Searcharxiv⌕ Search

arXiv subjects

Hai Huang

Publications and source records attributed to Hai Huang.

At least 73 records · Page 4Linked to original sources
arXiv2024

Less is More: on the Over-Globalizing Problem in Graph Transformers

Graph Transformer, due to its global attention mechanism, has emerged as a new tool in dealing with graph-structured data. It is well recognized that the global attention mechanism considers a wider receptive field in a fully connected graph, leading many to believe that useful information can be extracted from all the nodes. In this paper, we challenge this belief: does the globalizing property always benefit Graph Transformers? We reveal the over-globalizing problem in Graph Transformer by presenting both empirical evidence and theoretical analysis, i.e., the current attention mechanism overly focuses on those distant nodes, while the near nodes, which actually contain most of the useful information, are relatively weakened. Then we propose a novel Bi-Level Global Graph Transformer with Collaborative Training (CoBFormer), including the inter-cluster and intra-cluster Transformers, to prevent the over-globalizing problem while keeping the ability to extract valuable information from distant nodes. Moreover, the collaborative training is proposed to improve the model's generalization ability with a theoretical guarantee. Extensive experiments on various graphs well validate the effectiveness of our proposed CoBFormer.

cs.LG↗
arXiv2024

GCC: Generative Calibration Clustering

Deep clustering as an important branch of unsupervised representation learning focuses on embedding semantically similar samples into the identical feature space. This core demand inspires the exploration of contrastive learning and subspace clustering. However, these solutions always rely on the basic assumption that there are sufficient and category-balanced samples for generating valid high-level representation. This hypothesis actually is too strict to be satisfied for real-world applications. To overcome such a challenge, the natural strategy is utilizing generative models to augment considerable instances. How to use these novel samples to effectively fulfill clustering performance improvement is still difficult and under-explored. In this paper, we propose a novel Generative Calibration Clustering (GCC) method to delicately incorporate feature learning and augmentation into clustering procedure. First, we develop a discriminative feature alignment mechanism to discover intrinsic relationship across real and generated samples. Second, we design a self-supervised metric learning to generate more reliable cluster assignment to boost the conditional diffusion generation. Extensive experimental results on three benchmarks validate the effectiveness and advantage of our proposed method over the state-of-the-art methods.

cs.CV↗
arXiv2024

Composite Backdoor Attacks Against Large Language Models

Large language models (LLMs) have demonstrated superior performance compared to previous methods on various tasks, and often serve as the foundation models for many researches and services. However, the untrustworthy third-party LLMs may covertly introduce vulnerabilities for downstream tasks. In this paper, we explore the vulnerability of LLMs through the lens of backdoor attacks. Different from existing backdoor attacks against LLMs, ours scatters multiple trigger keys in different prompt components. Such a Composite Backdoor Attack (CBA) is shown to be stealthier than implanting the same multiple trigger keys in only a single component. CBA ensures that the backdoor is activated only when all trigger keys appear. Our experiments demonstrate that CBA is effective in both natural language processing (NLP) and multimodal tasks. For instance, with $3\%$ poisoning samples against the LLaMA-7B model on the Emotion dataset, our attack achieves a $100\%$ Attack Success Rate (ASR) with a False Triggered Rate (FTR) below $2.06\%$ and negligible model accuracy degradation. Our work highlights the necessity of increased security research on the trustworthiness of foundation LLMs.

cs.CR↗
arXiv2023

Rethinking Evaluation Metrics of Open-Vocabulary Segmentaion

In this paper, we highlight a problem of evaluation metrics adopted in the open-vocabulary segmentation. That is, the evaluation process still heavily relies on closed-set metrics on zero-shot or cross-dataset pipelines without considering the similarity between predicted and ground truth categories. To tackle this issue, we first survey eleven similarity measurements between two categorical words using WordNet linguistics statistics, text embedding, and language models by comprehensive quantitative analysis and user study. Built upon those explored measurements, we designed novel evaluation metrics, namely Open mIoU, Open AP, and Open PQ, tailored for three open-vocabulary segmentation tasks. We benchmarked the proposed evaluation metrics on 12 open-vocabulary methods of three segmentation tasks. Even though the relative subjectivity of similarity distance, we demonstrate that our metrics can still well evaluate the open ability of the existing open-vocabulary segmentation methods. We hope that our work can bring with the community new thinking about how to evaluate the open ability of models. The evaluation code is released in github.

