Searcharxiv⌕ Search

arXiv subjects

Min Chen

Publications and source records attributed to Min Chen.

At least 145 records · Page 8Linked to original sources
arXiv2023

PrivGraph: Differentially Private Graph Data Publication by Exploiting Community Information

Graph data is used in a wide range of applications, while analyzing graph data without protection is prone to privacy breach risks. To mitigate the privacy risks, we resort to the standard technique of differential privacy to publish a synthetic graph. However, existing differentially private graph synthesis approaches either introduce excessive noise by directly perturbing the adjacency matrix, or suffer significant information loss during the graph encoding process. In this paper, we propose an effective graph synthesis algorithm PrivGraph by exploiting the community information. Concretely, PrivGraph differentially privately partitions the private graph into communities, extracts intra-community and inter-community information, and reconstructs the graph from the extracted graph information. We validate the effectiveness of PrivGraph on six real-world graph datasets and seven commonly used graph metrics.

cs.CR↗
arXiv2023

ORL-AUDITOR: Dataset Auditing in Offline Deep Reinforcement Learning

Data is a critical asset in AI, as high-quality datasets can significantly improve the performance of machine learning models. In safety-critical domains such as autonomous vehicles, offline deep reinforcement learning (offline DRL) is frequently used to train models on pre-collected datasets, as opposed to training these models by interacting with the real-world environment as the online DRL. To support the development of these models, many institutions make datasets publicly available with opensource licenses, but these datasets are at risk of potential misuse or infringement. Injecting watermarks to the dataset may protect the intellectual property of the data, but it cannot handle datasets that have already been published and is infeasible to be altered afterward. Other existing solutions, such as dataset inference and membership inference, do not work well in the offline DRL scenario due to the diverse model behavior characteristics and offline setting constraints. In this paper, we advocate a new paradigm by leveraging the fact that cumulative rewards can act as a unique identifier that distinguishes DRL models trained on a specific dataset. To this end, we propose ORL-AUDITOR, which is the first trajectory-level dataset auditing mechanism for offline RL scenarios. Our experiments on multiple offline DRL models and tasks reveal the efficacy of ORL-AUDITOR, with auditing accuracy over 95% and false positive rates less than 2.88%. We also provide valuable insights into the practical implementation of ORL-AUDITOR by studying various parameter settings. Furthermore, we demonstrate the auditing capability of ORL-AUDITOR on open-source datasets from Google and DeepMind, highlighting its effectiveness in auditing published datasets. ORL-AUDITOR is open-sourced at https://github.com/link-zju/ORL-Auditor.

cs.CR↗
arXiv2023

Simulation-Based Optimization of User Interfaces for Quality-Assuring Machine Learning Model Predictions

Quality-sensitive applications of machine learning (ML) require quality assurance (QA) by humans before the predictions of an ML model can be deployed. QA for ML (QA4ML) interfaces require users to view a large amount of data and perform many interactions to correct errors made by the ML model. An optimized user interface (UI) can significantly reduce interaction costs. While UI optimization can be informed by user studies evaluating design options, this approach is not scalable because there are typically numerous small variations that can affect the efficiency of a QA4ML interface. Hence, we propose using simulation to evaluate and aid the optimization of QA4ML interfaces. In particular, we focus on simulating the combined effects of human intelligence in initiating appropriate interaction commands and machine intelligence in providing algorithmic assistance for accelerating QA4ML processes. As QA4ML is usually labor-intensive, we use the simulated task completion time as the metric for UI optimization under different interface and algorithm setups. We demonstrate the usage of this UI design method in several QA4ML applications.

cs.HC↗
arXiv2023

Fine-grained Spatio-Temporal Distribution Prediction of Mobile Content Delivery in 5G Ultra-Dense Networks

