Search arXiv⌕ Search

arXiv subjects

Li Zhang

Publications and source records attributed to Li Zhang.

At least 613 records · Page 34Linked to original sources

First extraction of the proton mass radius and scattering length $\left|α_{ρ^0 p}\right| $ from $ρ^0$ photoproduction

As it involves the lightest physical states excited from the vacuum by the vector quark current, near-threshold $ρ^0$ photoproduction is considered a possible way to research the proton radius and the absolute value of the scattering lengths of the $ρ^0$--proton interaction. In this work, under the assumption of a scalar form factor of dipole form, the value of the proton mass radius is calculated as $0.85\pm 0.06 \text{ fm }$ by fitting the differential cross section of the $γp \rightarrow ρ^0 p$ reaction at near-threshold energy. For light vector meson photoproduction, because the exchange of a scalar quark--antiquark pair is not suppressed and should dominate the scalar gluon exchange, the radius we extract from $ρ^0$ photoproduction is likely to represent the quark radius of the proton. This fact may explain why the value obtained in this work is very near the proton charge radius. Moreover, the absolute value of the $ρ^0$--proton scattering length $|α_{ρ^0 p}|= 0.31 \pm 0.06 \text{ fm}$ is obtained for the first time within the vector meson dominance model. This result disobeys the rule that the absolute value of the vector meson and proton scattering length $|α_{V p}|$ increases with the meson's mass, which can be attributed to treating the $ρ^0$ meson as a point in the analysis. These results provide useful theoretical information for an in-depth understanding of proton structure and the proton--vector meson interaction.

nucl-th↗

TSUP Speaker Diarization System for Conversational Short-phrase Speaker Diarization Challenge

This paper describes the TSUP team's submission to the ISCSLP 2022 conversational short-phrase speaker diarization (CSSD) challenge which particularly focuses on short-phrase conversations with a new evaluation metric called conversational diarization error rate (CDER). In this challenge, we explore three kinds of typical speaker diarization systems, which are spectral clustering(SC) based diarization, target-speaker voice activity detection(TS-VAD) and end-to-end neural diarization(EEND) respectively. Our major findings are summarized as follows. First, the SC approach is more favored over the other two approaches under the new CDER metric. Second, tuning on hyperparameters is essential to CDER for all three types of speaker diarization systems. Specifically, CDER becomes smaller when the length of sub-segments setting longer. Finally, multi-system fusion through DOVER-LAP will worsen the CDER metric on the challenge data. Our submitted SC system eventually ranks the third place in the challenge.

cs.SD↗

Generative Model Watermarking Based on Human Visual System

Intellectual property protection of deep neural networks is receiving attention from more and more researchers, and the latest research applies model watermarking to generative models for image processing. However, the existing watermarking methods designed for generative models do not take into account the effects of different channels of sample images on watermarking. As a result, the watermarking performance is still limited. To tackle this problem, in this paper, we first analyze the effects of embedding watermark information on different channels. Then, based on the characteristics of human visual system (HVS), we introduce two HVS-based generative model watermarking methods, which are realized in RGB color space and YUV color space respectively. In RGB color space, the watermark is embedded into the R and B channels based on the fact that HVS is more sensitive to G channel. In YUV color space, the watermark is embedded into the DCT domain of U and V channels based on the fact that HVS is more sensitive to brightness changes. Experimental results demonstrate the effectiveness of the proposed work, which improves the fidelity of the model to be protected and has good universality compared with previous methods.

cs.CV↗

Lattice-aligned gallium oxynitride nanolayer for GaN surface enhancement and function extension

Gallium nitride (GaN), as a promising alternative semiconductor to silicon, is of well-established use in photoelectronic and electronic technology. However, the vulnerable GaN surface has been a critical restriction that hinders the development of GaN-based devices, especially regarding device stability and reliability. Here, we overcome this challenge by converting the GaN surface into a gallium oxynitride (GaON) epitaxial nanolayer through an in-situ two-step "oxidation-reconfiguration" process. The oxygen plasma treatment overcomes the chemical inertness of the GaN surface, and the sequential thermal annealing manipulates the kinetic-thermodynamic reaction pathways to create a metastable GaON nanolayer with wurtzite lattice. This GaN-derived GaON nanolayer is a tailored structure for surface reinforcement and possesses several advantages, including wide bandgap, high thermodynamic stability, and large valence band offset with GaN substrate. These enhanced physical properties can be further leveraged to enable GaN-based applications in new scenarios, such as complementary logic integrated circuits, photoelectrochemical water splitting, and ultraviolet photoelectric conversion, making GaON a versatile functionality extender.