cs.CV↗
arXiv2023

CVPR 2023 Text Guided Video Editing Competition

Humans watch more than a billion hours of video per day. Most of this video was edited manually, which is a tedious process. However, AI-enabled video-generation and video-editing is on the rise. Building on text-to-image models like Stable Diffusion and Imagen, generative AI has improved dramatically on video tasks. But it's hard to evaluate progress in these video tasks because there is no standard benchmark. So, we propose a new dataset for text-guided video editing (TGVE), and we run a competition at CVPR to evaluate models on our TGVE dataset. In this paper we present a retrospective on the competition and describe the winning method. The competition dataset is available at https://sites.google.com/view/loveucvpr23/track4.

cs.CV↗
arXiv2023

Prompt Backdoors in Visual Prompt Learning

Fine-tuning large pre-trained computer vision models is infeasible for resource-limited users. Visual prompt learning (VPL) has thus emerged to provide an efficient and flexible alternative to model fine-tuning through Visual Prompt as a Service (VPPTaaS). Specifically, the VPPTaaS provider optimizes a visual prompt given downstream data, and downstream users can use this prompt together with the large pre-trained model for prediction. However, this new learning paradigm may also pose security risks when the VPPTaaS provider instead provides a malicious visual prompt. In this paper, we take the first step to explore such risks through the lens of backdoor attacks. Specifically, we propose BadVisualPrompt, a simple yet effective backdoor attack against VPL. For example, poisoning $5\%$ CIFAR10 training data leads to above $99\%$ attack success rates with only negligible model accuracy drop by $1.5\%$. In particular, we identify and then address a new technical challenge related to interactions between the backdoor trigger and visual prompt, which does not exist in conventional, model-level backdoors. Moreover, we provide in-depth analyses of seven backdoor defenses from model, prompt, and input levels. Overall, all these defenses are either ineffective or impractical to mitigate our BadVisualPrompt, implying the critical vulnerability of VPL.

cs.CR↗
arXiv2023

Parallel Computing Based Solution for Reliability-Constrained Distribution Network Planning

The main goal of distribution network (DN) expansion planning is essentially to achieve minimal investment constrained with specified reliability requirements. The reliability-constrained distribution network planning (RcDNP) problem can be cast an instance of mixed-integer linear programming (MILP) which involves ultra-heavy computation burden especially for large scale DNs. In this paper, we propose a parallel computing-based acceleration algorithm for solve RcDNP problem. The RcDNP is decomposed into a backbone grid and several lateral grid problems with coordination. Then a parallelizable augmented Lagrangian algorithm with acceleration strategy is developed to solve the coordination planning problems. In this method, the lateral grid problems are solved in parallel through coordinating with the backbone grid planning problem. To address the presence of nonconvexity, Gauss-Seidel iteration is adopted on the convex hull of the feasible region constructed by the decomposition method. Under mild conditions, the optimality and convergence of this algorithm is proven. The numerical tests show the proposed method can significantly reduce the computation time and make the RcDNP applicable for real-world problems.

eess.SY↗
arXiv2022

On the Privacy Risks of Cell-Based NAS Architectures

Existing studies on neural architecture search (NAS) mainly focus on efficiently and effectively searching for network architectures with better performance. Little progress has been made to systematically understand if the NAS-searched architectures are robust to privacy attacks while abundant work has already shown that human-designed architectures are prone to privacy attacks. In this paper, we fill this gap and systematically measure the privacy risks of NAS architectures. Leveraging the insights from our measurement study, we further explore the cell patterns of cell-based NAS architectures and evaluate how the cell patterns affect the privacy risks of NAS-searched architectures. Through extensive experiments, we shed light on how to design robust NAS architectures against privacy attacks, and also offer a general methodology to understand the hidden correlation between the NAS-searched architectures and other privacy risks.

