arXiv subjects

Zihao Wang

Publications and source records attributed to Zihao Wang.

At least 145 records · Page 8Linked to original sources

arXiv2025

Prompt Injection attack against LLM-integrated Applications

Large Language Models (LLMs), renowned for their superior proficiency in language comprehension and generation, stimulate a vibrant ecosystem of applications around them. However, their extensive assimilation into various services introduces significant security risks. This study deconstructs the complexities and implications of prompt injection attacks on actual LLM-integrated applications. Initially, we conduct an exploratory analysis on ten commercial applications, highlighting the constraints of current attack strategies in practice. Prompted by these limitations, we subsequently formulate HouYi, a novel black-box prompt injection attack technique, which draws inspiration from traditional web injection attacks. HouYi is compartmentalized into three crucial elements: a seamlessly-incorporated pre-constructed prompt, an injection prompt inducing context partition, and a malicious payload designed to fulfill the attack objectives. Leveraging HouYi, we unveil previously unknown and severe attack outcomes, such as unrestricted arbitrary LLM usage and uncomplicated application prompt theft. We deploy HouYi on 36 actual LLM-integrated applications and discern 31 applications susceptible to prompt injection. 10 vendors have validated our discoveries, including Notion, which has the potential to impact millions of users. Our investigation illuminates both the possible risks of prompt injection attacks and the possible tactics for mitigation.

cs.CR↗

arXiv2025

A Comprehensive Survey on Generative AI for Video-to-Music Generation

The burgeoning growth of video-to-music generation can be attributed to the ascendancy of multimodal generative models. However, there is a lack of literature that comprehensively combs through the work in this field. To fill this gap, this paper presents a comprehensive review of video-to-music generation using deep generative AI techniques, focusing on three key components: conditioning input construction, conditioning mechanism, and music generation frameworks. We categorize existing approaches based on their designs for each component, clarifying the roles of different strategies. Preceding this, we provide a fine-grained categorization of video and music modalities, illustrating how different categories influence the design of components within the generation pipelines. Furthermore, we summarize available multimodal datasets and evaluation metrics while highlighting ongoing challenges in the field.

eess.AS↗

arXiv2025

Reason to Contrast: A Cascaded Multimodal Retrieval Framework

Traditional multimodal retrieval systems rely primarily on bi-encoder architectures, where performance is closely tied to embedding dimensionality. Recent work, Think-Then-Embed (TTE), shows that incorporating multimodal reasoning to elicit additional informative tokens before embedding can further improve retrieval. In this paper, we extend this paradigm with TTE-v2, a hybrid multimodal retrieval framework that introduces reasoning-driven performance scaling based on additional input token budget rather than model or embedding size. Our approach augments the initial multimodal retrieval with additional reasoning steps for reranking, enabling more expressive query-candidate interactions at test time. The reranking stage further provides fine-grained supervision for hard negative mining and false negative filtering, creating a feedback loop that effectively strengthens the upstream retriever. This cascaded design delivers substantial test-time improvements based on intermediate reasoning token scaling. Experiments on the MMEB-V2 benchmark demonstrate that TTE-v2-7B achieves a new state-of-the-art accuracy of 75.7%, and that TTE-v2-2B matches or surpasses leading 7B models trained with significantly larger external data. Our results highlight the promise of token-wise scaling as an alternative scaling paradigm for multimodal retrieval.

cs.IR↗

arXiv2025

From Bits to Boardrooms: A Cutting-Edge Multi-Agent LLM Framework for Business Excellence

Large Language Models (LLMs) have shown promising potential in business applications, particularly in enterprise decision support and strategic planning, yet current approaches often struggle to reconcile intricate operational analyses with overarching strategic goals across diverse market environments, leading to fragmented workflows and reduced collaboration across organizational levels. This paper introduces BusiAgent, a novel multi-agent framework leveraging LLMs for advanced decision-making in complex corporate environments. BusiAgent integrates three core innovations: an extended Continuous Time Markov Decision Process (CTMDP) for dynamic agent modeling, a generalized entropy measure to optimize collaborative efficiency, and a multi-level Stackelberg game to handle hierarchical decision processes. Additionally, contextual Thompson sampling is employed for prompt optimization, supported by a comprehensive quality assurance system to mitigate errors. Extensive empirical evaluations across diverse business scenarios validate BusiAgent's efficacy, demonstrating its capacity to generate coherent, client-focused solutions that smoothly integrate granular insights with high-level strategy, significantly outperforming established approaches in both solution quality and user satisfaction. By fusing cutting-edge AI technologies with deep business insights, BusiAgent marks a substantial step forward in AI-driven enterprise decision-making, empowering organizations to navigate complex business landscapes more effectively.

