arXiv subjects

Jiahao Wu

Publications and source records attributed to Jiahao Wu.

At least 55 records · Page 3Linked to original sources

arXiv2025

A Comprehensive Survey in LLM(-Agent) Full Stack Safety: Data, Training and Deployment

The remarkable success of Large Language Models (LLMs) has illuminated a promising pathway toward achieving Artificial General Intelligence for both academic and industrial communities, owing to their unprecedented performance across various applications. As LLMs continue to gain prominence in both research and commercial domains, their security and safety implications have become a growing concern, not only for researchers and corporations but also for every nation. Currently, existing surveys on LLM safety primarily focus on specific stages of the LLM lifecycle, e.g., deployment phase or fine-tuning phase, lacking a comprehensive understanding of the entire "lifechain" of LLMs. To address this gap, this paper introduces, for the first time, the concept of "full-stack" safety to systematically consider safety issues throughout the entire process of LLM training, deployment, and eventual commercialization. Compared to the off-the-shelf LLM safety surveys, our work demonstrates several distinctive advantages: (I) Comprehensive Perspective. We define the complete LLM lifecycle as encompassing data preparation, pre-training, post-training, deployment and final commercialization. To our knowledge, this represents the first safety survey to encompass the entire lifecycle of LLMs. (II) Extensive Literature Support. Our research is grounded in an exhaustive review of over 800+ papers, ensuring comprehensive coverage and systematic organization of security issues within a more holistic understanding. (III) Unique Insights. Through systematic literature analysis, we have developed reliable roadmaps and perspectives for each chapter. Our work identifies promising research directions, including safety in data generation, alignment techniques, model editing, and LLM-based agent systems. These insights provide valuable guidance for researchers pursuing future work in this field.

cs.CR↗

arXiv2025

Xinyu AI Search: Enhanced Relevance and Comprehensive Results with Rich Answer Presentations

Traditional search engines struggle to synthesize fragmented information for complex queries, while generative AI search engines face challenges in relevance, comprehensiveness, and presentation. To address these limitations, we introduce Xinyu AI Search, a novel system that incorporates a query-decomposition graph to dynamically break down complex queries into sub-queries, enabling stepwise retrieval and generation. Our retrieval pipeline enhances diversity through multi-source aggregation and query expansion, while filtering and re-ranking strategies optimize passage relevance. Additionally, Xinyu AI Search introduces a novel approach for fine-grained, precise built-in citation and innovates in result presentation by integrating timeline visualization and textual-visual choreography. Evaluated on recent real-world queries, Xinyu AI Search outperforms eight existing technologies in human assessments, excelling in relevance, comprehensiveness, and insightfulness. Ablation studies validate the necessity of its key sub-modules. Our work presents the first comprehensive framework for generative AI search engines, bridging retrieval, generation, and user-centric presentation.

cs.IR↗

arXiv2025

Safe Delta: Consistently Preserving Safety when Fine-Tuning LLMs on Diverse Datasets

Large language models (LLMs) have shown great potential as general-purpose AI assistants across various domains. To fully leverage this potential in specific applications, many companies provide fine-tuning API services, enabling users to upload their own data for LLM customization. However, fine-tuning services introduce a new safety threat: user-uploaded data, whether harmful or benign, can break the model's alignment, leading to unsafe outputs. Moreover, existing defense methods struggle to address the diversity of fine-tuning datasets (e.g., varying sizes, tasks), often sacrificing utility for safety or vice versa. To address this issue, we propose Safe Delta, a safety-aware post-training defense method that adjusts the delta parameters (i.e., the parameter change before and after fine-tuning). Specifically, Safe Delta estimates the safety degradation, selects delta parameters to maximize utility while limiting overall safety loss, and applies a safety compensation vector to mitigate residual safety loss. Through extensive experiments on four diverse datasets with varying settings, our approach consistently preserves safety while ensuring that the utility gain from benign datasets remains unaffected.

cs.LG↗

arXiv2025

SemDiff: Generating Natural Unrestricted Adversarial Examples via Semantic Attributes Optimization in Diffusion Models

Unrestricted adversarial examples (UAEs), allow the attacker to create non-constrained adversarial examples without given clean samples, posing a severe threat to the safety of deep learning models. Recent works utilize diffusion models to generate UAEs. However, these UAEs often lack naturalness and imperceptibility due to simply optimizing in intermediate latent noises. In light of this, we propose SemDiff, a novel unrestricted adversarial attack that explores the semantic latent space of diffusion models for meaningful attributes, and devises a multi-attributes optimization approach to ensure attack success while maintaining the naturalness and imperceptibility of generated UAEs. We perform extensive experiments on four tasks on three high-resolution datasets, including CelebA-HQ, AFHQ and ImageNet. The results demonstrate that SemDiff outperforms state-of-the-art methods in terms of attack success rate and imperceptibility. The generated UAEs are natural and exhibit semantically meaningful changes, in accord with the attributes' weights. In addition, SemDiff is found capable of evading different defenses, which further validates its effectiveness and threatening.

