Search arXiv⌕ Search

arXiv subjects

Jun Zhang

Publications and source records attributed to Jun Zhang.

At least 595 records · Page 33Linked to original sources
arXiv2024

VulMatch: Binary-level Vulnerability Detection Through Signature

Similar vulnerability repeats in real-world software products because of code reuse, especially in wildly reused third-party code and libraries. Detecting repeating vulnerabilities like 1-day and N-day vulnerabilities is an important cyber security task. Unfortunately, the state-of-the-art methods suffer from poor performance because they detect patch existence instead of vulnerability existence and infer the vulnerability signature directly from binary code. In this paper, we propose VulMatch to extract precise vulnerability-related binary instructions to generate the vulnerability-related signature. VulMatch detects vulnerability existence based on binary signatures. Unlike previous approaches, VulMatch accurately locates vulnerability-related instructions by utilizing source and binary codes. Our experiments were conducted using over 1000 vulnerable instances across seven open-source projects. VulMatch significantly outperformed the baseline tools Asm2vec and Palmtree. Besides the performance advantages over the baseline tools, VulMatch offers a better feature by providing explainable reasons during vulnerability detection. Our empirical studies demonstrate that VulMatch detects fine-grained vulnerability that the state-of-the-art tools struggle with. Our experiment on commercial firmware demonstrates VulMatch is able to find vulnerabilities in real-world scenario.

cs.CR↗
arXiv2023

Molecular CT: Unifying Geometry and Representation Learning for Molecules at Different Scales

Deep learning is changing many areas in molecular physics, and it has shown great potential to deliver new solutions to challenging molecular modeling problems. Along with this trend arises the increasing demand of expressive and versatile neural network architectures which are compatible with molecular systems. A new deep neural network architecture, Molecular Configuration Transformer (Molecular CT), is introduced for this purpose. Molecular CT is composed of a relation-aware encoder module and a computationally universal geometry learning unit, thus able to account for the relational constraints between particles meanwhile scalable to different particle numbers and invariant with respect to the trans-rotational transforms. The computational efficiency and universality make Molecular CT versatile for a variety of molecular learning scenarios and especially appealing for transferable representation learning across different molecular systems. As examples, we show that Molecular CT enables representational learning for molecular systems at different scales, and achieves comparable or improved results on common benchmarks using a more light-weighted structure compared to baseline models.

cs.LG↗
arXiv2023

Large Language Models Empowered Autonomous Edge AI for Connected Intelligence

The evolution of wireless networks gravitates towards connected intelligence, a concept that envisions seamless interconnectivity among humans, objects, and intelligence in a hyper-connected cyber-physical world. Edge artificial intelligence (Edge AI) is a promising solution to achieve connected intelligence by delivering high-quality, low-latency, and privacy-preserving AI services at the network edge. This article presents a vision of autonomous edge AI systems that automatically organize, adapt, and optimize themselves to meet users' diverse requirements, leveraging the power of large language models (LLMs), i.e., Generative Pretrained Transformer (GPT). By exploiting the powerful abilities of GPT in language understanding, planning, and code generation, as well as incorporating classic wisdom such as task-oriented communication and edge federated learning, we present a versatile framework that efficiently coordinates edge AI models to cater to users' personal demands while automatically generating code to train new models in a privacy-preserving manner. Experimental results demonstrate the system's remarkable ability to accurately comprehend user demands, efficiently execute AI models with minimal cost, and effectively create high-performance AI models at edge servers.

cs.IT↗
arXiv2023

Entanglement islands read perfect-tensor entanglement

In this paper, we make use of holographic Boundary Conformal Field Theory (BCFT) to simulate the black hole information problem in the semi-classical picture. We investigate the correlation between a portion of Hawking radiation and entanglement islands by the area of an entanglement wedge cross-section. Building on the understanding of the relationship between entanglement wedge cross-sections and perfect tensor entanglement as discussed in reference [1], we make an intriguing observation: in the semi-classical picture, the positioning of an entanglement island automatically yields a pattern of perfect tensor entanglement. Furthermore, the contribution of this perfect tensor entanglement, combined with the bipartite entanglement contribution, precisely determines the area of the entanglement wedge cross-section.