cs.AI↗

arXiv2025

SceneMaker: Open-set 3D Scene Generation with Decoupled De-occlusion and Pose Estimation Model

We propose a decoupled 3D scene generation framework called SceneMaker in this work. Due to the lack of sufficient open-set de-occlusion and pose estimation priors, existing methods struggle to simultaneously produce high-quality geometry and accurate poses under severe occlusion and open-set settings. To address these issues, we first decouple the de-occlusion model from 3D object generation, and enhance it by leveraging image datasets and collected de-occlusion datasets for much more diverse open-set occlusion patterns. Then, we propose a unified pose estimation model that integrates global and local mechanisms for both self-attention and cross-attention to improve accuracy. Besides, we construct an open-set 3D scene dataset to further extend the generalization of the pose estimation model. Comprehensive experiments demonstrate the superiority of our decoupled framework on both indoor and open-set scenes. Our codes and datasets is released at https://idea-research.github.io/SceneMaker/.

cs.CV↗

arXiv2025

TAO-Net: Two-stage Adaptive OOD Classification Network for Fine-grained Encrypted Traffic Classification

Encrypted traffic classification aims to identify applications or services by analyzing network traffic data. One of the critical challenges is the continuous emergence of new applications, which generates Out-of-Distribution (OOD) traffic patterns that deviate from known categories and are not well represented by predefined models. Current approaches rely on predefined categories, which limits their effectiveness in handling unknown traffic types. Although some methods mitigate this limitation by simply classifying unknown traffic into a single "Other" category, they fail to make a fine-grained classification. In this paper, we propose a Two-stage Adaptive OOD classification Network (TAO-Net) that achieves accurate classification for both In-Distribution (ID) and OOD encrypted traffic. The method incorporates an innovative two-stage design: the first stage employs a hybrid OOD detection mechanism that integrates transformer-based inter-layer transformation smoothness and feature analysis to effectively distinguish between ID and OOD traffic, while the second stage leverages large language models with a novel semantic-enhanced prompt strategy to transform OOD traffic classification into a generation task, enabling flexible fine-grained classification without relying on predefined labels. Experiments on three datasets demonstrate that TAO-Net achieves 96.81-97.70% macro-precision and 96.77-97.68% macro-F1, outperforming previous methods that only reach 44.73-86.30% macro-precision, particularly in identifying emerging network applications.

cs.LG↗

arXiv2025

Singing Timbre Popularity Assessment Based on Multimodal Large Foundation Model

Automated singing assessment is crucial for education and entertainment. However, existing systems face two fundamental limitations: reliance on reference tracks, which stifles creative expression, and the simplification of complex performances into non-diagnostic scores based solely on pitch and rhythm. We advocate for a shift from discriminative to descriptive evaluation, creating a complete ecosystem for reference-free, multi-dimensional assessment. First, we introduce Sing-MD, a large-scale dataset annotated by experts across four dimensions: breath control, timbre quality, emotional expression, and vocal technique. Our analysis reveals significant annotation inconsistencies among experts, challenging the validity of traditional accuracy-based metrics. Second, addressing the memory limitations of Multimodal Large Language Models (MLLMs) in analyzing full-length songs, we propose VocalVerse. This efficient hybrid architecture leverages a lightweight acoustic encoder to model global performance features and long-term dependencies. Third, to address automated metric shortcomings, we establish the H-TPR (Human-in-the-loop Tiered Perceptual Ranking) benchmark, which evaluates a model's ability to generate perceptually valid rankings rather than predicting noisy ground-truth scores.

cs.SD↗

arXiv2025

Multi-Accent Mandarin Dry-Vocal Singing Dataset: Benchmark for Singing Accent Recognition

Singing accent research is underexplored compared to speech accent studies, primarily due to the scarcity of suitable datasets. Existing singing datasets often suffer from detail loss, frequently resulting from the vocal-instrumental separation process. Additionally, they often lack regional accent annotations. To address this, we introduce the Multi-Accent Mandarin Dry-Vocal Singing Dataset (MADVSD). MADVSD comprises over 670 hours of dry vocal recordings from 4,206 native Mandarin speakers across nine distinct Chinese regions. In addition to each participant recording audio of three popular songs in their native accent, they also recorded phonetic exercises covering all Mandarin vowels and a full octave range. We validated MADVSD through benchmark experiments in singing accent recognition, demonstrating its utility for evaluating state-of-the-art speech models in singing contexts. Furthermore, we explored dialectal influences on singing accent and analyzed the role of vowels in accentual variations, leveraging MADVSD's unique phonetic exercises.