cs.LG↗

arXiv2025

Dataset Condensation for Recommendation

Training recommendation models on large datasets requires significant time and resources. It is desired to construct concise yet informative datasets for efficient training. Recent advances in dataset condensation show promise in addressing this problem by synthesizing small datasets. However, applying existing methods of dataset condensation to recommendation has limitations: (1) they fail to generate discrete user-item interactions, and (2) they could not preserve users' potential preferences. To address the limitations, we propose a lightweight condensation framework tailored for recommendation (DConRec), focusing on condensing user-item historical interaction sets. Specifically, we model the discrete user-item interactions via a probabilistic approach and design a pre-augmentation module to incorporate the potential preferences of users into the condensed datasets. While the substantial size of datasets leads to costly optimization, we propose a lightweight policy gradient estimation to accelerate the data synthesis. Experimental results on multiple real-world datasets have demonstrated the effectiveness and efficiency of our framework. Besides, we provide a theoretical analysis of the provable convergence of DConRec. Our implementation is available at: https://github.com/JiahaoWuGit/DConRec.

cs.IR↗

arXiv2025

Backdoor Graph Condensation

Graph condensation has recently emerged as a prevalent technique to improve the training efficiency for graph neural networks (GNNs). It condenses a large graph into a small one such that a GNN trained on this small synthetic graph can achieve comparable performance to a GNN trained on the large graph. However, while existing graph condensation studies mainly focus on the best trade-off between graph size and the GNNs' performance (model utility), they overlook the security issues of graph condensation. To bridge this gap, we first explore backdoor attack against the GNNs trained on the condensed graphs. We introduce an effective backdoor attack against graph condensation, termed BGC. This attack aims to (1) preserve the condensed graph quality despite trigger injection, and (2) ensure trigger efficacy through the condensation process, achieving a high attack success rate. Specifically, BGC consistently updates triggers during condensation and targets representative nodes for poisoning. Extensive experiments demonstrate the effectiveness of our attack. BGC achieves a high attack success rate (close to 1.0) and good model utility in all cases. Furthermore, the results against multiple defense methods demonstrate BGC's resilience under their defenses. Finally, we analyze the key hyperparameters that influence the attack performance. Our code is available at: https://github.com/JiahaoWuGit/BGC.

cs.LG↗

arXiv2025

Instant Gaussian Stream: Fast and Generalizable Streaming of Dynamic Scene Reconstruction via Gaussian Splatting

Building Free-Viewpoint Videos in a streaming manner offers the advantage of rapid responsiveness compared to offline training methods, greatly enhancing user experience. However, current streaming approaches face challenges of high per-frame reconstruction time (10s+) and error accumulation, limiting their broader application. In this paper, we propose Instant Gaussian Stream (IGS), a fast and generalizable streaming framework, to address these issues. First, we introduce a generalized Anchor-driven Gaussian Motion Network, which projects multi-view 2D motion features into 3D space, using anchor points to drive the motion of all Gaussians. This generalized Network generates the motion of Gaussians for each target frame in the time required for a single inference. Second, we propose a Key-frame-guided Streaming Strategy that refines each key frame, enabling accurate reconstruction of temporally complex scenes while mitigating error accumulation. We conducted extensive in-domain and cross-domain evaluations, demonstrating that our approach can achieve streaming with a average per-frame reconstruction time of 2s+, alongside a enhancement in view synthesis quality.

cs.CV↗

arXiv2025

Swift4D:Adaptive divide-and-conquer Gaussian Splatting for compact and efficient reconstruction of dynamic scene