hep-th↗
arXiv2023

Holographic coarse-grained states and the necessity of perfect entanglement

In the framework of the holographic principle, focusing on a central concept, conditional mutual information, we construct a class of coarse-grained states, which are intuitively connected to a family of thread configurations. These coarse-grained states characterize the entanglement structure of holographic systems at a coarse-grained level. Importantly, these coarse-grained states can be used to further reveal nontrivial requirements for the holographic entanglement structure. Specifically, we employ these coarse-grained states to probe the entanglement entropies of disconnected regions and the entanglement wedge cross-section dual to the inherent correlation in a bipartite mixed state. The investigations demonstrate the necessity of perfect tensor state entanglement. Moreover, in a certain sense, our work establishes the equivalence between the holographic entanglement of purification and the holographic balanced partial entropy. We also construct a thread configuration with the Multi-Scale Entanglement Renormalization Ansatz (MERA) structure, reexamining the connection between the MERA structure and kinematic space.

hep-th↗
arXiv2023

A Fully-automatic Side-scan Sonar SLAM Framework

Side-scan sonar (SSS) is a lightweight acoustic sensor that is frequently deployed on autonomous underwater vehicles (AUVs) to provide high-resolution seafloor images. However, using side-scan images to perform simultaneous localization and mapping (SLAM) remains a challenge when there is a lack of 3D bathymetric information and discriminant features in the side-scan images. To tackle this, we propose a feature-based SLAM framework using side-scan sonar, which is able to automatically detect and robustly match keypoints between paired side-scan images. We then use the detected correspondences as constraints to optimize the AUV pose trajectory. The proposed method is evaluated on real data collected by a Hugin AUV, using as a ground truth reference both manually-annotated keypoints and a 3D bathymetry mesh from multibeam echosounder (MBES). Experimental results demonstrate that our approach is able to reduce drifts from the dead-reckoning system. The framework is made publicly available for the benefit of the community.

cs.RO↗
arXiv2023

Urban Generative Intelligence (UGI): A Foundational Platform for Agents in Embodied City Environment

Urban environments, characterized by their complex, multi-layered networks encompassing physical, social, economic, and environmental dimensions, face significant challenges in the face of rapid urbanization. These challenges, ranging from traffic congestion and pollution to social inequality, call for advanced technological interventions. Recent developments in big data, artificial intelligence, urban computing, and digital twins have laid the groundwork for sophisticated city modeling and simulation. However, a gap persists between these technological capabilities and their practical implementation in addressing urban challenges in an systemic-intelligent way. This paper proposes Urban Generative Intelligence (UGI), a novel foundational platform integrating Large Language Models (LLMs) into urban systems to foster a new paradigm of urban intelligence. UGI leverages CityGPT, a foundation model trained on city-specific multi-source data, to create embodied agents for various urban tasks. These agents, operating within a textual urban environment emulated by city simulator and urban knowledge graph, interact through a natural language interface, offering an open platform for diverse intelligent and embodied agent development. This platform not only addresses specific urban issues but also simulates complex urban systems, providing a multidisciplinary approach to understand and manage urban complexity. This work signifies a transformative step in city science and urban intelligence, harnessing the power of LLMs to unravel and address the intricate dynamics of urban systems. The code repository with demonstrations will soon be released here https://github.com/tsinghua-fib-lab/UGI.

cs.AI↗
arXiv2023

Linear Attention via Orthogonal Memory

Efficient attentions have greatly improved the computational efficiency of Transformers. However, most existing linear attention mechanisms suffer from an \emph{efficiency degradation} problem, leading to inefficiencies in causal language modeling and hindering their application in long-range language models. This problem is more pronounced under language modeling with unbounded contexts. In this paper, we propose \textbf{L}inear \textbf{A}ttention \textbf{V}ia \textbf{O}rthogonal memory~(\shortname) to address these limitations, achieving strong performance while maintaining linear complexity. \shortname employs orthogonal decomposition to compress a context into a fixed-size orthogonal memory while effectively minimizing redundancy within the context. Given that orthogonal memory compresses global information, we further dissect the context to amplify fine-grained local information. Additionally, we embed the relative position encoding into \shortname to improve the extrapolation ability. Experimental results show that \shortname greatly improves the efficiency of the causal language model with the best extrapolation performance and outperforms other efficient baselines. Further, we endeavor to employ \shortname for unbounded language modeling and successfully scale the context length to 128K.

