arXiv subjects

Zhicheng Zhang

Publications and source records attributed to Zhicheng Zhang.

At least 55 records · Page 3Linked to original sources

arXiv2025

Time-Efficient Quantum Entropy Estimator via Samplizer

Entropy is a measure of the randomness of a system. Estimating the entropy of a quantum state is a basic problem in quantum information. In this paper, we introduce a time-efficient quantum approach to estimating the von Neumann entropy $S(ρ)$ and Rényi entropy $S_α(ρ)$ of an $N$-dimensional quantum state $ρ$, given access to independent samples of $ρ$. Specifically, we provide the following: 1. A quantum estimator for $S(ρ)$ with time complexity $\tilde O(N^2)$, improving the prior best time complexity $\tilde O(N^6)$ by Acharya, Issa, Shende, and Wagner (2020) and Bavarian, Mehraba, and Wright (2016). 2. A quantum estimator for $S_α(ρ)$ with time complexity $\tilde O(N^{4/α-2})$ for $0<α<1$ and $\tilde O(N^{4-2/α})$ for $α>1$, improving the prior best time complexity $\tilde O(N^{6/α})$ for $0<α<1$ and $\tilde O(N^6)$ for $α>1$ by Acharya, Issa, Shende, and Wagner (2020), though at a cost of a slightly larger sample complexity. Moreover, these estimators are naturally extensible to the low-rank case. We also provide a sample lower bound for estimating $S_α(ρ)$. Technically, our method is quite different from the previous ones that are based on weak Schur sampling and Young diagrams. At the heart of our construction, is a novel tool called samplizer, which can "samplize" a quantum query algorithm to a quantum algorithm with similar behavior using only samples of quantum states; this suggests a unified framework for estimating quantum entropies. Specifically, when a quantum oracle $U$ block-encodes a mixed quantum state $ρ$, any quantum query algorithm using $Q$ queries to $U$ can be samplized to a $δ$-close (in the diamond norm) quantum algorithm using $\tildeΘ(Q^2/δ)$ samples of $ρ$. Moreover, this samplization is proven to be optimal, up to a polylogarithmic factor.

quant-ph↗

arXiv2025

VidEmo: Affective-Tree Reasoning for Emotion-Centric Video Foundation Models

Understanding and predicting emotion from videos has gathered significant attention in recent studies, driven by advancements in video large language models (VideoLLMs). While advanced methods have made progress in video emotion analysis, the intrinsic nature of emotions poses significant challenges. Emotions are characterized by dynamic and cues-dependent properties, making it difficult to understand complex and evolving emotional states with reasonable rationale. To tackle these challenges, we propose a novel affective cues-guided reasoning framework that unifies fundamental attribute perception, expression analysis, and high-level emotional understanding in a stage-wise manner. At the core of our approach is a family of video emotion foundation models (VidEmo), specifically designed for emotion reasoning and instruction-following. These models undergo a two-stage tuning process: first, curriculum emotion learning for injecting emotion knowledge, followed by affective-tree reinforcement learning for emotion reasoning. Moreover, we establish a foundational data infrastructure and introduce a emotion-centric fine-grained dataset (Emo-CFG) consisting of 2.1M diverse instruction-based samples. Emo-CFG includes explainable emotional question-answering, fine-grained captions, and associated rationales, providing essential resources for advancing emotion understanding tasks. Experimental results demonstrate that our approach achieves competitive performance, setting a new milestone across 15 face perception tasks.

cs.CV↗

arXiv2025

Quantum Lower Bounds by Sample-to-Query Lifting

The polynomial method by Beals, Buhrman, Cleve, Mosca, and de Wolf (FOCS 1998, J. ACM 2001), the adversary method by Ambainis (STOC 2000, J. Comput. Syst. Sci. 2002), and the compressed oracle method by Zhandry (CRYPTO 2019) have been shown to be powerful in proving quantum query lower bounds for a wide variety of problems. In this paper, we propose a new method for proving quantum query lower bounds by a quantum sample-to-query lifting theorem, which is from an information theory perspective. Using this method, we obtain the following new results: 1. A quadratic relation between quantum sample and query complexities regarding quantum property testing, which is optimal and saturated by quantum state discrimination. Here, the sample complexity is measured given sample access to the quantum state to be tested, while the query complexity is measured given query access to an oracle that block-encodes the quantum state. 2. A matching lower bound $\widetilde Ω(β)$ for quantum Gibbs sampling at inverse temperature $β$, showing that the quantum Gibbs sampler by Gilyén, Su, Low, and Wiebe (STOC 2019) is optimal. 3. A new lower bound $\widetilde Ω(1/\sqrtΔ)$ for the entanglement entropy problem with gap $Δ$, which was recently studied by She and Yuen (ITCS 2023). 4. A series of quantum query lower bounds for matrix spectrum testing, based on the sample lower bounds for quantum state spectrum testing by O'Donnell and Wright (STOC 2015, Comm. Math. Phys. 2021). In addition, we also provide unified proofs for some known lower bounds that have been proven previously via different techniques, including those for phase/amplitude estimation and Hamiltonian simulation.