Novel view synthesis has long been a practical but challenging task, although the introduction of numerous methods to solve this problem, even combining advanced representations like 3D Gaussian Splatting, they still struggle to recover high-quality results and often consume too much storage memory and training time. In this paper we propose Swift4D, a divide-and-conquer 3D Gaussian Splatting method that can handle static and dynamic primitives separately, achieving a good trade-off between rendering quality and efficiency, motivated by the fact that most of the scene is the static primitive and does not require additional dynamic properties. Concretely, we focus on modeling dynamic transformations only for the dynamic primitives which benefits both efficiency and quality. We first employ a learnable decomposition strategy to separate the primitives, which relies on an additional parameter to classify primitives as static or dynamic. For the dynamic primitives, we employ a compact multi-resolution 4D Hash mapper to transform these primitives from canonical space into deformation space at each timestamp, and then mix the static and dynamic primitives to produce the final output. This divide-and-conquer method facilitates efficient training and reduces storage redundancy. Our method not only achieves state-of-the-art rendering quality while being 20X faster in training than previous SOTA methods with a minimum storage requirement of only 30MB on real-world datasets. Code is available at https://github.com/WuJH2001/swift4d.

cs.CV↗

arXiv2025

AI Generations: From AI 1.0 to AI 4.0

This paper proposes that Artificial Intelligence (AI) progresses through several overlapping generations: AI 1.0 (Information AI), AI 2.0 (Agentic AI), AI 3.0 (Physical AI), and now a speculative AI 4.0 (Conscious AI). Each of these AI generations is driven by shifting priorities among algorithms, computing power, and data. AI 1.0 ushered in breakthroughs in pattern recognition and information processing, fueling advances in computer vision, natural language processing, and recommendation systems. AI 2.0 built on these foundations through real-time decision-making in digital environments, leveraging reinforcement learning and adaptive planning for agentic AI applications. AI 3.0 extended intelligence into physical contexts, integrating robotics, autonomous vehicles, and sensor-fused control systems to act in uncertain real-world settings. Building on these developments, AI 4.0 puts forward the bold vision of self-directed AI capable of setting its own goals, orchestrating complex training regimens, and possibly exhibiting elements of machine consciousness. This paper traces the historical foundations of AI across roughly seventy years, mapping how changes in technological bottlenecks from algorithmic innovation to high-performance computing to specialized data, have spurred each generational leap. It further highlights the ongoing synergies among AI 1.0, 2.0, 3.0, and 4.0, and explores the profound ethical, regulatory, and philosophical challenges that arise when artificial systems approach (or aspire to) human-like autonomy. Ultimately, understanding these evolutions and their interdependencies is pivotal for guiding future research, crafting responsible governance, and ensuring that AI transformative potential benefits society as a whole.

cs.AI↗

arXiv2025

TF-DCon: Leveraging Large Language Models (LLMs) to Empower Training-Free Dataset Condensation for Content-Based Recommendation

Modern techniques in Content-based Recommendation (CBR) leverage item content information to provide personalized services to users, but suffer from resource-intensive training on large datasets. To address this issue, we explore the dataset condensation for textual CBR in this paper. The goal of dataset condensation is to synthesize a small yet informative dataset, upon which models can achieve performance comparable to those trained on large datasets. While existing condensation approaches are tailored to classification tasks for continuous data like images or embeddings, direct application of them to CBR has limitations. To bridge this gap, we investigate efficient dataset condensation for content-based recommendation. Inspired by the remarkable abilities of large language models (LLMs) in text comprehension and generation, we leverage LLMs to empower the generation of textual content during condensation. To handle the interaction data involving both users and items, we devise a dual-level condensation method: content-level and user-level. At content-level, we utilize LLMs to condense all contents of an item into a new informative title. At user-level, we design a clustering-based synthesis module, where we first utilize LLMs to extract user interests. Then, the user interests and user embeddings are incorporated to condense users and generate interactions for condensed users. Notably, the condensation paradigm of this method is forward and free from iterative optimization on the synthesized dataset. Extensive empirical findings from our study, conducted on three authentic datasets, substantiate the efficacy of the proposed method. Particularly, we are able to approximate up to 97% of the original performance while reducing the dataset size by 95% (i.e., on dataset MIND).

cs.IR↗

arXiv2024

The second order Huang-Yang formula to the 3D Fermi gas: the Gross-Pitaevskii regime

For a system of $N$ Fermions of spin $1/2$, with its interaction potential of scattering length $a$, the classical Huang-Yang formula states that the energy density $e(ρ)$ is of the form \begin{equation*} e(ρ)=\frac{3}{5}(3π^2)^{\frac{2}{3}}ρ^{\frac{5}{3}}+2πaρ^2 +\frac{12}{35}(11-2\ln2)3^{\frac{1}{3}}π^{\frac{2}{3}}a^2ρ^{\frac{7}{3}} +o(ρ^{\frac{7}{3}}). \end{equation*} We consider a general system of $N$ Fermions of spin $1/\mathbf{q}$ with the scattering length of the interaction potential at the scale $a\propto N^{-1}$, that is, in the Gross-Pitaevskii regime. We prove the 2nd order ground state energy approximation corresponding to the Huang-Yang formula. The thermodynamic limit case shares a similar logic, and could be dealt with in a separate paper.