cs.CR↗
arXiv2022

Stabilization of three-dimensional charge order through interplanar orbital hybridization in Pr$_x$Y$_{1-x}$Ba$_2$Cu$_3$O$_{6+δ}$

The shape of 3$d$-orbitals often governs the electronic and magnetic properties of correlated transition metal oxides. In the superconducting cuprates, the planar confinement of the $d_{x^2-y^2}$ orbital dictates the two-dimensional nature of the unconventional superconductivity and a competing charge order. Achieving orbital-specific control of the electronic structure to allow coupling pathways across adjacent planes would enable direct assessment of the role of dimensionality in the intertwined orders. Using Cu-$L_3$ and Pr-$M_5$ resonant x-ray scattering and first-principles calculations, we report a highly correlated three-dimensional charge order in Pr-substituted YBa$_2$Cu$_3$O$_{7}$, where the Pr $f$-electrons create a direct orbital bridge between CuO$_2$ planes. With this, we demonstrate that interplanar orbital engineering can be used to surgically control electronic phases in correlated oxides and other layered materials.

cond-mat.str-el↗
arXiv2021

Sliding Sequential CVAE with Time Variant Socially-aware Rethinking for Trajectory Prediction

Pedestrian trajectory prediction is a key technology in many applications such as video surveillance, social robot navigation, and autonomous driving, and significant progress has been made in this research topic. However, there remain two limitations of previous studies. First, with the continuation of time, the prediction error at each time step increases significantly, causing the final displacement error to be impossible to ignore. Second, the prediction results of multiple pedestrians might be impractical in the prediction horizon, i.e., the predicted trajectories might collide with each other. To overcome these limitations, this work proposes a novel trajectory prediction method called CSR, which consists of a cascaded conditional variational autoencoder (CVAE) module and a socially-aware regression module. The cascaded CVAE module first estimates the future trajectories in a sequential pattern. Specifically, each CVAE concatenates the past trajectories and the predicted points so far as the input and predicts the location at the following time step. Then, the socially-aware regression module generates offsets from the estimated future trajectories to produce the socially compliant final predictions, which are more reasonable and accurate results than the estimated trajectories. Moreover, considering the large model parameters of the cascaded CVAE module, a slide CVAE module is further exploited to improve the model efficiency using one shared CVAE, in a slidable manner. Experiments results demonstrate that the proposed method exhibits improvements over state-of-the-art method on the Stanford Drone Dataset (SDD) and ETH/UCY of approximately 38.0% and 22.2%, respectively.

cs.CV↗
arXiv2021

The Constraint of H0from Galaxy clusters and Hubble parameter data

Using comoving distance $d_c$ and angular diameter distance $d_A$, we recalculate parameters describing kinematical state of the universe, still combining the kinematical model of universe but not relying on dynamical equations for gravity. Comoving distance $d_c$ comes from Hubble data H(z) and is more reliable. Angular diameter distance $d_A$ comes from SZE (Sunyaev-Zel dovich Effect) and X-ray data, and needs calibration. In low redshift case, we use expansion of relation between luminosity distance and redshift about redshift $z$; in high redshift case, we take variable substitution $y=1/(1+z)$, and expand the relation between luminosity distance and redshift about variable $y$ in order to reduce computational errors. Finally we get the more precise value of Hubble parameter $H_0=69.13\pm 0.24{\kern 1pt} km{\kern 1pt} \cdot {s^{ - 1}} \cdot Mp{c^{ - 1}}$, corresponding to 0.4\% uncertainty in $68.3\%$ confidence region, also deceleration factor $q_0=-0.57\pm0.04 $ and acceleration rate $j_0=1.28\pm0.33$, and their statistical values and probability graph. We compare the values of ${H_0}$, ${q_0}$, ${j_0}$ with those obtained from other observation data and model.