quant-ph↗

arXiv2025

Initialize to Generalize: A Stronger Initialization Pipeline for Sparse-View 3DGS

Sparse-view 3D Gaussian Splatting (3DGS) often overfits to the training views, leading to artifacts like blurring in novel view rendering. Prior work addresses it either by enhancing the initialization (\emph{i.e.}, the point cloud from Structure-from-Motion (SfM)) or by adding training-time constraints (regularization) to the 3DGS optimization. Yet our controlled ablations reveal that initialization is the decisive factor: it determines the attainable performance band in sparse-view 3DGS, while training-time constraints yield only modest within-band improvements at extra cost. Given initialization's primacy, we focus our design there. Although SfM performs poorly under sparse views due to its reliance on feature matching, it still provides reliable seed points. Thus, building on SfM, our effort aims to supplement the regions it fails to cover as comprehensively as possible. Specifically, we design: (i) frequency-aware SfM that improves low-texture coverage via low-frequency view augmentation and relaxed multi-view correspondences; (ii) 3DGS self-initialization that lifts photometric supervision into additional points, compensating SfM-sparse regions with learned Gaussian centers; and (iii) point-cloud regularization that enforces multi-view consistency and uniform spatial coverage through simple geometric/visibility priors, yielding a clean and reliable point cloud. Our experiments on LLFF and Mip-NeRF360 demonstrate consistent gains in sparse-view settings, establishing our approach as a stronger initialization strategy. Code is available at https://github.com/zss171999645/ItG-GS.

cs.CV↗

arXiv2025

Fine-Tuning Vision-Language Models for Visual Navigation Assistance

We address vision-language-driven indoor navigation to assist visually impaired individuals in reaching a target location using images and natural language guidance. Traditional navigation systems are ineffective indoors due to the lack of precise location data. Our approach integrates vision and language models to generate step-by-step navigational instructions, enhancing accessibility and independence. We fine-tune the BLIP-2 model with Low Rank Adaptation (LoRA) on a manually annotated indoor navigation dataset. We propose an evaluation metric that refines the BERT F1 score by emphasizing directional and sequential variables, providing a more comprehensive measure of navigational performance. After applying LoRA, the model significantly improved in generating directional instructions, overcoming limitations in the original BLIP-2 model.

cs.CV↗

arXiv2025

AgentScope 1.0: A Developer-Centric Framework for Building Agentic Applications

Driven by rapid advancements of Large Language Models (LLMs), agents are empowered to combine intrinsic knowledge with dynamic tool use, greatly enhancing their capacity to address real-world tasks. In line with such an evolution, AgentScope introduces major improvements in a new version (1.0), towards comprehensively supporting flexible and efficient tool-based agent-environment interactions for building agentic applications. Specifically, we abstract foundational components essential for agentic applications and provide unified interfaces and extensible modules, enabling developers to easily leverage the latest progress, such as new models and MCPs. Furthermore, we ground agent behaviors in the ReAct paradigm and offer advanced agent-level infrastructure based on a systematic asynchronous design, which enriches both human-agent and agent-agent interaction patterns while improving execution efficiency. Building on this foundation, we integrate several built-in agents tailored to specific practical scenarios. AgentScope also includes robust engineering support for developer-friendly experiences. We provide a scalable evaluation module with a visual studio interface, making the development of long-trajectory agentic applications more manageable and easier to trace. In addition, AgentScope offers a runtime sandbox to ensure safe agent execution and facilitates rapid deployment in production environments. With these enhancements, AgentScope provides a practical foundation for building scalable, adaptive, and effective agentic applications.