math-ph↗

arXiv2024

Simultaneous upper and lower bounds of American-style option prices with hedging via neural networks

In this paper, we introduce two novel methods to solve the American-style option pricing problem and its dual form at the same time using neural networks. Without applying nested Monte Carlo, the first method uses a series of neural networks to simultaneously compute both the lower and upper bounds of the option price, and the second one accomplishes the same goal with one global network. The avoidance of extra simulations and the use of neural networks significantly reduce the computational complexity and allow us to price Bermudan options with frequent exercise opportunities in high dimensions, as illustrated by the provided numerical experiments. As a by-product, these methods also derive a hedging strategy for the option, which can also be used as a control variate for variance reduction.

q-fin.CP↗

arXiv2024

KH-PINN: Physics-informed neural networks for Kelvin-Helmholtz instability with spatiotemporal and magnitude multiscale

Prediction of Kelvin-Helmholtz instability (KHI) is crucial across various fields, requiring extensive high-fidelity data. However, experimental data are often sparse and noisy, while simulated data may lack credibility due to discrepancies with real-world configurations and parameters. This underscores the need for field reconstruction and parameter inference from sparse, noisy data, which constitutes inverse problems. Based on the physics-informed neural networks (PINNs), the KH-PINN framework is established in this work to solve the inverse problems of KHI flows. By incorporating the governing physical equations, KH-PINN reconstructs continuous flow fields and infer unknown transport parameters from sparse, noisy observed data. The 2D unsteady incompressible flows with both constant and variable densities are studied. To our knowledge, this is the first application of PINNs to unsteady incompressible flows with variable densities. To address the spatiotemporal multiscale issue and enhance the reconstruction accuracy of small-scale structures, the multiscale embedding (ME) strategy is adopted. To address the magnitude multiscale issue and enhance the reconstruction accuracy of small-magnitude velocities, which are critical for KHI problems, the small-velocity amplification (SVA) strategy is proposed. The results demonstrate that KH-PINN can accurately reconstruct the fields with complex, evolving vortices and infer unknown parameters across a broad range of Reynolds numbers. Additionally, the energy-decaying and entropy-increasing curves are accurately obtained. The effectiveness of ME and SVA is validated through comparative studies, and the anti-noise and few-shot learning capabilities of KH-PINN are also validated. The code for this work is available at https://github.com/CAME-THU/KH-PINN.

physics.flu-dyn↗

arXiv2024

HDRGS: High Dynamic Range Gaussian Splatting

Recent years have witnessed substantial advancements in the field of 3D reconstruction from 2D images, particularly following the introduction of the neural radiance field (NeRF) technique. However, reconstructing a 3D high dynamic range (HDR) radiance field, which aligns more closely with real-world conditions, from 2D multi-exposure low dynamic range (LDR) images continues to pose significant challenges. Approaches to this issue fall into two categories: grid-based and implicit-based. Implicit methods, using multi-layer perceptrons (MLP), face inefficiencies, limited solvability, and overfitting risks. Conversely, grid-based methods require significant memory and struggle with image quality and long training times. In this paper, we introduce Gaussian Splatting-a recent, high-quality, real-time 3D reconstruction technique-into this domain. We further develop the High Dynamic Range Gaussian Splatting (HDR-GS) method, designed to address the aforementioned challenges. This method enhances color dimensionality by including luminance and uses an asymmetric grid for tone-mapping, swiftly and precisely converting pixel irradiance to color. Our approach improves HDR scene recovery accuracy and integrates a novel coarse-to-fine strategy to speed up model convergence, enhancing robustness against sparse viewpoints and exposure extremes, and preventing local optima. Extensive testing confirms that our method surpasses current state-of-the-art techniques in both synthetic and real-world scenarios.

cs.CV↗

arXiv2024

Harmonizing Material Quantity and Terahertz Wave Interference Shielding Efficiency with Metallic Borophene Nanosheets

