges in three steps: mapping, merging and searching. Specifically, we first design mapping function between models to apply model merging on MLLMs with different architecture. Then we apply linear interpolation on model weights to actively adapt the asymmetry in the heterogeneous MLLMs. Finally in the hyper-parameter searching step, we propose an unsupervised hyper-parameter selection method for model merging. As the first model merging method capable of merging heterogeneous MLLMs without labeled data, extensive experiments on various model combinations demonstrated that AdaMMS outperforms previous model merging methods on various vision-language benchmarks.

cs.CL↗

arXiv2025

SPIN: Accelerating Large Language Model Inference with Heterogeneous Speculative Models

Speculative decoding has been shown as an effective way to accelerate Large Language Model (LLM) inference by using a Small Speculative Model (SSM) to generate candidate tokens in a so-called speculation phase, which are subsequently verified by the LLM in a verification phase. However, current state-of-the-art speculative decoding approaches have three key limitations: handling requests with varying difficulty using homogeneous SSMs, lack of robust support for batch processing, and insufficient holistic optimization for both speculation and verification phases. In this paper, we introduce SPIN, an efficient LLM inference serving system based on speculative decoding, designed to address these challenges through three main innovations. First, SPIN improves token speculation by using multiple heterogeneous SSMs, with a learning-based algorithm for SSM selection that operates without prior knowledge of request difficulty. Second, SPIN employs a request decomposition method to minimize batching overhead during LLM verification. Finally, SPIN orchestrates speculation and verification phases by pipelining their executions on GPUs to achieve further acceleration. Experimental results demonstrate that SPIN significantly outperforms state-of-the-art methods, achieving a performance increase of approximately 2.28X.

cs.DC↗

arXiv2025

CoSpace: Benchmarking Continuous Space Perception Ability for Vision-Language Models

Vision-Language Models (VLMs) have recently witnessed significant progress in visual comprehension. As the permitting length of image context grows, VLMs can now comprehend a broader range of views and spaces. Current benchmarks provide insightful analysis of VLMs in tasks involving complex visual instructions following, multi-image understanding and spatial reasoning. However, they usually focus on spatially irrelevant images or discrete images captured from varied viewpoints. The compositional characteristic of images captured from a static viewpoint remains underestimated. We term this characteristic as Continuous Space Perception. When observing a scene from a static viewpoint while shifting orientations, it produces a series of spatially continuous images, enabling the reconstruction of the entire space. In this paper, we present CoSpace, a multi-image visual understanding benchmark designed to assess the Continuous Space perception ability for VLMs. CoSpace contains 2,918 images and 1,626 question-answer pairs, covering seven types of tasks. We conduct evaluation across 19 proprietary and open-source VLMs. Results reveal that there exist pitfalls on the continuous space perception ability for most of the evaluated models, including proprietary ones. Interestingly, we find that the main discrepancy between open-source and proprietary models lies not in accuracy but in the consistency of responses. We believe that enhancing the ability of continuous space perception is essential for VLMs to perform effectively in real-world tasks and encourage further research to advance this capability.

cs.CV↗

arXiv2025

Critical dynamics and its interferometry in the one-dimensional p-wave-paired Aubry-André-Harper model

In this work, we focus on the critical dynamics of the one-dimensional quasiperiodic p-wave-paired Aubry-André-Harper model, which exhibits a transition point between the gapped critical and the gapless localized phases. First, we disclose that the dynamical exponent features two distinct plateaus in the gapless localized phase. Besides the plateau with dynamical exponent $z=1.388$ near the transition point, there is a second one with $z=1$ away from the transition point. Then we demonstrate that these two plateaus intrinsically affect the critical dynamics with moderate quench rate by employing both one-way and round-trip quench protocols. In the one-way quench protocol, we clarify that the Kibble-Zurek (KZ) regime consists of two sub-regimes with different KZ exponents, which is a direct consequence of the two-plateau structure of the dynamical exponent. We also diagnose the KZ, pre-saturated, and saturated regimes by varying the quench rate from slow to fast limits. While, in the round-trip quench protocol, we confirm the interference effect of two critical dynamics and show a narrow neck in the oscillatory density of defects. It turns out that the position of the neck depends on the ratio of to-and-fro quench rates and the p-wave superconductivity pairing amplitude. We show how the position of the neck can be used to quantitatively determine the turning point of the two KZ sub-regimes.