astro-ph.CO↗
arXiv2021

Search for $Q \sim 0$ order near a forbidden Bragg position in Bi$_{2.1}$Sr$_{1.9}$CaCu$_2$O$_{8+x}$ with resonant soft x-ray scattering

Identifying what broken symmetries are present in the cuprates has become a major area of research. Many authors have reported evidence for so-called "$Q \sim 0$" order that involves broken inversion, mirror, chiral, or time-reversal symmetry that is uniform in space. Not all these observations are well understood and new experimental probes are needed. Here we use resonant soft x-ray scattering (RSXS) to search for $Q \sim 0$ order in Bi$_{2.1}$Sr$_{1.9}$CaCu$_2$O$_{8+x}$ (Bi-2212) by measuring the region of a forbidden Bragg peak, $(0,0,3)$, which is normally extinguished by symmetry but may become allowed on resonance if valence band order is present. Using circularly polarized light, we found that this reflection becomes allowed on the Cu $L_3$ resonance for temperatures $T_c < T < T^\ast$, though remains absent in linear polarization and at other temperatures. This observation suggests the existence of spatially uniform valence band order near the pseudogap temperature. In addition, we observed periodic oscillations in the specular reflectivity from the sample surface that resemble thin film interference fringes, though no known film is present. These fringes are highly resonant, appear in all polarizations, and exhibit a period that depends on the location where the beam strikes the sample surface. We speculate that these fringes arise from interaction between some intrinsic valence band instability and extrinsic structural surface morphologies of the material. Our study supports the existence of some kind of $Q \sim 0$ broken symmetry state in Bi-2212 at intermediate temperatures, and calls for further study using a microfocused beam that could disentangle microscopic effects from macroscopic heterogeneities.

cond-mat.str-el↗
arXiv2021

Data Poisoning Attacks to Deep Learning Based Recommender Systems

Recommender systems play a crucial role in helping users to find their interested information in various web services such as Amazon, YouTube, and Google News. Various recommender systems, ranging from neighborhood-based, association-rule-based, matrix-factorization-based, to deep learning based, have been developed and deployed in industry. Among them, deep learning based recommender systems become increasingly popular due to their superior performance. In this work, we conduct the first systematic study on data poisoning attacks to deep learning based recommender systems. An attacker's goal is to manipulate a recommender system such that the attacker-chosen target items are recommended to many users. To achieve this goal, our attack injects fake users with carefully crafted ratings to a recommender system. Specifically, we formulate our attack as an optimization problem, such that the injected ratings would maximize the number of normal users to whom the target items are recommended. However, it is challenging to solve the optimization problem because it is a non-convex integer programming problem. To address the challenge, we develop multiple techniques to approximately solve the optimization problem. Our experimental results on three real-world datasets, including small and large datasets, show that our attack is effective and outperforms existing attacks. Moreover, we attempt to detect fake users via statistical analysis of the rating patterns of normal and fake users. Our results show that our attack is still effective and outperforms existing attacks even if such a detector is deployed.

cs.CR↗
arXiv2019

A GPU-accelerated package for simulation of flow in nanoporous source rocks with many-body dissipative particle dynamics