The 5G networks have extensively promoted the growth of mobile users and novel applications, and with the skyrocketing user requests for a large amount of popular content, the consequent content delivery services (CDSs) have been bringing a heavy load to mobile service providers. As a key mission in intelligent networks management, understanding and predicting the distribution of CDSs benefits many tasks of modern network services such as resource provisioning and proactive content caching for content delivery networks. However, the revolutions in novel ubiquitous network architectures led by ultra-dense networks (UDNs) make the task extremely challenging. Specifically, conventional methods face the challenges of insufficient spatio precision, lacking generalizability, and complex multi-feature dependencies of user requests, making their effectiveness unreliable in CDSs prediction under 5G UDNs. In this paper, we propose to adopt a series of encoding and sampling methods to model CDSs of known and unknown areas at a tailored fine-grained level. Moreover, we design a spatio-temporal-social multi-feature extraction framework for CDSs hotspots prediction, in which a novel edge-enhanced graph convolution block is proposed to encode dynamic CDSs networks based on the social relationships and the spatio features. Besides, we introduce the Long-Short Term Memory (LSTM) to further capture the temporal dependency. Extensive performance evaluations with real-world measurement data collected in two mobile content applications demonstrate the effectiveness of our proposed solution, which can improve the prediction area under the curve (AUC) by 40.5% compared to the state-of-the-art proposals at a spatio granularity of 76m, with up to 80% of the unknown areas.

cs.NI↗
arXiv2023

Radial Icicle Tree (RIT): Node Separation and Area Constancy

Icicles and sunbursts are two commonly-used visual representations of trees. While icicle trees can map data values faithfully to rectangles of different sizes, often some rectangles are too narrow to be noticed easily. When an icicle tree is transformed into a sunburst tree, the width of each rectangle becomes the length of an annular sector that is usually longer than the original width. While sunburst trees alleviate the problem of narrow rectangles in icicle trees, it no longer maintains the consistency of size encoding. At different tree depths, nodes of the same data values are displayed in annular sections of different sizes in a sunburst tree, though they are represented by rectangles of the same size in an icicle tree. Furthermore, two nodes from different subtrees could sometimes appear as a single node in both icicle trees and sunburst trees. In this paper, we propose a new visual representation, referred to as \emph{radial icicle tree} (RIT), which transforms the rectangular bounding box of an icicle tree into a circle, circular sector, or annular sector while introducing gaps between nodes and maintaining area constancy for nodes of the same size. We applied the new visual design to several datasets. Both the analytical design process and user-centered evaluation have confirmed that this new design has improved the design of icicles and sunburst trees without introducing any relative demerit.

cs.HC↗
arXiv2023

Collimated hot electron generation from sub-wavelength grating target irradiated by a femtosecond laser pulse of relativistic intensity

We investigate the production of hot electrons from the interaction of relativistically intense ($I> 10^{18} W/cm^{2}$) ultra-short (25 fs) laser pulses with sub-wavelength grating target. We measure the hot electron angular distribution and energy spectra for grating target and compare them with those from a planar mirror target. We observe that hot electrons are emitted in a collimated beam along the specular direction of the grating target. From the measured electron energy spectra we see electron temperature for grating is higher than the mirror, suggesting a higher electron yield and hence a stronger coupling with the laser. We performed numerical simulations which are in good agreement with experimental results, offer insights into the acceleration mechanism by resulting electric and magnetic fields. Such collimated fast electron beams have a wide range of applications in applied and fundamental science.

physics.plasm-ph↗
arXiv2023

Urban Visual Intelligence: Studying Cities with AI and Street-level Imagery

The visual dimension of cities has been a fundamental subject in urban studies, since the pioneering work of scholars such as Sitte, Lynch, Arnheim, and Jacobs. Several decades later, big data and artificial intelligence (AI) are revolutionizing how people move, sense, and interact with cities. This paper reviews the literature on the appearance and function of cities to illustrate how visual information has been used to understand them. A conceptual framework, Urban Visual Intelligence, is introduced to systematically elaborate on how new image data sources and AI techniques are reshaping the way researchers perceive and measure cities, enabling the study of the physical environment and its interactions with socioeconomic environments at various scales. The paper argues that these new approaches enable researchers to revisit the classic urban theories and themes, and potentially help cities create environments that are more in line with human behaviors and aspirations in the digital age.

cs.CV↗
arXiv2023

Explore More Guidance: A Task-aware Instruction Network for Sign Language Translation Enhanced with Data Augmentation