cs.CL↗
arXiv2023

Selective Knowledge Sharing for Privacy-Preserving Federated Distillation without A Good Teacher

While federated learning is promising for privacy-preserving collaborative learning without revealing local data, it remains vulnerable to white-box attacks and struggles to adapt to heterogeneous clients. Federated distillation (FD), built upon knowledge distillation--an effective technique for transferring knowledge from a teacher model to student models--emerges as an alternative paradigm, which provides enhanced privacy guarantees and addresses model heterogeneity. Nevertheless, challenges arise due to variations in local data distributions and the absence of a well-trained teacher model, which leads to misleading and ambiguous knowledge sharing that significantly degrades model performance. To address these issues, this paper proposes a selective knowledge sharing mechanism for FD, termed Selective-FD. It includes client-side selectors and a server-side selector to accurately and precisely identify knowledge from local and ensemble predictions, respectively. Empirical studies, backed by theoretical insights, demonstrate that our approach enhances the generalization capabilities of the FD framework and consistently outperforms baseline methods.

cs.LG↗
arXiv2023

Image Super-resolution Via Latent Diffusion: A Sampling-space Mixture Of Experts And Frequency-augmented Decoder Approach

The recent use of diffusion prior, enhanced by pre-trained text-image models, has markedly elevated the performance of image super-resolution (SR). To alleviate the huge computational cost required by pixel-based diffusion SR, latent-based methods utilize a feature encoder to transform the image and then implement the SR image generation in a compact latent space. Nevertheless, there are two major issues that limit the performance of latent-based diffusion. First, the compression of latent space usually causes reconstruction distortion. Second, huge computational cost constrains the parameter scale of the diffusion model. To counteract these issues, we first propose a frequency compensation module that enhances the frequency components from latent space to pixel space. The reconstruction distortion (especially for high-frequency information) can be significantly decreased. Then, we propose to use Sample-Space Mixture of Experts (SS-MoE) to achieve more powerful latent-based SR, which steadily improves the capacity of the model without a significant increase in inference costs. These carefully crafted designs contribute to performance improvements in largely explored 4x blind super-resolution benchmarks and extend to large magnification factors, i.e., 8x image SR benchmarks. The code is available at https://github.com/amandaluof/moe_sr.

cs.CV↗
arXiv2023

Sluggish and Chemically-Biased Interstitial Diffusion in Concentrated Solid Solution Alloys: Mechanisms and Methods

Interstitial diffusion is a pivotal process that governs the phase stability and irradiation response of materials in non-equilibrium conditions. In this work, we study sluggish and chemically-biased interstitial diffusion in Fe-Ni concentrated solid solution alloys (CSAs) by combining machine learning (ML) and kinetic Monte Carlo (kMC), where ML is used to accurately and efficiently predict the migration energy barriers on-the-fly. The ML-kMC reproduces the diffusivity that was reported by molecular dynamics results at high temperatures. With this powerful tool, we find that the observed sluggish diffusion and the "Ni-Ni-Ni"-biased diffusion in Fe-Ni alloys are ascribed to a unique "Barrier Lock" mechanism, whereas the "Fe-Fe-Fe"-biased diffusion is influenced by a "Component Dominance" mechanism. Inspired by the mentioned mechanisms, a practical AvgS-kMC method is proposed for conveniently and swiftly determining interstitial-mediated diffusivity by only relying on the mean energy barriers of migration patterns. Combining the AvgS-kMC with the differential evolutionary algorithm, an inverse design strategy for optimizing sluggish diffusion properties is applied to emphasize the crucial role of favorable migration patterns.

cond-mat.mtrl-sci↗
arXiv2023

Channel and Gradient-Importance Aware Device Scheduling for Over-the-Air Federated Learning

Federated learning (FL) is a popular privacy-preserving distributed training scheme, where multiple devices collaborate to train machine learning models by uploading local model updates. To improve communication efficiency, over-the-air computation (AirComp) has been applied to FL, which leverages analog modulation to harness the superposition property of radio waves such that numerous devices can upload their model updates concurrently for aggregation. However, the uplink channel noise incurs considerable model aggregation distortion, which is critically determined by the device scheduling and compromises the learned model performance. In this paper, we propose a probabilistic device scheduling framework for over-the-air FL, named PO-FL, to mitigate the negative impact of channel noise, where each device is scheduled according to a certain probability and its model update is reweighted using this probability in aggregation. We prove the unbiasedness of this aggregation scheme and demonstrate the convergence of PO-FL on both convex and non-convex loss functions. Our convergence bounds unveil that the device scheduling affects the learning performance through the communication distortion and global update variance. Based on the convergence analysis, we further develop a channel and gradient-importance aware algorithm to optimize the device scheduling probabilities in PO-FL. Extensive simulation results show that the proposed PO-FL framework with channel and gradient-importance awareness achieves faster convergence and produces better models than baseline methods.