cs.AI↗

arXiv2025

Making Teams and Influencing Agents: Efficiently Coordinating Decision Trees for Interpretable Multi-Agent Reinforcement Learning

Poor interpretability hinders the practical applicability of multi-agent reinforcement learning (MARL) policies. Deploying interpretable surrogates of uninterpretable policies enhances the safety and verifiability of MARL for real-world applications. However, if these surrogates are to interact directly with the environment within human supervisory frameworks, they must be both performant and computationally efficient. Prior work on interpretable MARL has either sacrificed performance for computational efficiency or computational efficiency for performance. To address this issue, we propose HYDRAVIPER, a decision tree-based interpretable MARL algorithm. HYDRAVIPER coordinates training between agents based on expected team performance, and adaptively allocates budgets for environment interaction to improve computational efficiency. Experiments on standard benchmark environments for multi-agent coordination and traffic signal control show that HYDRAVIPER matches the performance of state-of-the-art methods using a fraction of the runtime, and that it maintains a Pareto frontier of performance for different interaction budgets.

cs.MA↗

arXiv2025

Hot-Swap MarkBoard: An Efficient Black-box Watermarking Approach for Large-scale Model Distribution

Recently, Deep Learning (DL) models have been increasingly deployed on end-user devices as On-Device AI, offering improved efficiency and privacy. However, this deployment trend poses more serious Intellectual Property (IP) risks, as models are distributed on numerous local devices, making them vulnerable to theft and redistribution. Most existing ownership protection solutions (e.g., backdoor-based watermarking) are designed for cloud-based AI-as-a-Service (AIaaS) and are not directly applicable to large-scale distribution scenarios, where each user-specific model instance must carry a unique watermark. These methods typically embed a fixed watermark, and modifying the embedded watermark requires retraining the model. To address these challenges, we propose Hot-Swap MarkBoard, an efficient watermarking method. It encodes user-specific $n$-bit binary signatures by independently embedding multiple watermarks into a multi-branch Low-Rank Adaptation (LoRA) module, enabling efficient watermark customization without retraining through branch swapping. A parameter obfuscation mechanism further entangles the watermark weights with those of the base model, preventing removal without degrading model performance. The method supports black-box verification and is compatible with various model architectures and DL tasks, including classification, image generation, and text generation. Extensive experiments across three types of tasks and six backbone models demonstrate our method's superior efficiency and adaptability compared to existing approaches, achieving 100\% verification accuracy.

cs.CR↗

arXiv2025

Sparsity-Promoting Dynamic Mode Decomposition Applied to Sea Surface Temperature Fields

In this paper, we leverage Koopman mode decomposition to analyze the nonlinear and high-dimensional climate systems acting on the observed data space. The dynamics of atmospheric systems are assumed to be equation-free, with the linear evolution of observables derived from measured historical long-term time-series data snapshots, such as monthly sea surface temperature records, to construct a purely data-driven climate dynamics. In particular, sparsity-promoting dynamic mode decomposition is exploited to extract the dominant spatial and temporal modes, which are among the most significant coherent structures underlying climate variability, enabling a more efficient, interpretable, and low-dimensional representation of the system dynamics. We hope that the combined use of Koopman modes and sparsity-promoting techniques will provide insights into the significant climate modes, enabling reduced-order modeling of the climate system and offering a potential framework for predicting and controlling weather and climate variability.

eess.SY↗

arXiv2025

MODA: MOdular Duplex Attention for Multimodal Perception, Cognition, and Emotion Understanding

Multimodal large language models (MLLMs) recently showed strong capacity in integrating data among multiple modalities, empowered by a generalizable attention architecture. Advanced methods predominantly focus on language-centric tuning while less exploring multimodal tokens mixed through attention, posing challenges in high-level tasks that require fine-grained cognition and emotion understanding. In this work, we identify the attention deficit disorder problem in multimodal learning, caused by inconsistent cross-modal attention and layer-by-layer decayed attention activation. To address this, we propose a novel attention mechanism, termed MOdular Duplex Attention (MODA), simultaneously conducting the inner-modal refinement and inter-modal interaction. MODA employs a correct-after-align strategy to effectively decouple modality alignment from cross-layer token mixing. In the alignment phase, tokens are mapped to duplex modality spaces based on the basis vectors, enabling the interaction between visual and language modality. Further, the correctness of attention scores is ensured through adaptive masked attention, which enhances the model's flexibility by allowing customizable masking patterns for different modalities. Extensive experiments on 21 benchmark datasets verify the effectiveness of MODA in perception, cognition, and emotion tasks. Source code and demo are available in https://zzcheng.top/MODA.