Materials with electromagnetic interference (EMI) shielding in the terahertz (THz) regime, while minimizing the quantity used, are highly demanded for future information communication, healthcare and mineral resource exploration applications. Currently, there is often a trade-off between the amount of material used and the absolute EMI shielding effectiveness (EESt) for the EMI shielding materials. Here, we address this trade-off by harnessing the unique properties of two-dimensional (2D) beta12-borophene (beta12-Br) nanosheets. Leveraging beta12-Br's light weight and exceptional electron mobility characteristics, which represent among the highest reported values to date, we simultaneously achieve a THz EMI shield effectiveness (SE) of 70 dB and an EESt of 4.8E5 dB cm^2/g (@0.87 THz) using a beta12-Br polymer composite. This surpasses the values of previously reported THz shielding materials with an EESt less than 3E5 dB cm^2/g and a SE smaller than 60 dB, while only needs 0.1 wt.% of these materials to realize the same SE value. Furthermore, by capitalizing on the composite's superior mechanical properties, with 158% tensile strain at a Young's modulus of 33 MPa, we demonstrate the high-efficiency shielding performances of conformably coated surfaces based on beta12-Br nanosheets, suggesting their great potential in EMI shielding area.

physics.app-ph↗

arXiv2024

Adapt2Reward: Adapting Video-Language Models to Generalizable Robotic Rewards via Failure Prompts

For a general-purpose robot to operate in reality, executing a broad range of instructions across various environments is imperative. Central to the reinforcement learning and planning for such robotic agents is a generalizable reward function. Recent advances in vision-language models, such as CLIP, have shown remarkable performance in the domain of deep learning, paving the way for open-domain visual recognition. However, collecting data on robots executing various language instructions across multiple environments remains a challenge. This paper aims to transfer video-language models with robust generalization into a generalizable language-conditioned reward function, only utilizing robot video data from a minimal amount of tasks in a singular environment. Unlike common robotic datasets used for training reward functions, human video-language datasets rarely contain trivial failure videos. To enhance the model's ability to distinguish between successful and failed robot executions, we cluster failure video features to enable the model to identify patterns within. For each cluster, we integrate a newly trained failure prompt into the text encoder to represent the corresponding failure mode. Our language-conditioned reward function shows outstanding generalization to new environments and new instructions for robot planning and reinforcement learning.

cs.CV↗

arXiv2024

Wideband Coherent Microwave Conversion via Magnon Nonlinearity in Hybrid Quantum System

Frequency conversion is a widely realized physical process in nonlinear systems of optics and electronics. As an emerging nonlinear platform, spintronic devices have the potential to achieve stronger frequency conversion. Here, we demonstrated a microwave frequency conversion method in a hybrid quantum system, integrating nitrogen-vacancy centers in diamond with magnetic thin film CoFeB. We achieve a conversion bandwidth ranging from 0.1 to 12GHz, presenting an up to $\mathrm{25^{th}}$ order frequency conversion and further display the application of this method for frequency detection and qubits coherent control. Distinct from traditional frequency conversion techniques based on nonlinear electric response, our approach employs nonlinear magnetic response in spintronic devices. The nonlinearity, originating from the symmetry breaking such as domain walls in magnetic films, presents that our method can be adapted to hybrid systems of other spintronic devices and spin qubits, expanding the application scope of spintronic devices and providing a promising on-chip platform for coupling quantum systems.

quant-ph↗

arXiv2024

FlamePINN-1D: Physics-informed neural networks to solve forward and inverse problems of 1D laminar flames

Given the existence of various forward and inverse problems in combustion studies and applications that necessitate distinct methods for resolution, a framework to solve them in a unified way is critically needed. A promising approach is the integration of machine learning methods with governing equations of combustion systems, which exhibits superior generality and few-shot learning ability compared to purely data-driven methods. In this work, the FlamePINN-1D framework is proposed to solve the forward and inverse problems of 1D laminar flames based on physics-informed neural networks. Three cases with increasing complexity have been tested: Case 1 are freely-propagating premixed (FPP) flames with simplified physical models, while Case 2 and Case 3 are FPP and counterflow premixed (CFP) flames with detailed models, respectively. For forward problems, FlamePINN-1D aims to solve the flame fields and infer the unknown eigenvalues (such as laminar flame speeds) under the constraints of governing equations and boundary conditions. For inverse problems, FlamePINN-1D aims to reconstruct the continuous fields and infer the unknown parameters (such as transport and chemical kinetics parameters) from noisy sparse observations of the flame. Our results strongly validate these capabilities of FlamePINN-1D across various flames and working conditions. Compared to traditional methods, FlamePINN-1D is differentiable and mesh-free, exhibits no discretization errors, and is easier to implement for inverse problems. The inverse problem results also indicate the possibility of optimizing chemical mechanisms from measurements of laboratory 1D flames. Furthermore, some proposed strategies, such as hard constraints and thin-layer normalization, are proven to be essential for the robust learning of FlamePINN-1D. The code for this paper is partially available at https://github.com/CAME-THU/FlamePINN-1D.

cs.LG↗