physics.app-ph↗

NWPU-ASLP System for the VoicePrivacy 2022 Challenge

This paper presents the NWPU-ASLP speaker anonymization system for VoicePrivacy 2022 Challenge. Our submission does not involve additional Automatic Speaker Verification (ASV) model or x-vector pool. Our system consists of four modules, including feature extractor, acoustic model, anonymization module, and neural vocoder. First, the feature extractor extracts the Phonetic Posteriorgram (PPG) and pitch from the input speech signal. Then, we reserve a pseudo speaker ID from a speaker look-up table (LUT), which is subsequently fed into a speaker encoder to generate the pseudo speaker embedding that is not corresponding to any real speaker. To ensure the pseudo speaker is distinguishable, we further average the randomly selected speaker embedding and weighted concatenate it with the pseudo speaker embedding to generate the anonymized speaker embedding. Finally, the acoustic model outputs the anonymized mel-spectrogram from the anonymized speaker embedding and a modified version of HifiGAN transforms the mel-spectrogram into the anonymized speech waveform. Experimental results demonstrate the effectiveness of our proposed anonymization system.

eess.AS↗

Spatial-temporal Transformers for EEG Emotion Recognition

Electroencephalography (EEG) is a popular and effective tool for emotion recognition. However, the propagation mechanisms of EEG in the human brain and its intrinsic correlation with emotions are still obscure to researchers. This work proposes four variant transformer frameworks~(spatial attention, temporal attention, sequential spatial-temporal attention and simultaneous spatial-temporal attention) for EEG emotion recognition to explore the relationship between emotion and spatial-temporal EEG features. Specifically, spatial attention and temporal attention are to learn the topological structure information and time-varying EEG characteristics for emotion recognition respectively. Sequential spatial-temporal attention does the spatial attention within a one-second segment and temporal attention within one sample sequentially to explore the influence degree of emotional stimulation on EEG signals of diverse EEG electrodes in the same temporal segment. The simultaneous spatial-temporal attention, whose spatial and temporal attention are performed simultaneously, is used to model the relationship between different spatial features in different time segments. The experimental results demonstrate that simultaneous spatial-temporal attention leads to the best emotion recognition accuracy among the design choices, indicating modeling the correlation of spatial and temporal features of EEG signals is significant to emotion recognition.

cs.RO↗

Backend Ensemble for Speaker Verification and Spoofing Countermeasure

This paper describes the NPU system submitted to Spoofing Aware Speaker Verification Challenge 2022. We particularly focus on the \textit{backend ensemble} for speaker verification and spoofing countermeasure from three aspects. Firstly, besides simple concatenation, we propose circulant matrix transformation and stacking for speaker embeddings and countermeasure embeddings. With the stacking operation of newly-defined circulant embeddings, we almost explore all the possible interactions between speaker embeddings and countermeasure embeddings. Secondly, we attempt different convolution neural networks to selectively fuse the embeddings' salient regions into channels with convolution kernels. Finally, we design parallel attention in 1D convolution neural networks to learn the global correlation in channel dimensions as well as to learn the important parts in feature dimensions. Meanwhile, we embed squeeze-and-excitation attention in 2D convolutional neural networks to learn the global dependence among speaker embeddings and countermeasure embeddings. Experimental results demonstrate that all the above methods are effective. After fusion of four well-trained models enhanced by the mentioned methods, the best SASV-EER, SPF-EER and SV-EER we achieve are 0.559\%, 0.354\% and 0.857\% on the evaluation set respectively. Together with the above contributions, our submission system achieves the fifth place in this challenge.