quant-ph↗

arXiv2025

StableToolBench: Towards Stable Large-Scale Benchmarking on Tool Learning of Large Language Models

Large Language Models (LLMs) have witnessed remarkable advancements in recent years, prompting the exploration of tool learning, which integrates LLMs with external tools to address diverse real-world challenges. Assessing the capability of LLMs to utilise tools necessitates large-scale and stable benchmarks. However, previous works relied on either hand-crafted online tools with limited scale, or large-scale real online APIs suffering from instability of API status. To address this problem, we introduce StableToolBench, a benchmark evolving from ToolBench, proposing a virtual API server and stable evaluation system. The virtual API server contains a caching system and API simulators which are complementary to alleviate the change in API status. Meanwhile, the stable evaluation system designs solvable pass and win rates using GPT-4 as the automatic evaluator to eliminate the randomness during evaluation. Experimental results demonstrate the stability of StableToolBench, and further discuss the effectiveness of API simulators, the caching system, and the evaluator system.

cs.CL↗

arXiv2025

Revealing the electron-spin fluctuation coupling by photoemission in CaKFe4As4

Electron-boson coupling in unconventional superconductors is one of the key parameters in understanding the superconducting pairing symmetry. Here, we report definitive photoemission evidence of electron-spin exciton coupling in the iron-based superconductor CaKFe4As4, obtained via high-resolution ARPES. Our study identifies a distinct kink structure on the α band, observable only in the superconducting phase and closely linked with the superconductivity, indicative of strong electron-boson interactions. Notably, this kink structure corresponds to two distinct bosonic modes at 11 meV and 13 meV, aligning with spin resonance modes previously observed in inelastic neutron scattering experiments. This alignment underscores the significant role of antiferromagnetic fluctuations in the pairing mechanism of this superconductor. Furthermore, the unique momentum-dependent and orbital-selective properties of the coupling revealed by ARPES provide profound insights into the pairing symmetry, suggesting predominantly s_+- wave pairing facilitated by spin fluctuations. Our findings not only highlight the pivotal role of spin resonance in the superconductivity of CaKFe4As4 but also enhance understanding of the electron-spin exciton interactions in unconventional superconductors.

cond-mat.supr-con↗

arXiv2025

HoSNN: Adversarially-Robust Homeostatic Spiking Neural Networks with Adaptive Firing Thresholds

While spiking neural networks (SNNs) offer a promising neurally-inspired model of computation, they are vulnerable to adversarial attacks. We present the first study that draws inspiration from neural homeostasis to design a threshold-adapting leaky integrate-and-fire (TA-LIF) neuron model and utilize TA-LIF neurons to construct the adversarially robust homeostatic SNNs (HoSNNs) for improved robustness. The TA-LIF model incorporates a self-stabilizing dynamic thresholding mechanism, offering a local feedback control solution to the minimization of each neuron's membrane potential error caused by adversarial disturbance. Theoretical analysis demonstrates favorable dynamic properties of TA-LIF neurons in terms of the bounded-input bounded-output stability and suppressed time growth of membrane potential error, underscoring their superior robustness compared with the standard LIF neurons. When trained with weak FGSM attacks (attack budget = 2/255) and tested with much stronger PGD attacks (attack budget = 8/255), our HoSNNs significantly improve model accuracy on several datasets: from 30.54% to 74.91% on FashionMNIST, from 0.44% to 35.06% on SVHN, from 0.56% to 42.63% on CIFAR10, from 0.04% to 16.66% on CIFAR100, over the conventional LIF-based SNNs.

cs.NE↗

arXiv2025

Earlier Tokens Contribute More: Learning Direct Preference Optimization From Temporal Decay Perspective