Mesoscopic simulations of hydrocarbon flow in source shales are challenging, in part due to the heterogeneous shale pores with sizes ranging from a few nanometers to a few micrometers. Additionally, the sub-continuum fluid-fluid and fluid-solid interactions in nano- to micro-scale shale pores, which are physically and chemically sophisticated, must be captured. To address those challenges, we present a GPU-accelerated package for simulation of flow in nano- to micro-pore networks with a many-body dissipative particle dynamics (mDPD) mesoscale model. Based on a fully distributed parallel paradigm, the code offloads all intensive workloads on GPUs. Other advancements, such as smart particle packing and no-slip boundary condition in complex pore geometries, are also implemented for the construction and the simulation of the realistic shale pores from 3D nanometer-resolution stack images. Our code is validated for accuracy and compared against the CPU counterpart for speedup. In our benchmark tests, the code delivers nearly perfect strong scaling and weak scaling (with up to 512 million particles) on up to 512 K20X GPUs on Oak Ridge National Laboratory's (ORNL) Titan supercomputer. Moreover, a single-GPU benchmark on ORNL's SummitDev and IBM's AC922 suggests that the host-to-device NVLink can boost performance over PCIe by a remarkable 40\%. Lastly, we demonstrate, through a flow simulation in realistic shale pores, that the CPU counterpart requires 840 Power9 cores to rival the performance delivered by our package with four V100 GPUs on ORNL's Summit architecture. This simulation package enables quick-turnaround and high-throughput mesoscopic numerical simulations for investigating complex flow phenomena in nano- to micro-porous rocks with realistic pore geometries.

physics.comp-ph↗
arXiv2018

Phase field and level set methods for modeling solute precipitation and/or dissolution

The dynamics of solid-liquid interfaces controlled by solute precipitation and/or dissolution due to the chemical reaction at the interface were computed in two dimensions using a phase field models. Sharp-interface asymptotic analysis demonstrated that the phase field solutions should converge to the proper sharp-interface precipitation/dissolution limit. For the purpose of comparison, the numerical solution of the sharp-interface model for solute precipitation/dissolution was directly solved using a level set method. In general, the phase field results are found in good agreement with the level set results for all reaction rates and geometry configurations investigated. Present study supports the applications of both methods to more complicated and realistic reactive systems, including the nuclear waste release and mineral precipitation and dissolution.

physics.comp-ph↗
arXiv2018

Persistent low-energy phonon broadening near the charge order $q$-vector in bilayer cuprate Bi$_2$Sr$_2$CaCu$_2$O$_{8+δ}$

We report a persistent low-energy phonon broadening around $q_{B} \sim 0.28$ r.l.u. along the Cu-O bond direction in the high-$T_c$ cuprate Bi$_2$Sr$_2$CaCu$_2$O$_{8+δ}$ (Bi-2212). We show that such broadening exists both inside and outside the conventional charge density wave (CDW) phase, via temperature dependent measurements in both underdoped and heavily overdoped samples. Combining inelastic hard x-ray scattering, diffuse scattering, angle-resolved photoemission spectroscopy, and resonant soft x-ray scattering at the Cu $L_3$-edge, we exclude the presence of a CDW in the heavily overdoped Bi-2212 similar to that observed in the underdoped systems. Finally, we discuss the origin of such anisotropic low-energy phonon broadening, and its potential precursory role to the CDW phase in the underdoped region.

cond-mat.str-el↗
arXiv2015

Ultrasensitive and broadband MoS2 photodetector driven by ferroelectrics

Photodetectors based on two dimensional materials have attracted growing interest. However, the sensitivity is still unsatisfactory even under high gate voltage. Here we demonstrate a MoS2 photodetector with a poly(vinylidene fluoride-trifluoroethylene) ferroelectric layer in place of the oxide layer in a traditional field effect transistor. The dark current of the photodetector is strongly suppressed by ferroelectric polarization. A high detectivity 2.21012 Jones) and photoresponsitivity (2570 A W) detector has been achieved under ZERO gate bias at a wavelength of 635 nm. Most strikingly, the band gap of few-layer MoS2 can be tuned by the ultra-high electrostatic field from the ferroelectric polarization. With this characteristic, photoresponse wavelengths of the photodetector are extended into the near infrared (0.85-1.55m). A ferroelectrics optoelectronics hybrid structure is an effective way to achieve high performance 2D electronic optoelectronic devices.

cond-mat.mes-hall↗