cs.SD↗

Extraction-based Deep Learning Reconstruction of Interior Tomography

Interior tomography is a typical strategy for radiation dose reduction in computed tomography, where only a certain region-of-interest (ROI) is scanned. However, given the truncated projection data, ROI reconstruction by conventional analytical algorithms may suffer from severe cupping artifacts. In this paper, we proposed a new extraction-based deep learning method for the reconstruction of interior tomography. Our approach works in dual domains where a sinogram-domain network (SDNet) estimates the contribution of the exterior region to the truncated projection and an image-domain network (IDNet) further mitigates artifacts. Unlike the previous extrapolation-based methods, SDNet is intended to obtain a complete ROI-only sinogram via extraction instead of a fully non-truncated sinogram for both the ROI and exterior regions. Our experiments validated the proposed method and the results indicate that the proposed method can disclose more reliable structures. It achieved better image quality with better generalization performance than extrapolation-based methods.

physics.med-ph↗

Dynamic Graph Message Passing Networks for Visual Recognition

Modelling long-range dependencies is critical for scene understanding tasks in computer vision. Although convolution neural networks (CNNs) have excelled in many vision tasks, they are still limited in capturing long-range structured relationships as they typically consist of layers of local kernels. A fully-connected graph, such as the self-attention operation in Transformers, is beneficial for such modelling, however, its computational overhead is prohibitive. In this paper, we propose a dynamic graph message passing network, that significantly reduces the computational complexity compared to related works modelling a fully-connected graph. This is achieved by adaptively sampling nodes in the graph, conditioned on the input, for message passing. Based on the sampled nodes, we dynamically predict node-dependent filter weights and the affinity matrix for propagating information between them. This formulation allows us to design a self-attention module, and more importantly a new Transformer-based backbone network, that we use for both image classification pretraining, and for addressing various downstream tasks (object detection, instance and semantic segmentation). Using this model, we show significant improvements with respect to strong, state-of-the-art baselines on four different tasks. Our approach also outperforms fully-connected graphs while using substantially fewer floating-point operations and parameters. Code and models will be made publicly available at https://github.com/fudan-zvg/DGMN2

cs.CV↗

Dynamic Graph Message Passing Networks

Modelling long-range dependencies is critical for scene understanding tasks in computer vision. Although CNNs have excelled in many vision tasks, they are still limited in capturing long-range structured relationships as they typically consist of layers of local kernels. A fully-connected graph is beneficial for such modelling, however, its computational overhead is prohibitive. We propose a dynamic graph message passing network, that significantly reduces the computational complexity compared to related works modelling a fully-connected graph. This is achieved by adaptively sampling nodes in the graph, conditioned on the input, for message passing. Based on the sampled nodes, we dynamically predict node-dependent filter weights and the affinity matrix for propagating information between them. Using this model, we show significant improvements with respect to strong, state-of-the-art baselines on three different tasks and backbone architectures. Our approach also outperforms fully-connected graphs while using substantially fewer floating-point operations and parameters. The project website is http://www.robots.ox.ac.uk/~lz/dgmn/

cs.CV↗

Model-Guided Multi-Contrast Deep Unfolding Network for MRI Super-resolution Reconstruction

Magnetic resonance imaging (MRI) with high resolution (HR) provides more detailed information for accurate diagnosis and quantitative image analysis. Despite the significant advances, most existing super-resolution (SR) reconstruction network for medical images has two flaws: 1) All of them are designed in a black-box principle, thus lacking sufficient interpretability and further limiting their practical applications. Interpretable neural network models are of significant interest since they enhance the trustworthiness required in clinical practice when dealing with medical images. 2) most existing SR reconstruction approaches only use a single contrast or use a simple multi-contrast fusion mechanism, neglecting the complex relationships between different contrasts that are critical for SR improvement. To deal with these issues, in this paper, a novel Model-Guided interpretable Deep Unfolding Network (MGDUN) for medical image SR reconstruction is proposed. The Model-Guided image SR reconstruction approach solves manually designed objective functions to reconstruct HR MRI. We show how to unfold an iterative MGDUN algorithm into a novel model-guided deep unfolding network by taking the MRI observation matrix and explicit multi-contrast relationship matrix into account during the end-to-end optimization. Extensive experiments on the multi-contrast IXI dataset and BraTs 2019 dataset demonstrate the superiority of our proposed model.

