arXiv subjects

Xin Du

Publications and source records attributed to Xin Du.

At least 55 records · Page 3Linked to original sources

arXiv2024

Import competition and domestic vertical integration: Theory and Evidence from Chinese firms

What impact does import competition have on firms' production organizational choices? Existing literature has predominantly focused on the relationship between import competition and firms' global production networks, with less attention given to domestic. We first develop a Nash-bargaining model to guide our empirical analysis, then utilize tariff changes as an exogenous shock to test our theoretical hypotheses using a database of Chinese listed firms from 2000 to 2023. Our findings indicate that a decrease in downstream tariffs lead to an increase in vertical integration. In our mechanism tests, we discover that a reduction in upstream tariffs also enhances this effect. Moreover, the impact of tariff reductions on vertical integration is primarily observed in industries with high asset specificity, indicating that asset-specificity is a crucial mechanism. We further explore whether import competition encourages vertical integration for technological acquisition purpose, the effect is found only among high-tech firms, while it's absent in non-high-tech firms. Our research provides new perspectives and evidence on how firms optimize their production organization in the process of globalization.

econ.GN↗

arXiv2024

An Asynchronous Multi-core Accelerator for SNN inference

Spiking Neural Networks (SNNs) are extensively utilized in brain-inspired computing and neuroscience research. To enhance the speed and energy efficiency of SNNs, several many-core accelerators have been developed. However, maintaining the accuracy of SNNs often necessitates frequent explicit synchronization among all cores, which presents a challenge to overall efficiency. In this paper, we propose an asynchronous architecture for Spiking Neural Networks (SNNs) that eliminates the need for inter-core synchronization, thus enhancing speed and energy efficiency. This approach leverages the pre-determined dependencies of neuromorphic cores established during compilation. Each core is equipped with a scheduler that monitors the status of its dependencies, allowing it to safely advance to the next timestep without waiting for other cores. This eliminates the necessity for global synchronization and minimizes core waiting time despite inherent workload imbalances. Comprehensive evaluations using five different SNN workloads show that our architecture achieves a 1.86x speedup and a 1.55x increase in energy efficiency compared to state-of-the-art synchronization architectures.

cs.NE↗

arXiv2024

Bottleneck-Minimal Indexing for Generative Document Retrieval

We apply an information-theoretic perspective to reconsider generative document retrieval (GDR), in which a document $x \in X$ is indexed by $t \in T$, and a neural autoregressive model is trained to map queries $Q$ to $T$. GDR can be considered to involve information transmission from documents $X$ to queries $Q$, with the requirement to transmit more bits via the indexes $T$. By applying Shannon's rate-distortion theory, the optimality of indexing can be analyzed in terms of the mutual information, and the design of the indexes $T$ can then be regarded as a {\em bottleneck} in GDR. After reformulating GDR from this perspective, we empirically quantify the bottleneck underlying GDR. Finally, using the NQ320K and MARCO datasets, we evaluate our proposed bottleneck-minimal indexing method in comparison with various previous indexing methods, and we show that it outperforms those methods.

cs.IR↗

arXiv2024

Semi-supervised Anomaly Detection via Adaptive Reinforcement Learning-Enabled Method with Causal Inference for Sensor Signals

Semi-supervised anomaly detection for sensor signals is critical in ensuring system reliability in smart manufacturing. However, existing methods rely heavily on data correlation, neglecting causality and leading to potential misinterpretations due to confounding factors. Moreover, while current reinforcement learning-based methods can effectively identify known and unknown anomalies with limited labeled samples, these methods still face several challenges, such as under-utilization of priori knowledge, lack of model flexibility, and deficient reward feedback during environmental interactions. To address the above problems, this paper innovatively constructs a counterfactual causal reinforcement learning model, termed Triple-Assisted Causal Reinforcement Learning Anomaly Detector (Tri-CRLAD). The model leverages causal inference to extract the intrinsic causal feature in data, enhancing the agent's utilization of prior knowledge and improving its generalization capability. In addition, Tri-CRLAD features a triple decision support mechanism, including a sampling strategy based on historical similarity, an adaptive threshold smoothing adjustment strategy, and an adaptive decision reward mechanism. These mechanisms further enhance the flexibility and generalization ability of the model, enabling it to effectively respond to various complex and dynamically changing environments. Experimental results across seven diverse sensor signal datasets demonstrate that Tri-CRLAD outperforms nine state-of-the-art baseline methods. Notably, Tri-CRLAD achieves up to a 23\% improvement in anomaly detection stability with minimal known anomaly samples, highlighting its potential in semi-supervised anomaly detection scenarios. Our code is available at https://github.com/Aoudsung/Tri-CRLAD.