Direct Preference Optimization (DPO) has gained attention as an efficient alternative to reinforcement learning from human feedback (RLHF) for aligning large language models (LLMs) with human preferences. Despite its advantages, DPO suffers from a length bias, generating responses longer than those from the reference model. Existing solutions like SimPO and SamPO address this issue but uniformly treat the contribution of rewards across sequences, overlooking temporal dynamics. To this end, we propose an enhanced preference optimization method that incorporates a temporal decay factor controlled by a gamma parameter. This dynamic weighting mechanism adjusts the influence of each reward based on its position in the sequence, prioritizing earlier tokens that are more critical for alignment. By adaptively focusing on more relevant feedback, our approach mitigates overfitting to less pertinent data and remains responsive to evolving human preferences. Experimental results on several benchmarks show that our approach consistently outperforms vanilla DPO by 5.9-8.8 points on AlpacaEval 2 and 3.3-9.7 points on Arena-Hard across different model architectures and sizes. Furthermore, additional experiments on mathematical and reasoning benchmarks (MMLU, GSM8K, and MATH) confirm that our method enhances performance without compromising general capabilities. Our codebase would be available at \url{https://github.com/LotuSrc/D2PO}.

cs.CL↗

arXiv2025

Quantum fluctuations-driven Melting Transitions in Two-dimensional Superconductors

Quantum fluctuations are pivotal in driving quantum phase transitions, exemplified by the quantum melting of Wigner crystals into Fermi liquids in electron systems. However, their impact on superconducting systems near zero temperature, particularly in the superconductor-insulator/metal transition, remains poorly understood. In this study, through electric transport measurements on the two-dimensional (2D) superconductor (SnS)1.17NbS2, we demonstrate that quantum fluctuations induce vortex displacement from their mean position, leading to the quantum melting of vortex solid near zero temperature. Quantitative analysis reveals the magnetic field-induced anomalous metal originates from this quantum melting transition, with energy dissipation governed by quantum fluctuations-driven vortex displacements. Remarkably, further extending this analysis to various 2D superconductors yields the same results, and many properties of anomalous metal can be qualitatively understood within the framework of quantum melting. The connection between the quantum melting of vortex solids and dissipative anomalous metal opens a novel pathway towards understanding quantum phase transitions through vortex dynamics, providing new insights on both fields.

cond-mat.supr-con↗

arXiv2025

LLM-USO: Large Language Model-based Universal Sizing Optimizer

The design of analog circuits is a cornerstone of integrated circuit (IC) development, requiring the optimization of complex, interconnected sub-structures such as amplifiers, comparators, and buffers. Traditionally, this process relies heavily on expert human knowledge to refine design objectives by carefully tuning sub-components while accounting for their interdependencies. Existing methods, such as Bayesian Optimization (BO), offer a mathematically driven approach for efficiently navigating large design spaces. However, these methods fall short in two critical areas compared to human expertise: (i) they lack the semantic understanding of the sizing solution space and its direct correlation with design objectives before optimization, and (ii) they fail to reuse knowledge gained from optimizing similar sub-structures across different circuits. To overcome these limitations, we propose the Large Language Model-based Universal Sizing Optimizer (LLM-USO), which introduces a novel method for knowledge representation to encode circuit design knowledge in a structured text format. This representation enables the systematic reuse of optimization insights for circuits with similar sub-structures. LLM-USO employs a hybrid framework that integrates BO with large language models (LLMs) and a learning summary module. This approach serves to: (i) infuse domain-specific knowledge into the BO process and (ii) facilitate knowledge transfer across circuits, mirroring the cognitive strategies of expert designers. Specifically, LLM-USO constructs a knowledge summary mechanism to distill and apply design insights from one circuit to related ones. It also incorporates a knowledge summary critiquing mechanism to ensure the accuracy and quality of the summaries and employs BO-guided suggestion filtering to identify optimal design points efficiently.

