arXiv subjects

Jian Zhou

Publications and source records attributed to Jian Zhou.

At least 109 records · Page 6Linked to original sources

arXiv2024

BSDB-Net: Band-Split Dual-Branch Network with Selective State Spaces Mechanism for Monaural Speech Enhancement

Although the complex spectrum-based speech enhancement(SE) methods have achieved significant performance, coupling amplitude and phase can lead to a compensation effect, where amplitude information is sacrificed to compensate for the phase that is harmful to SE. In addition, to further improve the performance of SE, many modules are stacked onto SE, resulting in increased model complexity that limits the application of SE. To address these problems, we proposed a dual-path network based on compressed frequency using Mamba. First, we extract amplitude and phase information through parallel dual branches. This approach leverages structured complex spectra to implicitly capture phase information and solves the compensation effect by decoupling amplitude and phase, and the network incorporates an interaction module to suppress unnecessary parts and recover missing components from the other branch. Second, to reduce network complexity, the network introduces a band-split strategy to compress the frequency dimension. To further reduce complexity while maintaining good performance, we designed a Mamba-based module that models the time and frequency dimensions under linear complexity. Finally, compared to baselines, our model achieves an average 8.3 times reduction in computational complexity while maintaining superior performance. Furthermore, it achieves a 25 times reduction in complexity compared to transformer-based models.

cs.SD↗

arXiv2024

Asymmetric electron distribution induced intrinsically strong anisotropy of thermal transport in bulk CrOCl

Anisotropic heat transfer offers promising solutions to the efficient heat dissipation in the realm of electronic device thermal management. However, the fundamental origin of the anisotropy of thermal transport remains mysterious. In this paper, by combining frequency domain thermoreflectance (FDTR) technique and first-principles-based multiscale simulations, we report the intrinsic anisotropy of thermal transport in bulk CrOCl, and further trace the origin of the anisotropy back to the fundamental electronic structures. The in-plane and cross-plane thermal conductivities ($κ$) at 300 K are found to be 21.6 and 2.18 Wm$^{-1}$K$^{-1}$, respectively, showcasing a strong $κ_\mathrm{in-plane}/κ_\mathrm{cross-plane}$ ratio of $\sim$10. Deep analysis of orbital-resolved electronic structures reveals that electrons are mainly distributed along the in-plane direction with limited interlayer distribution along the cross-plane direction, fundamentally leading to the intrinsic anisotropy of thermal transport in bulk CrOCl. The insight gained in this work sheds light on the design of advanced thermal functional materials.

physics.app-ph↗

arXiv2024

High-temperature Phonon Coherence and Tunneling Effect in Semiconductor Superlattices

Phonons, the quanta of lattice vibrations, are primary heat carriers for semiconductors and dielectrics. The demand of effective phonon manipulation urgently emerges, because the thermal management is crucial for the ongoing development of micro/nano semiconductor devices towards higher integration and power densities1, 2. Phonons also show wave-particle duality, while they are commonly treated as particle flows in current semiconductor structures3, 4. However, it sees constraints when the structure size reduces to nano and atomic scales, where the wave behavior of phonons begins to dominate, and studies of these phonon behaviors and their manipulations become long-standing challenges in experiments5. Here we show the experimental realization of coherent phonon transport, a wave-based thermal conduction fashion, in semiconductor structures. We report the successful observation of robust phonon coherence and tunneling effect in InAs/AlAs superlattices over an extensive temperature range up to 500 K, a breakthrough towards practical-application temperature for semiconductors compared with cryogenic conditions6. Our results demonstrate that the phonon coherence is robust even at a record-high interface density due to the dominating long-wavelength phonons, and the first-principles calculations clearly reveal their wave-particle duality. This revelation heralds a promising pathway towards efficient thermal phonon engineering at extreme scales, holding implications for a broad spectrum of semiconductor device applications, including microelectronics, optoelectronics, and thermoelectrics.

cond-mat.mtrl-sci↗

arXiv2024

Co-Fix3D: Enhancing 3D Object Detection with Collaborative Refinement

3D object detection in driving scenarios faces the challenge of complex road environments, which can lead to the loss or incompleteness of key features, thereby affecting perception performance. To address this issue, we propose an advanced detection framework called Co-Fix3D. Co-Fix3D integrates Local and Global Enhancement (LGE) modules to refine Bird's Eye View (BEV) features. The LGE module uses Discrete Wavelet Transform (DWT) for pixel-level local optimization and incorporates an attention mechanism for global optimization. To handle varying detection difficulties, we adopt multi-head LGE modules, enabling each module to focus on targets with different levels of detection complexity, thus further enhancing overall perception capability. Experimental results show that on the nuScenes dataset's LiDAR benchmark, Co-Fix3D achieves 69.4\% mAP and 73.5\% NDS, while on the multimodal benchmark, it achieves 72.3\% mAP and 74.7\% NDS. The source code is publicly available at \href{https://github.com/rubbish001/Co-Fix3d}{https://github.com/rubbish001/Co-Fix3d}.