cs.LG↗

arXiv2024

Correlation Dimension of Natural Language in a Statistical Manifold

The correlation dimension of natural language is measured by applying the Grassberger-Procaccia algorithm to high-dimensional sequences produced by a large-scale language model. This method, previously studied only in a Euclidean space, is reformulated in a statistical manifold via the Fisher-Rao distance. Language exhibits a multifractal, with global self-similarity and a universal dimension around 6.5, which is smaller than those of simple discrete random sequences and larger than that of a Barabási-Albert process. Long memory is the key to producing self-similarity. Our method is applicable to any probabilistic model of real-world discrete sequences, and we show an application to music data.

cs.CL↗

arXiv2024

Rearrangements during slow compression of a jammed two-dimensional emulsion

As amorphous materials get jammed, both geometric and dynamic heterogeneity are observed. We investigate the correlation between the local geometric heterogeneity and local rearrangements in a slowly compressed bidisperse quasi-two-dimensional emulsion system. The compression is driven by evaporation of the continuous phase, and causes the area packing fraction to increase from 0.88 to 0.99. We quantify the structural heterogeneity of the system using the radical Voronoi tessellation following the method of [Rieser et al., Phys. Rev. Lett. 116, 088001 (2016)]. We define two structural quantities characterizing local structure, the first which considers nearest neighbors and the second of which includes information from second nearest neighbors. We find that droplets in heterogeneous local regions are more likely to have local rearrangements. These rearrangements are generally T1 events where two droplets converge toward a void, and two droplets move away from the void to make room for the converging droplets. Thus the presence of the voids tends to orient the T1 events. The presence of a correlation between the structural quantities and the rearrangement dynamics remains qualitatively unchanged over the entire range of packing fractions observed.

cond-mat.soft↗

arXiv2023

Co-Training Realized Volatility Prediction Model with Neural Distributional Transformation

This paper shows a novel machine learning model for realized volatility (RV) prediction using a normalizing flow, an invertible neural network. Since RV is known to be skewed and have a fat tail, previous methods transform RV into values that follow a latent distribution with an explicit shape and then apply a prediction model. However, knowing that shape is non-trivial, and the transformation result influences the prediction model. This paper proposes to jointly train the transformation and the prediction model. The training process follows a maximum-likelihood objective function that is derived from the assumption that the prediction residuals on the transformed RV time series are homogeneously Gaussian. The objective function is further approximated using an expectation-maximum algorithm. On a dataset of 100 stocks, our method significantly outperforms other methods using analytical or naive neural-network transformations.

cs.CE↗

arXiv2023

DegUIL: Degree-aware Graph Neural Networks for Long-tailed User Identity Linkage

User identity linkage (UIL), matching accounts of a person on different social networks, is a fundamental task in cross-network data mining. Recent works have achieved promising results by exploiting graph neural networks (GNNs) to capture network structure. However, they rarely analyze the realistic node-level bottlenecks that hinder UIL's performance. First, node degrees in a graph vary widely and are long-tailed. A significant fraction of tail nodes with small degrees are underrepresented due to limited structural information, degrading linkage performance seriously. The second bottleneck usually overlooked is super head nodes. It is commonly accepted that head nodes perform well. However, we find that some of them with super high degrees also have difficulty aligning counterparts, due to noise introduced by the randomness of following friends in real-world social graphs. In pursuit of learning ideal representations for these two groups of nodes, this paper proposes a degree-aware model named DegUIL to narrow the degree gap. To this end, our model complements missing neighborhoods for tail nodes and discards redundant structural information for super head nodes in embeddings respectively. Specifically, the neighboring bias is predicted and corrected locally by two modules, which are trained using the knowledge from structurally adequate head nodes. As a result, ideal neighborhoods are obtained for meaningful aggregation in GNNs. Extensive experiments demonstrate the superiority of our model. Our data and code can be found at https://github.com/Longmeix/DegUIL.