cs.SD↗

arXiv2025

DefenSee: Dissecting Threat from Sight and Text -- A Multi-View Defensive Pipeline for Multi-modal Jailbreaks

Multi-modal large language models (MLLMs), capable of processing text, images, and audio, have been widely adopted in various AI applications. However, recent MLLMs integrating images and text remain highly vulnerable to coordinated jailbreaks. Existing defenses primarily focus on the text, lacking robust multi-modal protection. As a result, studies indicate that MLLMs are more susceptible to malicious or unsafe instructions, unlike their text-only counterparts. In this paper, we proposed DefenSee, a robust and lightweight multi-modal black-box defense technique that leverages image variants transcription and cross-modal consistency checks, mimicking human judgment. Experiments on popular multi-modal jailbreak and benign datasets show that DefenSee consistently enhances MLLM robustness while better preserving performance on benign tasks compared to SOTA defenses. It reduces the ASR of jailbreak attacks to below 1.70% on MiniGPT4 using the MM-SafetyBench benchmark, significantly outperforming prior methods under the same conditions.

cs.CR↗

arXiv2025

Mavors: Multi-granularity Video Representation for Multimodal Large Language Model

Long-context video understanding in multimodal large language models (MLLMs) faces a critical challenge: balancing computational efficiency with the retention of fine-grained spatio-temporal patterns. Existing approaches (e.g., sparse sampling, dense sampling with low resolution, and token compression) suffer from significant information loss in temporal dynamics, spatial details, or subtle interactions, particularly in videos with complex motion or varying resolutions. To address this, we propose $\mathbf{Mavors}$, a novel framework that introduces $\mathbf{M}$ulti-gr$\mathbf{a}$nularity $\mathbf{v}$ide$\mathbf{o}$ $\mathbf{r}$epre$\mathbf{s}$entation for holistic long-video modeling. Specifically, Mavors directly encodes raw video content into latent representations through two core components: 1) an Intra-chunk Vision Encoder (IVE) that preserves high-resolution spatial features via 3D convolutions and Vision Transformers, and 2) an Inter-chunk Feature Aggregator (IFA) that establishes temporal coherence across chunks using transformer-based dependency modeling with chunk-level rotary position encodings. Moreover, the framework unifies image and video understanding by treating images as single-frame videos via sub-image decomposition. Experiments across diverse benchmarks demonstrate Mavors' superiority in maintaining both spatial fidelity and temporal continuity, significantly outperforming existing methods in tasks requiring fine-grained spatio-temporal reasoning.

cs.CV↗

arXiv2025

Chemical vapor deposition growth of continuous monolayer antiferromagnetic CrOCl films

The discovery of two-dimensional magnetic materials has provided an ideal platform for exploring physical phenomena in the two-dimensional limit. However, intrinsic two-dimensional antiferromagnetic materials have been rarely reported, limiting systematic studies of their electronic properties. The discovery of novel intrinsic two-dimensional antiferromagnets and the development of robust synthesis strategies, therefore, remain significant challenges. Here, we report the chemical vapor deposition synthesis of CrOCl monolayer films and nanosheets that exhibit excellent air stability. The CrOCl morphology is tunable, ranging from two-dimensional nanosheets to three-dimensional flower-like structures, with lateral sizes ranging from several microns to continuous monolayer films. Structural characterization confirms the materials composition and high crystalline quality. Furthermore, magnetic measurements, supported by theoretical calculations, reveal a Néel temperature for CrOCl of ~14 K. This work provides a reliable route for preparing two-dimensional antiferromagnetic materials.

cond-mat.mes-hall↗

arXiv2025

PRBench: Large-Scale Expert Rubrics for Evaluating High-Stakes Professional Reasoning

Frontier model progress is often measured by academic benchmarks, which offer a limited view of performance in real-world professional contexts. Existing evaluations often fail to assess open-ended, economically consequential tasks in high-stakes domains like Legal and Finance, where practical returns are paramount. To address this, we introduce Professional Reasoning Bench (PRBench), a realistic, open-ended, and difficult benchmark of real-world problems in Finance and Law. We open-source its 1,100 expert-authored tasks and 19,356 expert-curated criteria, making it, to our knowledge, the largest public, rubric-based benchmark for both legal and finance domains. We recruit 182 qualified professionals, holding JDs, CFAs, or 6+ years of experience, who contributed tasks inspired by their actual workflows. This process yields significant diversity, with tasks spanning 114 countries and 47 US jurisdictions. Our expert-curated rubrics are validated through a rigorous quality pipeline, including independent expert validation. Subsequent evaluation of 20 leading models reveals substantial room for improvement, with top scores of only 0.39 (Finance) and 0.37 (Legal) on our Hard subsets. We further catalog associated economic impacts of the prompts and analyze performance using human-annotated rubric categories. Our analysis shows that models with similar overall scores can diverge significantly on specific capabilities. Common failure modes include inaccurate judgments, a lack of process transparency and incomplete reasoning, highlighting critical gaps in their reliability for professional adoption.