eess.IV↗

Stable interaction-induced Anderson-like localization embedded in standing waves

We uncover the interaction-induced \emph{stable self-localization} of bosons in disorder-free superlattices. In these nonthermalized multi-particle states, one of the particles forms a superposition of multiple standing waves, so that it provides a quasirandom potential to localize the other particles. We derive effective Hamiltonians for self-localized states and find their energy level spacings obeying the Poisson statistics for Anderson-like localization. Surprisingly, we find that the correlated self-localization can be solely induced by interaction in the well-studied nonintegrable Bose-Hubbard models, which has been overlooked for a long time. We propose a dynamical scheme to detect self-localization, where long-time quantum walks of a single particle form a superposition of multiple standing waves for trapping the subsequently loaded particles. Our work provides an experimentally feasible way to realize stable Anderson-like localization in translation-invariant disorder-free systems.

quant-ph↗

Data-Driven Deep Supervision for Skin Lesion Classification

Automatic classification of pigmented, non-pigmented, and depigmented non-melanocytic skin lesions have garnered lots of attention in recent years. However, imaging variations in skin texture, lesion shape, depigmentation contrast, lighting condition, etc. hinder robust feature extraction, affecting classification accuracy. In this paper, we propose a new deep neural network that exploits input data for robust feature extraction. Specifically, we analyze the convolutional network's behavior (field-of-view) to find the location of deep supervision for improved feature extraction. To achieve this, first, we perform activation mapping to generate an object mask, highlighting the input regions most critical for classification output generation. Then the network layer whose layer-wise effective receptive field matches the approximated object shape in the object mask is selected as our focus for deep supervision. Utilizing different types of convolutional feature extractors and classifiers on three melanoma detection datasets and two vitiligo detection datasets, we verify the effectiveness of our new method.

cs.CV↗

Scalable Nanophotonic-Electronic Spiking Neural Networks

Spiking neural networks (SNN) provide a new computational paradigm capable of highly parallelized, real-time processing. Photonic devices are ideal for the design of high-bandwidth, parallel architectures matching the SNN computational paradigm. Co-integration of CMOS and photonic elements allow low-loss photonic devices to be combined with analog electronics for greater flexibility of nonlinear computational elements. As such, we designed and simulated an optoelectronic spiking neuron circuit on a monolithic silicon photonics (SiPh) process that replicates useful spiking behaviors beyond the leaky integrate-and-fire (LIF). Additionally, we explored two learning algorithms with the potential for on-chip learning using Mach-Zehnder Interferometric (MZI) meshes as synaptic interconnects. A variation of Random Backpropagation (RPB) was experimentally demonstrated on-chip and matched the performance of a standard linear regression on a simple classification task. Meanwhile, the Contrastive Hebbian Learning (CHL) rule was applied to a simulated neural network composed of MZI meshes for a random input-output mapping task. The CHL-trained MZI network performed better than random guessing but does not match the performance of the ideal neural network (without the constraints imposed by the MZI meshes). Through these efforts, we demonstrate that co-integrated CMOS and SiPh technologies are well-suited to the design of scalable SNN computing architectures.

cs.NE↗

Hierarchical Reinforcement Learning Based Video Semantic Coding for Segmentation

The rapid development of intelligent tasks, e.g., segmentation, detection, classification, etc, has brought an urgent need for semantic compression, which aims to reduce the compression cost while maintaining the original semantic information. However, it is impractical to directly integrate the semantic metric into the traditional codecs since they cannot be optimized in an end-to-end manner. To solve this problem, some pioneering works have applied reinforcement learning to implement image-wise semantic compression. Nevertheless, video semantic compression has not been explored since its complex reference architectures and compression modes. In this paper, we take a step forward to video semantic compression and propose the Hierarchical Reinforcement Learning based task-driven Video Semantic Coding, named as HRLVSC. Specifically, to simplify the complex mode decision of video semantic coding, we divided the action space into frame-level and CTU-level spaces in a hierarchical manner, and then explore the best mode selection for them progressively with the cooperation of frame-level and CTU-level agents. Moreover, since the modes of video semantic coding will exponentially increase with the number of frames in a Group of Pictures (GOP), we carefully investigate the effects of different mode selections for video semantic coding and design a simple but effective mode simplification strategy for it. We have validated our HRLVSC on the video segmentation task with HEVC reference software HM16.19. Extensive experimental results demonstrated that our HRLVSC can achieve over 39% BD-rate saving for video semantic coding under the Low Delay P configuration.