cs.AR↗

arXiv2025

Contrastive Private Data Synthesis via Weighted Multi-PLM Fusion

Substantial quantity and high quality are the golden rules of making a good training dataset with sample privacy protection equally important. Generating synthetic samples that resemble high-quality private data while ensuring Differential Privacy (DP), a formal privacy guarantee, promises scalability and practicality. However, existing methods relying on pre-trained models for data synthesis %that avoid fine-tuning large pre-trained generative models often struggle in data-deficient scenarios, suffering from limited sample size, inevitable generation noise and existing pre-trained model bias. To address these challenges, we propose a novel contrAstive private data Synthesis via Weighted multiple Pre-trained language models (PLM) framework, named as WASP. WASP utilizes limited private samples for more accurate private data distribution estimation via a Top-Q voting mechanism, and leverages low-quality synthetic samples for contrastive generation via collaboration among dynamically weighted multiple pre-trained models.Extensive experiments on 6 well-developed datasets with 6 open-source and 3 closed-source PLMs demonstrate the superiority of WASP in improving model performance over diverse downstream tasks. Code is available at https://anonymous.4open.science/r/WASP.

cs.LG↗

arXiv2025

Kibble-Zurek scaling immune to anti-Kibble-Zurek behavior in driven open systems at the limit of loss difference

We investigate the dissipative quench dynamics in a family of two-band fermionic systems by linearly ramping the staggered on-site energy. In the Lindblad formalism, we present an analytical solution in the presence of uniform loss or loss difference on bipartite lattices, which tells that dissipation exponentially suppresses the Kibble-Zurek (KZ) scaling behavior and the quantum jump term of the dissipation is responsible for the anti-KZ (AKZ) behavior. Interestingly, we find two different scaling behaviors at the limit of loss difference. Both scaling behaviors arise from the gapless Liouvillian. But one is accompanied by impulse stage rendered by the criticality of the system, so that it is ascribed to the universal KZ scaling law. Another depends on the dissipation strength and there is no impulse stage in it. We also point out a convenient way to observe the two new scaling behaviors by counting the number of residual particles in the end, since it is immune to the influence of AKZ behavior. We illustrate our findings through the prototypical one-dimensional Rice-Mele model first. Then, in the one-dimensional Shockley model and the two-dimensional Haldane model for Chern insulators, we show that the two scaling behaviors can appear together or separately with appropriate quench protocols.

quant-ph↗

arXiv2025

Orthogonal Geometry of Magneto-Optical Kerr Effect Enabled by Magnetization Multipole of Berry Curvature

The Magneto-Optical Kerr Effect (MOKE) is a fundamental tool in magnetometry, pivotal for advancing research in optics, magnetism, and spintronics as a direct probe of magnetization. Traditional MOKE measurements primarily detect the magnetization components parallel to the Poynting vector, which can only access the magnitude but not the direction of the orthogonal component. In this study, we introduce an orthogonal MOKE geometry in which the Kerr signal detects both the magnitude and direction of the magnetization component perpendicular to the Poynting vector. We demonstrate the broad applicability of this orthogonal geometry through the MOKE measurements in cubic ferromagnets and van der Waals ferromagnet. We theoretically show that the orthogonal MOKE geometry is enabled by the multipolar structure of Berry curvature in the magnetization space, which generally induces a Voigt vector orthogonal to the magnetization, thereby accounting for the unique magnetization angle dependence distinct from conventional MOKE. The establishment of the orthogonal MOKE geometry not only introduces a new paradigm for magneto-optical measurements but also provides a framework for exploring the magnetization multipoles of Berry curvature across the electromagnetic spectrum.

physics.optics↗

arXiv2025

Agent Hospital: A Simulacrum of Hospital with Evolvable Medical Agents

The recent rapid development of large language models (LLMs) has sparked a new wave of technological revolution in medical artificial intelligence (AI). While LLMs are designed to understand and generate text like a human, autonomous agents that utilize LLMs as their "brain" have exhibited capabilities beyond text processing such as planning, reflection, and using tools by enabling their "bodies" to interact with the environment. We introduce a simulacrum of hospital called Agent Hospital that simulates the entire process of treating illness, in which all patients, nurses, and doctors are LLM-powered autonomous agents. Within the simulacrum, doctor agents are able to evolve by treating a large number of patient agents without the need to label training data manually. After treating tens of thousands of patient agents in the simulacrum (human doctors may take several years in the real world), the evolved doctor agents outperform state-of-the-art medical agent methods on the MedQA benchmark comprising US Medical Licensing Examination (USMLE) test questions. Our methods of simulacrum construction and agent evolution have the potential in benefiting a broad range of applications beyond medical AI.

cs.AI↗