cs.CL↗

arXiv2025

Diff-V2M: A Hierarchical Conditional Diffusion Model with Explicit Rhythmic Modeling for Video-to-Music Generation

Video-to-music (V2M) generation aims to create music that aligns with visual content. However, two main challenges persist in existing methods: (1) the lack of explicit rhythm modeling hinders audiovisual temporal alignments; (2) effectively integrating various visual features to condition music generation remains non-trivial. To address these issues, we propose Diff-V2M, a general V2M framework based on a hierarchical conditional diffusion model, comprising two core components: visual feature extraction and conditional music generation. For rhythm modeling, we begin by evaluating several rhythmic representations, including low-resolution mel-spectrograms, tempograms, and onset detection functions (ODF), and devise a rhythmic predictor to infer them directly from videos. To ensure contextual and affective coherence, we also extract semantic and emotional features. All features are incorporated into the generator via a hierarchical cross-attention mechanism, where emotional features shape the affective tone via the first layer, while semantic and rhythmic features are fused in the second cross-attention layer. To enhance feature integration, we introduce timestep-aware fusion strategies, including feature-wise linear modulation (FiLM) and weighted fusion, allowing the model to adaptively balance semantic and rhythmic cues throughout the diffusion process. Extensive experiments identify low-resolution ODF as a more effective signal for modeling musical rhythm and demonstrate that Diff-V2M outperforms existing models on both in-domain and out-of-domain datasets, achieving state-of-the-art performance in terms of objective metrics and subjective comparisons. Demo and code are available at https://Tayjsl97.github.io/Diff-V2M-Demo/.

cs.SD↗

arXiv2025

CryptPEFT: Efficient and Private Neural Network Inference via Parameter-Efficient Fine-Tuning

Publicly available large pretrained models (i.e., backbones) and lightweight adapters for parameter-efficient fine-tuning (PEFT) have become standard components in modern machine learning pipelines. However, preserving the privacy of both user inputs and fine-tuned adapters -- often trained on sensitive data -- during inference remains a significant challenge. Applying cryptographic techniques, such as multi-party computation (MPC), to PEFT settings still incurs substantial encrypted computation across both the backbone and adapter, mainly due to the inherent two-way communication between them. To address this limitation, we propose CryptPEFT, the first PEFT solution specifically designed for private inference scenarios. CryptPEFT introduces a novel one-way communication (OWC) architecture that confines encrypted computation solely to the adapter, significantly reducing both computational and communication overhead. To maintain strong model utility under this constraint, we explore the design space of OWC-compatible adapters and employ an automated architecture search algorithm to optimize the trade-off between private inference efficiency and model utility. We evaluated CryptPEFT using Vision Transformer backbones across widely used image classification datasets. Our results show that CryptPEFT significantly outperforms existing baselines, delivering speedups ranging from $20.62\times$ to $291.48\times$ in simulated wide-area network (WAN) and local-area network (LAN) settings. On CIFAR-100, CryptPEFT attains 85.47% accuracy with just 2.26 seconds of inference latency. These findings demonstrate that CryptPEFT offers an efficient and privacy-preserving solution for modern PEFT-based inference.

cs.CR↗

arXiv2025

Game-TARS: Pretrained Foundation Models for Scalable Generalist Multimodal Game Agents

We present Game-TARS, a generalist game agent trained with a unified, scalable action space anchored to human-aligned native keyboard-mouse inputs. Unlike API- or GUI-based approaches, this paradigm enables large-scale continual pre-training across heterogeneous domains, including OS, web, and simulation games. Game-TARS is pre-trained on over 500B tokens with diverse trajectories and multimodal data. Key techniques include a decaying continual loss to reduce causal confusion and an efficient Sparse-Thinking strategy that balances reasoning depth and inference cost. Experiments show that Game-TARS achieves about 2 times the success rate over the previous sota model on open-world Minecraft tasks, is close to the generality of fresh humans in unseen web 3d games, and outperforms GPT-5, Gemini-2.5-Pro, and Claude-4-Sonnet in FPS benchmarks. Scaling results on training-time and test-time confirm that the unified action space sustains improvements when scaled to cross-game and multimodal data. Our results demonstrate that simple, scalable action representations combined with large-scale pre-training provide a promising path toward generalist agents with broad computer-use abilities.