Sign language recognition and translation first uses a recognition module to generate glosses from sign language videos and then employs a translation module to translate glosses into spoken sentences. Most existing works focus on the recognition step, while paying less attention to sign language translation. In this work, we propose a task-aware instruction network, namely TIN-SLT, for sign language translation, by introducing the instruction module and the learning-based feature fuse strategy into a Transformer network. In this way, the pre-trained model's language ability can be well explored and utilized to further boost the translation performance. Moreover, by exploring the representation space of sign language glosses and target spoken language, we propose a multi-level data augmentation scheme to adjust the data distribution of the training set. We conduct extensive experiments on two challenging benchmark datasets, PHOENIX-2014-T and ASLG-PC12, on which our method outperforms former best solutions by 1.65 and 1.42 in terms of BLEU-4. Our code is published at https://github.com/yongcaoplus/TIN-SLT.

cs.CL↗
arXiv2023

Pay More Attention to Relation Exploration for Knowledge Base Question Answering

Knowledge base question answering (KBQA) is a challenging task that aims to retrieve correct answers from large-scale knowledge bases. Existing attempts primarily focus on entity representation and final answer reasoning, which results in limited supervision for this task. Moreover, the relations, which empirically determine the reasoning path selection, are not fully considered in recent advancements. In this study, we propose a novel framework, RE-KBQA, that utilizes relations in the knowledge base to enhance entity representation and introduce additional supervision. We explore guidance from relations in three aspects, including (1) distinguishing similar entities by employing a variational graph auto-encoder to learn relation importance; (2) exploring extra supervision by predicting relation distributions as soft labels with a multi-task scheme; (3) designing a relation-guided re-ranking algorithm for post-processing. Experimental results on two benchmark datasets demonstrate the effectiveness and superiority of our framework, improving the F1 score by 5.7% from 40.5 to 46.3 on CWQ and 5.8% from 62.8 to 68.5 on WebQSP, better or on par with state-of-the-art methods.

cs.CL↗
arXiv2023

FACE-AUDITOR: Data Auditing in Facial Recognition Systems

Few-shot-based facial recognition systems have gained increasing attention due to their scalability and ability to work with a few face images during the model deployment phase. However, the power of facial recognition systems enables entities with moderate resources to canvas the Internet and build well-performed facial recognition models without people's awareness and consent. To prevent the face images from being misused, one straightforward approach is to modify the raw face images before sharing them, which inevitably destroys the semantic information, increases the difficulty of retroactivity, and is still prone to adaptive attacks. Therefore, an auditing method that does not interfere with the facial recognition model's utility and cannot be quickly bypassed is urgently needed. In this paper, we formulate the auditing process as a user-level membership inference problem and propose a complete toolkit FACE-AUDITOR that can carefully choose the probing set to query the few-shot-based facial recognition model and determine whether any of a user's face images is used in training the model. We further propose to use the similarity scores between the original face images as reference information to improve the auditing performance. Extensive experiments on multiple real-world face image datasets show that FACE-AUDITOR can achieve auditing accuracy of up to $99\%$. Finally, we show that FACE-AUDITOR is robust in the presence of several perturbation mechanisms to the training images or the target models. The source code of our experiments can be found at \url{https://github.com/MinChen00/Face-Auditor}.

cs.CR↗
arXiv2023

Assessing Cross-Cultural Alignment between ChatGPT and Human Societies: An Empirical Study

The recent release of ChatGPT has garnered widespread recognition for its exceptional ability to generate human-like responses in dialogue. Given its usage by users from various nations and its training on a vast multilingual corpus that incorporates diverse cultural and societal norms, it is crucial to evaluate its effectiveness in cultural adaptation. In this paper, we investigate the underlying cultural background of ChatGPT by analyzing its responses to questions designed to quantify human cultural differences. Our findings suggest that, when prompted with American context, ChatGPT exhibits a strong alignment with American culture, but it adapts less effectively to other cultural contexts. Furthermore, by using different prompts to probe the model, we show that English prompts reduce the variance in model responses, flattening out cultural differences and biasing them towards American culture. This study provides valuable insights into the cultural implications of ChatGPT and highlights the necessity of greater diversity and cultural awareness in language technologies.