cs.SI↗

arXiv2023

Digital Twin Brain: a simulation and assimilation platform for whole human brain

In this work, we present a computing platform named digital twin brain (DTB) that can simulate spiking neuronal networks of the whole human brain scale and more importantly, a personalized biological brain structure. In comparison to most brain simulations with a homogeneous global structure, we highlight that the sparseness, couplingness and heterogeneity in the sMRI, DTI and PET data of the brain has an essential impact on the efficiency of brain simulation, which is proved from the scaling experiments that the DTB of human brain simulation is communication-intensive and memory-access intensive computing systems rather than computation-intensive. We utilize a number of optimization techniques to balance and integrate the computation loads and communication traffics from the heterogeneous biological structure to the general GPU-based HPC and achieve leading simulation performance for the whole human brain-scaled spiking neuronal networks. On the other hand, the biological structure, equipped with a mesoscopic data assimilation, enables the DTB to investigate brain cognitive function by a reverse-engineering method, which is demonstrated by a digital experiment of visual evaluation on the DTB. Furthermore, we believe that the developing DTB will be a promising powerful platform for a large of research orients including brain-inspiredintelligence, rain disease medicine and brain-machine interface.

cs.NE↗

arXiv2023

Frequency-limited H$_2$ Model Order Reduction Based on Relative Error

Frequency-limited model order reduction aims to approximate a high-order model with a reduced-order model that maintains high fidelity within a specific frequency range. Beyond this range, a decrease in accuracy is acceptable due to the nature of the problem. The quality of the reduced-order model is typically evaluated using absolute or relative measures of approximation error. Relative error, which represents the percentage error, becomes particularly relevant when reducing a plant model for the purpose of designing a reduced-order controller. This paper derives the necessary conditions for achieving a local optimum of the frequency-limited H2 norm for the relative error system. Based on these optimality conditions, an oblique projection algorithm is proposed to ensure a small relative error within the desired frequency interval. Unlike existing algorithms, the proposed approach does not necessitate solving large-scale Lyapunov and Ricatti equations. Instead, the proposed algorithm relies on solving sparse-dense Sylvester equations, which typically emerge in the majority of H2 model order reduction algorithms, but can be efficiently solved. To evaluate the performance of the proposed algorithm, a comparison is conducted with three existing techniques: frequency-limited balanced truncation, frequency-limited balanced stochastic truncation, and frequency-limited iterative Rational Krylov algorithm. The comparative analysis focuses on designing reduced-order controllers for high-order plants. Numerical results confirm that the reduced-order controllers obtained using the proposed algorithm ensure superior robust closed-loop stability.

eess.SY↗

arXiv2023

Exploring charge and spin fluctuations in infinite-layer cuprate SrCuO$_{2}$ from a phonon perspective

The infinite-layer cuprate $A$CuO$_2$ ($A=$ Ca, Sr, Ba) has the simplest crystal structure among numerous cuprate superconductors and can serve a prototypical system to explore the unconventional superconductivity. Based on the first-principles electronic structure calculations, we have studied the electronic and magnetic properties of the infinite-layer cuprate SrCuO$_{2}$ from a phonon perspective. We find that interesting fluctuations of charges, electrical dipoles, and local magnetic moments can be induced by the zero-point vibrations of phonon modes in SrCuO$_{2}$ upon the hole doping. Among all optical phonon modes of SrCuO$_{2}$ in the antiferromagnetic Néel state, only the $A_{1}$$_g$ mode that involves the full-breathing O vibrations along the Cu-O bonds can cause significant fluctuations of local magnetic moments on O atoms and dramatic charge redistributions between Cu and O atoms. Notably, due to the zero-point vibration of the $A_{1g}$ mode, both the charge fluctuations on Cu and the electrical dipoles on O show a dome-like evolution with increasing hole doping, quite similar to the experimentally observed behavior of the superconducting $T_c$; in comparison, the fluctuations of local magnetic moments on O display a monotonic enhancement along with the hole doping. Further analyses indicate that around the optimal doping, there exist a large softening in the frequency of the $A_{1g}$ phonon mode and a van Hove singularity in the electronic structure close to the Fermi level, suggesting potential electron-phonon coupling. Our work reveals the important role of the full-breathing O phonon mode playing in the infinite-layer SrCuO$_{2}$, which may provide new insights in understanding the cuprate superconductivity.

cond-mat.supr-con↗

arXiv2023

Frequency-dependent Switching Control for Disturbance Attenuation of Linear Systems