cs.AI↗

arXiv2025

The Early Bird Catches the Leak: Unveiling Timing Side Channels in LLM Serving Systems

The wide deployment of Large Language Models (LLMs) has given rise to strong demands for optimizing their inference performance. Today's techniques serving this purpose primarily focus on reducing latency and improving throughput through algorithmic and hardware enhancements, while largely overlooking their privacy side effects, particularly in a multi-user environment. In our research, for the first time, we discovered a set of new timing side channels in LLM systems, arising from shared caches and GPU memory allocations, which can be exploited to infer both confidential system prompts and those issued by other users. These vulnerabilities echo security challenges observed in traditional computing systems, highlighting an urgent need to address potential information leakage in LLM serving infrastructures. In this paper, we report novel attack strategies designed to exploit such timing side channels inherent in LLM deployments, specifically targeting the Key-Value (KV) cache and semantic cache widely used to enhance LLM inference performance. Our approach leverages timing measurements and classification models to detect cache hits, allowing an adversary to infer private prompts with high accuracy. We also propose a token-by-token search algorithm to efficiently recover shared prompt prefixes in the caches, showing the feasibility of stealing system prompts and those produced by peer users. Our experimental studies on black-box testing of popular online LLM services demonstrate that such privacy risks are completely realistic, with significant consequences. Our findings underscore the need for robust mitigation to protect LLM systems against such emerging threats.

cs.CR↗

arXiv2025

Online Rubrics Elicitation from Pairwise Comparisons

Rubrics provide a flexible way to train LLMs on open-ended long-form answers where verifiable rewards are not applicable and human preferences provide coarse signals. Prior work shows that reinforcement learning with rubric-based rewards leads to consistent gains in LLM post-training. Most existing approaches rely on rubrics that remain static over the course of training. Such static rubrics, however, are vulnerable to reward-hacking type behaviors and fail to capture emergent desiderata that arise during training. We introduce Online Rubrics Elicitation (OnlineRubrics), a method that dynamically curates evaluation criteria in an online manner through pairwise comparisons of responses from current and reference policies. This online process enables continuous identification and mitigation of errors as training proceeds. Empirically, this approach yields consistent improvements of up to 8% over training exclusively with static rubrics across AlpacaEval, GPQA, ArenaHard as well as the validation sets of expert questions and rubrics. We qualitatively analyze the elicited criteria and identify prominent themes such as transparency, practicality, organization, and reasoning.

cs.CL↗

arXiv2025

SHAPE. I. A SOM-SED hybrid approach for efficient galaxy parameter estimation leveraging JWST

With the launch and application of next-generation ground- and space-based telescopes, astronomy has entered the era of big data, necessitating more efficient and robust data analysis methods. Most traditional parameter estimation methods are unable to reconcile differences between photometric systems. Ideally, we would like to optimally rely on high-quality observation data provided by, e.g., JWST, for calibrating and improving upcoming wide-field surveys such as the China Space Station Telescope (CSST) and Euclid. To this end, we introduce a new approach (SHAPE, SOM-SED Hybrid Approach for efficient Parameter Estimation) that can bridge different photometric systems and efficiently estimate key galaxy parameters, such as stellar mass ($M_\star$) and star formation rate (SFR), leveraging data from a large and deep JWST/NIRCam and MIRI survey (PRIMER). As a test of the methodology, we focus on galaxies at $z\sim 1.5-2.5$. To mitigate discrepancies between input colors and the training set, we replace the default SOM weights with stacked SEDs from each cell, extending the applicability of our model to other photometric catalogs (e.g., COSMOS2020). By incorporating a SED library (SED Lib), we apply this JWST-calibrated model to the COSMOS2020 catalog. Despite the limited sample size and potential template-related uncertainties, SOM-derived parameters exhibit a good agreement with results from SED-fitting using extended photometry. Under identical photometric constraints from CSST and Euclid bands, our method outperforms traditional SED-fitting techniques in SFR estimation, exhibiting both a reduced bias (-0.01 vs. 0.18) and a smaller $σ_{\rm NMAD}$ (0.25 vs. 0.35). With its computational efficiency capable of processing $10^6$ sources per CPU per hour during the estimation phase, this JWST-calibrated estimator holds significant promise for next-generation wide-field surveys.

astro-ph.GA↗