cs.CL↗
arXiv2023

Boundary Unlearning

The practical needs of the ``right to be forgotten'' and poisoned data removal call for efficient \textit{machine unlearning} techniques, which enable machine learning models to unlearn, or to forget a fraction of training data and its lineage. Recent studies on machine unlearning for deep neural networks (DNNs) attempt to destroy the influence of the forgetting data by scrubbing the model parameters. However, it is prohibitively expensive due to the large dimension of the parameter space. In this paper, we refocus our attention from the parameter space to the decision space of the DNN model, and propose Boundary Unlearning, a rapid yet effective way to unlearn an entire class from a trained DNN model. The key idea is to shift the decision boundary of the original DNN model to imitate the decision behavior of the model retrained from scratch. We develop two novel boundary shift methods, namely Boundary Shrink and Boundary Expanding, both of which can rapidly achieve the utility and privacy guarantees. We extensively evaluate Boundary Unlearning on CIFAR-10 and Vggface2 datasets, and the results show that Boundary Unlearning can effectively forget the forgetting class on image classification and face recognition tasks, with an expected speed-up of $17\times$ and $19\times$, respectively, compared with retraining from the scratch.

cs.CV↗
arXiv2023

Enhancing Vital Sign Estimation Performance of FMCW MIMO Radar by Prior Human Shape Recognition

Radio technology enabled contact-free human posture and vital sign estimation is promising for health monitoring. Radio systems at millimeter-wave (mmWave) frequencies advantageously bring large bandwidth, multi-antenna array and beam steering capability. \textit{However}, the human point cloud obtained by mmWave radar and utilized for posture estimation is likely to be sparse and incomplete. Additionally, human's random body movements deteriorate the estimation of breathing and heart rates, therefore the information of the chest location and a narrow radar beam toward the chest are demanded for more accurate vital sign estimation. In this paper, we propose a pipeline aiming to enhance the vital sign estimation performance of mmWave FMCW MIMO radar. The first step is to recognize human body part and posture, where we exploit a trained Convolutional Neural Networks (CNN) to efficiently process the imperfect human form point cloud. The CNN framework outputs the key point of different body parts, and was trained by using RGB image reference and Augmentative Ellipse Fitting Algorithm (AEFA). The next step is to utilize the chest information of the prior estimated human posture for vital sign estimation. While CNN is initially trained based on the frame-by-frame point clouds of human for posture estimation, the vital signs are extracted through beamforming toward the human chest. The numerical results show that this spatial filtering improves the estimation of the vital signs in regard to lowering the level of side harmonics and detecting the harmonics of vital signs efficiently, i.e., peak-to-average power ratio in the harmonics of vital signal is improved up to 0.02 and 0.07dB for the studied cases.

eess.SP↗
arXiv2023

Higher-harmonic generation in boron-doped silicon from band carriers and bound-dopant photoionization

We investigate ultrafast harmonic generation (HG) in Si:B, driven by intense pump pulses with fields reaching ~100 kV/cm and a carrier frequency of 300 GHz, at 4 K and 300 K, both experimentally and theoretically. We report several novel findings concerning the nonlinear charge carrier dynamics in intense sub-THz fields. (i) Harmonics of order up to n=9 are observed at room temperature, while at low temperature we can resolve harmonics reaching even n=13. The susceptibility per charge carrier at moderate field strength is as high as for charge carriers in graphene, considered to be one of the materials with the strongest sub-THz nonlinear response. (ii) For T=300 K, where the charge carriers bound to acceptors are fully thermally ionized into the valence subbands, the susceptibility values decrease with increasing field strength. Simulations incorporating multi-valence-band Monte-Carlo and finite-difference-time-domain (FDTD) propagation show that here, the HG process becomes increasingly dominated by energy-dependent scattering rates over the contribution from band non-parabolicity, due to the onset of optical-phonon emission, which ultimately leads to the saturation at high fields. (iii) At T=4 K, where the majority of charges are bound to acceptors, we observe a drastic rise of the HG yields for internal pump fields of 30 kV/cm, as one reaches the threshold for tunnel ionization. We disentangle the HG contributions in this case into contributions from the initial 'generational'- and subsequent band-nonlinearities, and show that scattering seriously degrades any coherent recollision during the subsequent oscillation of the holes.

physics.optics↗
arXiv2023

Non-homogeneous boundary value problems for coupled KdV-KdV systems posed on the half line