The generalized Kalman-Yakubovich-Popov lemma as established by Iwasaki and Hara in 2005 marks a milestone in the analysis and synthesis of linear systems from a finite-frequency perspective. Given a pre-specified frequency band, it allows us to produce passive controllers with excellent in-band disturbance attenuation performance at the expense of some of the out-of-band performance. This paper focuses on control design of linear systems in the presence of disturbances with non-strictly or non-stationary limited frequency spectrum. We first propose a class of frequency-dependent excited energy functions (FD-EEF) as well as frequency-dependent excited power functions (FD-EPF), which possess a desirable frequency-selectiveness property with regard to the in-band and out-of-band excited energy as well as excited power of the system. Based upon a group of frequency-selective passive controllers, we then develop a frequency-dependent switching control (FDSC) scheme that selects the most appropriate controller at runtime. We show that our FDSC scheme is capable to approximate the solid in-band performance while maintaining acceptable out-of-band performance with regard to global time horizons as well as localized time horizons. The method is illustrated by a commonly used benchmark model.

eess.SY↗

arXiv2023

Physical-layer Adversarial Robustness for Deep Learning-based Semantic Communications

End-to-end semantic communications (ESC) rely on deep neural networks (DNN) to boost communication efficiency by only transmitting the semantics of data, showing great potential for high-demand mobile applications. We argue that central to the success of ESC is the robust interpretation of conveyed semantics at the receiver side, especially for security-critical applications such as automatic driving and smart healthcare. However, robustifying semantic interpretation is challenging as ESC is extremely vulnerable to physical-layer adversarial attacks due to the openness of wireless channels and the fragileness of neural models. Toward ESC robustness in practice, we ask the following two questions: Q1: For attacks, is it possible to generate semantic-oriented physical-layer adversarial attacks that are imperceptible, input-agnostic and controllable? Q2: Can we develop a defense strategy against such semantic distortions and previously proposed adversaries? To this end, we first present MobileSC, a novel semantic communication framework that considers the computation and memory efficiency in wireless environments. Equipped with this framework, we propose SemAdv, a physical-layer adversarial perturbation generator that aims to craft semantic adversaries over the air with the abovementioned criteria, thus answering the Q1. To better characterize the realworld effects for robust training and evaluation, we further introduce a novel adversarial training method SemMixed to harden the ESC against SemAdv attacks and existing strong threats, thus answering the Q2. Extensive experiments on three public benchmarks verify the effectiveness of our proposed methods against various physical adversarial attacks. We also show some interesting findings, e.g., our MobileSC can even be more robust than classical block-wise communication systems in the low SNR regime.

eess.SP↗

arXiv2023

Universal Adversarial Backdoor Attacks to Fool Vertical Federated Learning in Cloud-Edge Collaboration

Vertical federated learning (VFL) is a cloud-edge collaboration paradigm that enables edge nodes, comprising resource-constrained Internet of Things (IoT) devices, to cooperatively train artificial intelligence (AI) models while retaining their data locally. This paradigm facilitates improved privacy and security for edges and IoT devices, making VFL an essential component of Artificial Intelligence of Things (AIoT) systems. Nevertheless, the partitioned structure of VFL can be exploited by adversaries to inject a backdoor, enabling them to manipulate the VFL predictions. In this paper, we aim to investigate the vulnerability of VFL in the context of binary classification tasks. To this end, we define a threat model for backdoor attacks in VFL and introduce a universal adversarial backdoor (UAB) attack to poison the predictions of VFL. The UAB attack, consisting of universal trigger generation and clean-label backdoor injection, is incorporated during the VFL training at specific iterations. This is achieved by alternately optimizing the universal trigger and model parameters of VFL sub-problems. Our work distinguishes itself from existing studies on designing backdoor attacks for VFL, as those require the knowledge of auxiliary information not accessible within the split VFL architecture. In contrast, our approach does not necessitate any additional data to execute the attack. On the LendingClub and Zhongyuan datasets, our approach surpasses existing state-of-the-art methods, achieving up to 100\% backdoor task performance while maintaining the main task performance. Our results in this paper make a major advance to revealing the hidden backdoor risks of VFL, hence paving the way for the future development of secure AIoT.