cs.CV↗

arXiv2024

Uncertainty-Aware Decision-Making and Planning for Autonomous Forced Merging

In this paper, we develop an uncertainty-aware decision-making and motion-planning method for an autonomous ego vehicle in forced merging scenarios, considering the motion uncertainty of surrounding vehicles. The method dynamically captures the uncertainty of surrounding vehicles by online estimation of their acceleration bounds, enabling a reactive but rapid understanding of the uncertainty characteristics of the surrounding vehicles. By leveraging these estimated bounds, a non-conservative forward occupancy of surrounding vehicles is predicted over a horizon, which is incorporated in both the decision-making process and the motion-planning strategy, to enhance the resilience and safety of the planned reference trajectory. The method successfully fulfills the tasks in challenging forced merging scenarios, and the properties are illustrated by comparison with several alternative approaches.

cs.RO↗

arXiv2024

Constraining axion-gluon coupling in monohadron processes

The axion-gluon coupling can be constrained directly through hard exclusive processes at the LHC. Specifically, we study the associated production of a long-lived axion with a $ρ^0$ meson in ultra-peripheral $AA$ collisions and in $pp$ collisions. With the axion escaped from the detector, the final state is characterized by a mono-hadron signature. The main background in our analysis originates from the $ρ^0+π^0$ process, where the photons from the $π^0$ decay are undetected due to limited detector performance. Our analysis yields an exclusion limit of the axion-gluon coupling that is comparable to the limit obtained from the mono-jet process at the LHC.

hep-ph↗

arXiv2024

Blockchain-Envisioned UAV-Aided Disaster Relief Networks: Challenges and Solutions

Natural or man-made disasters pose significant challenges for delivering critical relief to affected populations due to disruptions in critical infrastructures and logistics networks. Unmanned aerial vehicles (UAVs)-aided disaster relief networks (UDRNs) leverage UAVs to assist existing ground relief networks by swiftly assessing affected areas and timely delivering lifesaving supplies. To meet the growing demands for collaborative, trust-free, and transparent UDRN services, blockchain-based UDRNs emerge as a promising approach through immutable ledgers and distributed smart contracts. However, several efficiency and security challenges hinder the deployment of blockchain-based UDRNs, including the lack of cooperation between smart contracts, lack of dynamic audit for smart contract vulnerabilities, and low forensics robustness against transaction malleability attacks. Towards efficient and secure blockchain-based UDRNs, this paper presents potential solutions: (i) a series of collaborative smart contracts for coordinated relief management, (ii) a dynamic contract audit mechanism to prevent known/unknown contract vulnerabilities; and (iii) a robust transaction forensics strategy with on/off-chain cooperation to resist transaction malleability attacks. Our prototype implementation and experimental results demonstrate the feasibility and effectiveness of our approach. Lastly, we outline key open research issues crucial to advancing this emerging field.

cs.CR↗

arXiv2024

Pressure induced nonmonotonic evolution of superconductivity in 6R-TaS2 with a natural bulk van der Waals heterostructure

The natural bulk van der Waals heterostructures compound 6R-TaS2 consists of alternate stacking 1T- and 1H-TaS2 monolayers, creating a unique system that incorporates charge-density-wave (CDW) order and superconductivity (SC) in distinct monolayers. Here, after confirming that the 2D nature of the lattice is preserved up to 8 GPa in 6R-TaS2, we documented an unusual evolution of CDW and SC by conducting highpressure electronic transport measurements. Upon compression, we observe a gradual suppression of CDW within the 1T-layers, while the SC exhibits a dome-shaped behavior that terminates at a critical pressure Pc around 2.9 GPa. By taking account of the fact that the substantial suppression of SC is concomitant with the complete collapse of CDW order at Pc, we argue that the 6R-TaS2 behaves like a stack of Josephson junctions and thus the suppressed superconductivity can be attributed to the weakening of Josephson coupling associated with the presence of CDW fluctuations in the 1T-layers. Furthermore, the SC reversely enhances above Pc, implying the development of emergent superconductivity in the 1T-layers after the melting of T-layer CDW orders. These results show that the 6R-TaS2 not only provides a promising platform to explore emergent phenomena but also serves as a model system to study the complex interactions between competing electronic states.

cond-mat.supr-con↗

arXiv2024

Probing quark orbital angular momentum at EIC and EicC

We propose to detect signals from quark orbital angular momentum (OAM) through exclusive $π^0$ production in electron-(longitudinally-polarized) proton collisions. Our analysis demonstrates that the $\sin 2ϕ$ azimuthal angular correlation between the transverse momentum of the scattered electron and the recoil proton serves as a sensitive probe of quark OAM. Additionally, we present a numerical estimate of the asymmetry associated with this correlation for the kinematics accessible at EIC and EicC. This study aims to pave the way for the first experimental study of quark OAM in relation to the Jaffe-Manohar spin sum rule.