In this article, we study an initial-boundary-value problem of a coupled KdV-KdV system on the half line $ \mathbb{R}^+ $ with non-homogeneous boundary conditions: \begin{equation*} \left\{ \begin{array}{l} u_t+v_x+u u_x+v_{xxx}=0, \quad v_t+u_x+(vu)_x+u_{xxx}=0, \quad u(x,0)=ϕ(x),\quad v(x,0)=ψ(x), \quad u(0,t)=h_1(t),\quad v(0,t)=h_2(t),\quad v_x(0,t)=h_3(t), \end{array} \right. \qquad x,\,t>0. \end{equation*} It is shown that the problem is locally unconditionally well-posed in $H^s(\mathbb{R}^+)\times H^s(\mathbb{R}^+)$ for $s> -\frac34 $ with initial data $(ϕ,ψ)$ in $H^s(\mathbb{R}^+)\times H^{s}(\mathbb{R}^+)$ and boundary data $(h_1,h_2,h_3) $ in $H^{\frac{s+1}{3}}(\mathbb{R}^+)\times H^{\frac{s+1}{3}}(\mathbb{R}^+)\times H^{\frac{s}{3}}(\mathbb{R}^+)$. The approach developed in this paper can also be applied to study more general KdV-KdV systems posed on the half line.

math.AP↗
arXiv2022

High-order nonlinear terahertz probing of the two-band superconductor MgB$_2$: Third- and fifth-order harmonic generation

We report on high-order harmonic generation in the two-band superconductor MgB$_2$ driven by intense terahertz electromagnetic pulses. Third- and fifth-order harmonics are resolved in time domain and investigated as a function of temperature and in applied magnetic fields crossing the superconducting phase boundary. The high-order harmonics in the superconducting phase reflects nonequilibrium dynamics of the superconducting order parameter in MgB$_2$, which is probed via nonlinear coupling to the terahertz field. The observed temperature and field dependence of the nonlinear response allows to establish the superconducting phase diagram.

cond-mat.supr-con↗
arXiv2022

Dynamical interplay between superconductivity and charge-density-wave: a nonlinear terahertz study of coherently-driven 2H-NbSe2 and La2-xSrxCuO4

2H-NbSe2 is an archetypal system in which superconductivity and charge-density-wave (CDW) coexist and compete macroscopically with each other. In particular, this interplay also manifests in their dynamical fluctuations. As a result, the superconducting amplitude fluctuations (i.e. Higgs mode) is pushed below the quasiparticle continuum, allowing it to become a coherent excitation observable by Raman scattering. In the present study, we coherently drive the collective oscillations of the two orders and visualize their interplay in the driven states in the time domain. We find that both collective modes contribute to terahertz third harmonic generation (THG) and the THG signals interfere below Tc, leading to an anti-resonance of the integrated THG signal. The dynamical Ginzburg-Landau model suggests that around the anti-resonance a periodic energy transfer between the driven Higgs oscillations and the driven CDW oscillations is possible. In addition to 2H-NbSe2, we also studied an underdoped La2-xSrxCuO4 (x ~ 0.12) driven beyond the perturbative regime of THG. A similar interference between two sources of THG is observed below Tc. While there might be additional sources of THG in these experiments, our results illustrate the roles of coupled modes in the terahertz THG process and the tantalizing possibility of coherent control via such couplings.

cond-mat.supr-con↗
arXiv2022

CasFusionNet: A Cascaded Network for Point Cloud Semantic Scene Completion by Dense Feature Fusion

Semantic scene completion (SSC) aims to complete a partial 3D scene and predict its semantics simultaneously. Most existing works adopt the voxel representations, thus suffering from the growth of memory and computation cost as the voxel resolution increases. Though a few works attempt to solve SSC from the perspective of 3D point clouds, they have not fully exploited the correlation and complementarity between the two tasks of scene completion and semantic segmentation. In our work, we present CasFusionNet, a novel cascaded network for point cloud semantic scene completion by dense feature fusion. Specifically, we design (i) a global completion module (GCM) to produce an upsampled and completed but coarse point set, (ii) a semantic segmentation module (SSM) to predict the per-point semantic labels of the completed points generated by GCM, and (iii) a local refinement module (LRM) to further refine the coarse completed points and the associated labels from a local perspective. We organize the above three modules via dense feature fusion in each level, and cascade a total of four levels, where we also employ feature fusion between each level for sufficient information usage. Both quantitative and qualitative results on our compiled two point-based datasets validate the effectiveness and superiority of our CasFusionNet compared to state-of-the-art methods in terms of both scene completion and semantic segmentation. The codes and datasets are available at: https://github.com/JinfengX/CasFusionNet.

cs.CV↗