cs.LG↗

arXiv2023

Exploiting Input Sanitization for Regex Denial of Service

Web services use server-side input sanitization to guard against harmful input. Some web services publish their sanitization logic to make their client interface more usable, e.g., allowing clients to debug invalid requests locally. However, this usability practice poses a security risk. Specifically, services may share the regexes they use to sanitize input strings -- and regex-based denial of service (ReDoS) is an emerging threat. Although prominent service outages caused by ReDoS have spurred interest in this topic, we know little about the degree to which live web services are vulnerable to ReDoS. In this paper, we conduct the first black-box study measuring the extent of ReDoS vulnerabilities in live web services. We apply the Consistent Sanitization Assumption: that client-side sanitization logic, including regexes, is consistent with the sanitization logic on the server-side. We identify a service's regex-based input sanitization in its HTML forms or its API, find vulnerable regexes among these regexes, craft ReDoS probes, and pinpoint vulnerabilities. We analyzed the HTML forms of 1,000 services and the APIs of 475 services. Of these, 355 services publish regexes; 17 services publish unsafe regexes; and 6 services are vulnerable to ReDoS through their APIs (6 domains; 15 subdomains). Both Microsoft and Amazon Web Services patched their web services as a result of our disclosure. Since these vulnerabilities were from API specifications, not HTML forms, we proposed a ReDoS defense for a popular API validation library, and our patch has been merged. To summarize: in client-visible sanitization logic, some web services advertise ReDoS vulnerabilities in plain sight. Our results motivate short-term patches and long-term fundamental solutions.

cs.CR↗

arXiv2022

Relative Error-based Time-limited H2 Model Order Reduction via Oblique Projection

In time-limited model order reduction, a reduced-order approximation of the original high-order model is obtained that accurately approximates the original model within the desired limited time interval. Accuracy outside that time interval is not that important. The error incurred when a reduced-order model is used as a surrogate for the original model can be quantified in absolute or relative terms to access the performance of the model reduction algorithm. The relative error is generally more meaningful than an absolute error because if the original and reduced systems' responses are of small magnitude, the absolute error is small in magnitude as well. However, this does not necessarily mean that the reduced model is accurate. The relative error in such scenarios is useful and meaningful as it quantifies percentage error irrespective of the magnitude of the system's response. In this paper, the necessary conditions for a local optimum of the time-limited H2 norm of the relative error system are derived. Inspired by these conditions, an oblique projection algorithm is proposed that ensures small H2-norm relative error within the desired time interval. Unlike the existing relative error-based model reduction algorithms, the proposed algorithm does not require solutions of large-scale Lyapunov and Riccati equations. The proposed algorithm is compared with time-limited balanced truncation, time-limited balanced stochastic truncation, and time-limited iterative Rational Krylov algorithm. Numerical results confirm the superiority of the proposed algorithm over these existing algorithms.

eess.SY↗

arXiv2022

ClassPruning: Speed Up Image Restoration Networks by Dynamic N:M Pruning

Image restoration tasks have achieved tremendous performance improvements with the rapid advancement of deep neural networks. However, most prevalent deep learning models perform inference statically, ignoring that different images have varying restoration difficulties and lightly degraded images can be well restored by slimmer subnetworks. To this end, we propose a new solution pipeline dubbed ClassPruning that utilizes networks with different capabilities to process images with varying restoration difficulties. In particular, we use a lightweight classifier to identify the image restoration difficulty, and then the sparse subnetworks with different capabilities can be sampled based on predicted difficulty by performing dynamic N:M fine-grained structured pruning on base restoration networks. We further propose a novel training strategy along with two additional loss terms to stabilize training and improve performance. Experiments demonstrate that ClassPruning can help existing methods save approximately 40% FLOPs while maintaining performance.

cs.CV↗

arXiv2022

Risk-Driven Design of Perception Systems

Modern autonomous systems rely on perception modules to process complex sensor measurements into state estimates. These estimates are then passed to a controller, which uses them to make safety-critical decisions. It is therefore important that we design perception systems to minimize errors that reduce the overall safety of the system. We develop a risk-driven approach to designing perception systems that accounts for the effect of perceptual errors on the performance of the fully-integrated, closed-loop system. We formulate a risk function to quantify the effect of a given perceptual error on overall safety, and show how we can use it to design safer perception systems by including a risk-dependent term in the loss function and generating training data in risk-sensitive regions. We evaluate our techniques on a realistic vision-based aircraft detect and avoid application and show that risk-driven design reduces collision risk by 37% over a baseline system.

cs.RO↗