cs.LG↗
arXiv2023

Surgical Temporal Action-aware Network with Sequence Regularization for Phase Recognition

To assist surgeons in the operating theatre, surgical phase recognition is critical for developing computer-assisted surgical systems, which requires comprehensive understanding of surgical videos. Although existing studies made great progress, there are still two significant limitations worthy of improvement. First, due to the compromise of resource consumption, frame-wise visual features are extracted by 2D networks and disregard spatial and temporal knowledge of surgical actions, which hinders subsequent inter-frame modeling for phase prediction. Second, these works simply utilize ordinary classification loss with one-hot phase labels to optimize the phase predictions, and cannot fully explore surgical videos under inadequate supervision. To overcome these two limitations, we propose a Surgical Temporal Action-aware Network with sequence Regularization, named STAR-Net, to recognize surgical phases more accurately from input videos. Specifically, we propose an efficient multi-scale surgical temporal action (MS-STA) module, which integrates visual features with spatial and temporal knowledge of surgical actions at the cost of 2D networks. Moreover, we devise the dual-classifier sequence regularization (DSR) to facilitate the training of STAR-Net by the sequence guidance of an auxiliary classifier with a smaller capacity. Our STAR-Net with MS-STA and DSR can exploit visual features of surgical actions with effective regularization, thereby leading to the superior performance of surgical phase recognition. Extensive experiments on a large-scale gastrectomy surgery dataset and the public Cholec80 benchmark prove that our STAR-Net significantly outperforms state-of-the-arts of surgical phase recognition.

cs.CV↗
arXiv2023

KBioXLM: A Knowledge-anchored Biomedical Multilingual Pretrained Language Model

Most biomedical pretrained language models are monolingual and cannot handle the growing cross-lingual requirements. The scarcity of non-English domain corpora, not to mention parallel data, poses a significant hurdle in training multilingual biomedical models. Since knowledge forms the core of domain-specific corpora and can be translated into various languages accurately, we propose a model called KBioXLM, which transforms the multilingual pretrained model XLM-R into the biomedical domain using a knowledge-anchored approach. We achieve a biomedical multilingual corpus by incorporating three granularity knowledge alignments (entity, fact, and passage levels) into monolingual corpora. Then we design three corresponding training tasks (entity masking, relation masking, and passage relation prediction) and continue training on top of the XLM-R model to enhance its domain cross-lingual ability. To validate the effectiveness of our model, we translate the English benchmarks of multiple tasks into Chinese. Experimental results demonstrate that our model significantly outperforms monolingual and multilingual pretrained models in cross-lingual zero-shot and few-shot scenarios, achieving improvements of up to 10+ points. Our code is publicly available at https://github.com/ngwlh-gl/KBioXLM.

cs.CL↗
arXiv2023

DESI Legacy Imaging Surveys Data Release 9: Cosmological Constraints from Galaxy Clustering and Weak Lensing using the Minimal Bias Model