cs.CV↗

arXiv2025

Access Control Threatened by Quantum Entanglement

Access control is a cornerstone of computer security that prevents unauthorised access to resources. In this paper, we study access control in quantum computer systems. We present the first explicit scenario of a security breach when a classically secure access control system is straightforwardly adapted to the quantum setting. The breach is ultimately due to that quantum mechanics allows the phenomenon of entanglement and violates Mermin inequality, a multi-party variant of the celebrated Bell inequality. This reveals a threat from quantum entanglement to access control if existing computer systems integrate with quantum computing. To protect against such threat, we propose several new models of quantum access control, and rigorously analyse their security, flexibility and efficiency.

quant-ph↗

arXiv2025

Learning Instruction-Following Policies through Open-Ended Instruction Relabeling with Large Language Models

Developing effective instruction-following policies in reinforcement learning remains challenging due to the reliance on extensive human-labeled instruction datasets and the difficulty of learning from sparse rewards. In this paper, we propose a novel approach that leverages the capabilities of large language models (LLMs) to automatically generate open-ended instructions retrospectively from previously collected agent trajectories. Our core idea is to employ LLMs to relabel unsuccessful trajectories by identifying meaningful subtasks the agent has implicitly accomplished, thereby enriching the agent's training data and substantially alleviating reliance on human annotations. Through this open-ended instruction relabeling, we efficiently learn a unified instruction-following policy capable of handling diverse tasks within a single policy. We empirically evaluate our proposed method in the challenging Craftax environment, demonstrating clear improvements in sample efficiency, instruction coverage, and overall policy performance compared to state-of-the-art baselines. Our results highlight the effectiveness of utilizing LLM-guided open-ended instruction relabeling to enhance instruction-following reinforcement learning.

cs.LG↗

arXiv2025

Quantum algorithms for matrix geometric means

Matrix geometric means between two positive definite matrices can be defined from distinct perspectives - as solutions to certain nonlinear systems of equations, as points along geodesics in Riemannian geometry, and as solutions to certain optimisation problems. We devise quantum subroutines for the matrix geometric means, and construct solutions to the algebraic Riccati equation - an important class of nonlinear systems of equations appearing in machine learning, optimal control, estimation, and filtering. Using these subroutines, we present a new class of quantum learning algorithms, for both classical and quantum data, called quantum geometric mean metric learning, for weakly supervised learning and anomaly detection. The subroutines are also useful for estimating geometric Rényi relative entropies and the Uhlmann fidelity, in particular achieving optimal dependence on precision for the Uhlmann and Matsumoto fidelities. Finally, we provide a BQP-complete problem based on matrix geometric means that can be solved by our subroutines.

quant-ph↗

arXiv2025

Quantum Register Machine: Efficient Implementation of Quantum Recursive Programs

Quantum recursive programming has been recently introduced for describing sophisticated and complicated quantum algorithms in a compact and elegant way. However, implementation of quantum recursion involves intricate interplay between quantum control flow and recursive procedure calls. In this paper, we aim at resolving this fundamental challenge and develop a series of techniques to efficiently implement quantum recursive programs. Our main contributions include: 1. We propose a notion of quantum register machine, the first quantum architecture (including an instruction set) that provides instruction-level support for quantum control flow and recursive procedure calls at the same time. 2. Based on quantum register machine, we describe the first comprehensive implementation process of quantum recursive programs, including the compilation, the partial evaluation of quantum control flow, and the execution on the quantum register machine. 3. As a bonus, our efficient implementation of quantum recursive programs also offers automatic parallelisation of quantum algorithms. For implementing certain quantum algorithmic subroutine, like the widely used quantum multiplexor, we can even obtain exponential parallel speed-up (over the straightforward implementation) from this automatic parallelisation. This demonstrates that quantum recursive programming can be win-win for both modularity of programs and efficiency of their implementation.

quant-ph↗

arXiv2025

M3HF: Multi-agent Reinforcement Learning from Multi-phase Human Feedback of Mixed Quality