eess.IV↗

Fabrication and characterization of InSb nanosheet/hBN/graphite heterostructure devices

Semiconductor InSb nanosheet/hexagonal boron nitride (hBN)/graphite trilayers are fabricated, and single- and double-gate devices made from the trilayers are realized and characterized. The InSb nanosheets employed in the trilayer devices are epitaxially grown, free-standing, zincblende crystals and are in micrometer lateral sizes. The hBN and graphite flakes are obtained by exfoliation. Each trilayer is made by successively stacking an InSb nanosheet on an hBN flake and on a graphite flake using a home-made alignment stacking/transfer setup. The fabricated single- and double-gate devices are characterized by electrical and/or magnetotransport measurements. In all these devices, the graphite and hBN flakes are employed as the bottom gates and the gate dielectrics. The measurements of a fabricated single bottom-gate field-effect device show that the InSb nanosheet in the device has an electron field-effect mobility of ~7300 cm$^2$V$^{-1}$s$^{-1}$ and a low gate hysteresis of ~0.05 V at 1.9 K. The measurements of a double-gate Hall-bar device show that both the top and the bottom gate exhibit strong capacitive couplings to the InSb nanosheet channel and can thus tune the nanosheet channel conduction effectively. The electron Hall mobility in the InSb nanosheet of the Hall-bar device is extracted to be larger than 1.1$\times$10$^4$ cm$^2$V$^{-1}$s$^{-1}$ at a sheet electron density of ~6.1$\times$10$^{11}$ cm$^{-2}$ and 1.9 K and, thus, the device exhibits well-defined Shubnikov-de Haas oscillations.

cond-mat.mes-hall↗

Quantum oscillations in a hexagonal boron nitride-supported single crystalline InSb nanosheet

A gated Hall-bar device is made from an epitaxially grown, free-standing InSb nanosheet on a hexagonal boron nitride (hBN) dielectric/graphite gate structure and the electron transport properties in the InSb nanosheet are studied by gate-transfer characteristic and magnetotransport measurements at low temperatures. The measurements show that the carriers in the InSb nanosheet are of electrons and the carrier density in the nanosheet can be highly efficiently tuned by the graphite gate. The mobility of the electrons in the InSb nanosheet is extracted from low-field magneotransport measurements and a value of the mobility exceeding ~18000 cm$^2$V$^{-1}$s$^{-1}$ is found. High-field magentotransport measurements show well-defined Shubnikov-de Haas (SdH) oscillations in the longitudinal resistance of the InSb nanosheet. Temperature-dependent measurements of the SdH oscillations are carried out and key transport parameters, including the electron effective mass m$^*$$\sim$0.028 m$_0$ and the quantum lifetime $τ$$\sim$0.046 ps, in the InSb nanosheet are extracted. It is for the first time that such experimental measurements have been reported for a free-standing InSb nanosheet and the results obtained indicate that InSb nanosheet/hBN/graphite gate structures can be used to develop advanced quantum devices for novel physics studies and for quantum technology applications.

cond-mat.mes-hall↗

Learning Ego 3D Representation as Ray Tracing

A self-driving perception model aims to extract 3D semantic representations from multiple cameras collectively into the bird's-eye-view (BEV) coordinate frame of the ego car in order to ground downstream planner. Existing perception methods often rely on error-prone depth estimation of the whole scene or learning sparse virtual 3D representations without the target geometry structure, both of which remain limited in performance and/or capability. In this paper, we present a novel end-to-end architecture for ego 3D representation learning from an arbitrary number of unconstrained camera views. Inspired by the ray tracing principle, we design a polarized grid of "imaginary eyes" as the learnable ego 3D representation and formulate the learning process with the adaptive attention mechanism in conjunction with the 3D-to-2D projection. Critically, this formulation allows extracting rich 3D representation from 2D images without any depth supervision, and with the built-in geometry structure consistent w.r.t. BEV. Despite its simplicity and versatility, extensive experiments on standard BEV visual tasks (e.g., camera-based 3D object detection and BEV segmentation) show that our model outperforms all state-of-the-art alternatives significantly, with an extra advantage in computational efficiency from multi-task learning.

cs.CV↗