We present a tentative constraint on cosmological parameters $Ω_m$ and $σ_8$ from a joint analysis of galaxy clustering and galaxy-galaxy lensing from DESI Legacy Imaging Surveys Data Release 9 (DR9), covering approximately 10000 square degrees and spanning the redshift range of 0.1 to 0.9. To study the dependence of cosmological parameters on lens redshift, we divide lens galaxies into seven approximately volume-limited samples, each with an equal width in photometric redshift. To retrieve the intrinsic projected correlation function $w_{\rm p}(r_{\rm p})$ from the lens samples, we employ a novel method to account for redshift uncertainties. Additionally, we measured the galaxy-galaxy lensing signal $ΔΣ(r_{\rm p})$ for each lens sample, using source galaxies selected from the shear catalog by applying our \texttt{Fourier\_Quad} pipeline to DR9 images. We model these observables within the flat $Λ$CDM framework, employing the minimal bias model. To ensure the reliability of the minimal bias model, we apply conservative scale cuts: $r_{\rm p} > 8$ and $12 ~h^{-1}{\rm Mpc}$, for $w_{\rm p}(r_{\rm p})$ and $ΔΣ(r_{\rm p})$, respectively. Our findings suggest a mild tendency that $S_8 \equiv σ_8 \sqrt{Ω_m/0.3} $ increases with lens redshift, although this trend is only marginally significant. When we combine low redshift samples, the value of $S_8$ is determined to be $0.84 \pm 0.02$, consistent with the Planck results but significantly higher than the 3$\times$ 2pt analysis by 2-5$σ$. Despite the fact that further refinements in measurements and modeling could improve the accuracy of our results, the consistency with standard values demonstrates the potential of our method for more precise and accurate cosmology in the future.

astro-ph.CO↗
arXiv2023

Stochastic Coded Federated Learning: Theoretical Analysis and Incentive Mechanism Design

Federated learning (FL) has achieved great success as a privacy-preserving distributed training paradigm, where many edge devices collaboratively train a machine learning model by sharing the model updates instead of the raw data with a server. However, the heterogeneous computational and communication resources of edge devices give rise to stragglers that significantly decelerate the training process. To mitigate this issue, we propose a novel FL framework named stochastic coded federated learning (SCFL) that leverages coded computing techniques. In SCFL, before the training process starts, each edge device uploads a privacy-preserving coded dataset to the server, which is generated by adding Gaussian noise to the projected local dataset. During training, the server computes gradients on the global coded dataset to compensate for the missing model updates of the straggling devices. We design a gradient aggregation scheme to ensure that the aggregated model update is an unbiased estimate of the desired global update. Moreover, this aggregation scheme enables periodical model averaging to improve the training efficiency. We characterize the tradeoff between the convergence performance and privacy guarantee of SCFL. In particular, a more noisy coded dataset provides stronger privacy protection for edge devices but results in learning performance degradation. We further develop a contract-based incentive mechanism to coordinate such a conflict. The simulation results show that SCFL learns a better model within the given time and achieves a better privacy-performance tradeoff than the baseline methods. In addition, the proposed incentive mechanism grants better training performance than the conventional Stackelberg game approach.

cs.DC↗
arXiv2023

Improving Large-scale Deep Biasing with Phoneme Features and Text-only Data in Streaming Transducer

Deep biasing for the Transducer can improve the recognition performance of rare words or contextual entities, which is essential in practical applications, especially for streaming Automatic Speech Recognition (ASR). However, deep biasing with large-scale rare words remains challenging, as the performance drops significantly when more distractors exist and there are words with similar grapheme sequences in the bias list. In this paper, we combine the phoneme and textual information of rare words in Transducers to distinguish words with similar pronunciation or spelling. Moreover, the introduction of training with text-only data containing more rare words benefits large-scale deep biasing. The experiments on the LibriSpeech corpus demonstrate that the proposed method achieves state-of-the-art performance on rare word error rate for different scales and levels of bias lists.

cs.CL↗
arXiv2023

Learning Decentralized Traffic Signal Controllers with Multi-Agent Graph Reinforcement Learning

This paper considers optimal traffic signal control in smart cities, which has been taken as a complex networked system control problem. Given the interacting dynamics among traffic lights and road networks, attaining controller adaptivity and scalability stands out as a primary challenge. Capturing the spatial-temporal correlation among traffic lights under the framework of Multi-Agent Reinforcement Learning (MARL) is a promising solution. Nevertheless, existing MARL algorithms ignore effective information aggregation which is fundamental for improving the learning capacity of decentralized agents. In this paper, we design a new decentralized control architecture with improved environmental observability to capture the spatial-temporal correlation. Specifically, we first develop a topology-aware information aggregation strategy to extract correlation-related information from unstructured data gathered in the road network. Particularly, we transfer the road network topology into a graph shift operator by forming a diffusion process on the topology, which subsequently facilitates the construction of graph signals. A diffusion convolution module is developed, forming a new MARL algorithm, which endows agents with the capabilities of graph learning. Extensive experiments based on both synthetic and real-world datasets verify that our proposal outperforms existing decentralized algorithms.

cs.LG↗