Designing effective reward functions in multi-agent reinforcement learning (MARL) is a significant challenge, often leading to suboptimal or misaligned behaviors in complex, coordinated environments. We introduce Multi-agent Reinforcement Learning from Multi-phase Human Feedback of Mixed Quality ($\text{M}^3\text{HF}$), a novel framework that integrates multi-phase human feedback of mixed quality into the MARL training process. By involving humans with diverse expertise levels to provide iterative guidance, $\text{M}^3\text{HF}$ leverages both expert and non-expert feedback to continuously refine agents' policies. During training, we strategically pause agent learning for human evaluation, parse feedback using large language models to assign it appropriately and update reward functions through predefined templates and adaptive weights by using weight decay and performance-based adjustments. Our approach enables the integration of nuanced human insights across various levels of quality, enhancing the interpretability and robustness of multi-agent cooperation. Empirical results in challenging environments demonstrate that $\text{M}^3\text{HF}$ significantly outperforms state-of-the-art methods, effectively addressing the complexities of reward design in MARL and enabling broader human participation in the training process.

cs.MA↗

arXiv2025

Adaptive Differential Denoising for Respiratory Sounds Classification

Automated respiratory sound classification faces practical challenges from background noise and insufficient denoising in existing systems. We propose Adaptive Differential Denoising network, that integrates noise suppression and pathological feature preservation via three innovations: 1) Adaptive Frequency Filter with learnable spectral masks and soft shrink to eliminate noise while retaining diagnostic high-frequency components; 2) A Differential Denoise Layer using differential attention to reduce noise-induced variations through augmented sample comparisons; 3) A bias denoising loss jointly optimizing classification and robustness without clean labels. Experiments on the ICBHI2017 dataset show that our method achieves 65.53\% of the Score, which is improved by 1.99\% over the previous sota method. The code is available in https://github.com/deegy666/ADD-RSC

eess.AS↗

arXiv2025

Category-aware EEG image generation based on wavelet transform and contrast semantic loss

Reconstructing visual stimuli from EEG signals is a crucial step in realizing brain-computer interfaces. In this paper, we propose a transformer-based EEG signal encoder integrating the Discrete Wavelet Transform (DWT) and the gating mechanism. Guided by the feature alignment and category-aware fusion losses, this encoder is used to extract features related to visual stimuli from EEG signals. Subsequently, with the aid of a pre-trained diffusion model, these features are reconstructed into visual stimuli. To verify the effectiveness of the model, we conducted EEG-to-image generation and classification tasks using the THINGS-EEG dataset. To address the limitations of quantitative analysis at the semantic level, we combined WordNet-based classification and semantic similarity metrics to propose a novel semantic-based score, emphasizing the ability of our model to transfer neural activities into visual representations. Experimental results show that our model significantly improves semantic alignment and classification accuracy, which achieves a maximum single-subject accuracy of 43\%, outperforming other state-of-the-art methods. The source code and supplementary material is available at https://github.com/zes0v0inn/DWT_EEG_Reconstruction/tree/main.

cs.CV↗

arXiv2025

Hierarchical and Step-Layer-Wise Tuning of Attention Specialty for Multi-Instance Synthesis in Diffusion Transformers

Text-to-image (T2I) generation models often struggle with multi-instance synthesis (MIS), where they must accurately depict multiple distinct instances in a single image based on complex prompts detailing individual features. Traditional MIS control methods for UNet architectures like SD v1.5/SDXL fail to adapt to DiT-based models like FLUX and SD v3.5, which rely on integrated attention between image and text tokens rather than text-image cross-attention. To enhance MIS in DiT, we first analyze the mixed attention mechanism in DiT. Our token-wise and layer-wise analysis of attention maps reveals a hierarchical response structure: instance tokens dominate early layers, background tokens in middle layers, and attribute tokens in later layers. Building on this observation, we propose a training-free approach for enhancing MIS in DiT-based models with hierarchical and step-layer-wise attention specialty tuning (AST). AST amplifies key regions while suppressing irrelevant areas in distinct attention maps across layers and steps, guided by the hierarchical structure. This optimizes multimodal interactions by hierarchically decoupling the complex prompts with instance-based sketches. We evaluate our approach using upgraded sketch-based layouts for the T2I-CompBench and customized complex scenes. Both quantitative and qualitative results confirm our method enhances complex layout generation, ensuring precise instance placement and attribute representation in MIS.

cs.CV↗