hep-ph↗

arXiv2024

Bulk Photovoltaic Effects in Helimagnets

The bulk photovoltaic (BPV) effect that converts light into electric current is highly sensitive to the system symmetry and its electronic Bloch wavefunction. To create a sizable net electric current, it is necessary to break the centrosymmetry P in its host material. While prior studies mainly focus on P-broken nonmagnetic (time reversal T-reserved) and collinear antiferromagnetic PT systems, here we adopt magnetic group theory to scrutinize how helical spin polarization breaks P that is otherwise kept in its crystalline geometry. Such spin helix widely exists in intrinsic multiferroics and magnon excitation in collinear magnets. We demonstrate that magnetic symmetries are determined by spin spiral configurations, helical propagation wavevectors, magnetic chirality, and spin winding angles at the home site, offering effective and versatile manipulation of BPV generation in helimagnets. This is illustrated by tight-binding model calculations. Furthermore, we apply our theory to monolayer NiI2, a realistic showcase for helimagnetic material. Our first-principles calculations show that it could host observable direction-dependent BPV current, which could serve as a photoelectric probe to track the subtle magnetic configurations and potential multiferroic nature. This work reveals the fundamental mechanism of the interplay between spin spiral order and nonlinear optics process.

cond-mat.mtrl-sci↗

arXiv2024

Light-induced topological phase transition with tunable layer Hall effect in axion antiferromagnets

The intricate interplay between light and matter provides effective tools for manipulating topological phenomena. Here, we theoretically propose and computationally show that circularly polarized light hold the potential to transform the axion insulating phase into quantum anomalous Hall state in MnBi2Te4 thin films, featuring tunable Chern numbers (ranging up to 2). In particular, we reveal the spatial rearrangement of the hidden layer-resolved anomalous Hall effect under light driven Floquet-engineering. Notably, upon Bi2Te3 layer intercalation, the anomalous Hall conductance predominantly localizes in the nonmagnetic Bi2Te3 layers that hold zero Berry curvature in the intact state, suggesting significant magnetic proximity effect. Additionally, we estimate variations in the magneto-optical Kerr effect, giving a contactless method for detecting topological transitions. Our work not only presents a strategy to investigate emergent topological phases, but also sheds light on the possible applications of the layer Hall effect in topological antiferromagnetic spintronics.

cond-mat.mes-hall↗

arXiv2024

Boosting Medical Image-based Cancer Detection via Text-guided Supervision from Reports

The absence of adequately sufficient expert-level tumor annotations hinders the effectiveness of supervised learning based opportunistic cancer screening on medical imaging. Clinical reports (that are rich in descriptive textual details) can offer a "free lunch'' supervision information and provide tumor location as a type of weak label to cope with screening tasks, thus saving human labeling workloads, if properly leveraged. However, predicting cancer only using such weak labels can be very changeling since tumors are usually presented in small anatomical regions compared to the whole 3D medical scans. Weakly semi-supervised learning (WSSL) utilizes a limited set of voxel-level tumor annotations and incorporates alongside a substantial number of medical images that have only off-the-shelf clinical reports, which may strike a good balance between minimizing expert annotation workload and optimizing screening efficacy. In this paper, we propose a novel text-guided learning method to achieve highly accurate cancer detection results. Through integrating diagnostic and tumor location text prompts into the text encoder of a vision-language model (VLM), optimization of weakly supervised learning can be effectively performed in the latent space of VLM, thereby enhancing the stability of training. Our approach can leverage clinical knowledge by large-scale pre-trained VLM to enhance generalization ability, and produce reliable pseudo tumor masks to improve cancer detection. Our extensive quantitative experimental results on a large-scale cancer dataset, including 1,651 unique patients, validate that our approach can reduce human annotation efforts by at least 70% while maintaining comparable cancer detection accuracy to competing fully supervised methods (AUC value 0.961 versus 0.966).

cs.CV↗

arXiv2024

Learning-based Rigid Tube Model Predictive Control

This paper is concerned with model predictive control (MPC) of discrete-time linear systems subject to bounded additive disturbance and mixed constraints on the state and input, whereas the true disturbance set is unknown. Unlike most existing work on robust MPC, we propose an algorithm incorporating online learning that builds on prior knowledge of the disturbance, i.e., a known but conservative disturbance set. We approximate the true disturbance set at each time step with a parameterised set, which is referred to as a quantified disturbance set, using disturbance realisations. A key novelty is that the parameterisation of these quantified disturbance sets enjoys desirable properties such that the quantified disturbance set and its corresponding rigid tube bounding disturbance propagation can be efficiently updated online. We provide statistical gaps between the true and quantified disturbance sets, based on which, probabilistic recursive feasibility of MPC optimisation problems is discussed. Numerical simulations are provided to demonstrate the effectiveness of our proposed algorithm and compare with conventional robust MPC algorithms.

math.OC↗

arXiv2024

Accessing the gluon GTMD $\boldsymbol{F_{1,4}}$ in exclusive $\boldsymbol{π^0}$ production in $\boldsymbol{ep}$ collisions

We demonstrate that the longitudinal single target-spin asymmetry in exclusive $π^0$ production in $ep$ collisions can give access to the imaginary part of the gluon generalized transverse momentum distribution (GTMD) $F_{1,4}$. Such a longitudinal spin asymmetry that results from the Coulomb-nuclear interference effect, leads to a characteristic azimuthal angular correlation of $\sin 2ϕ$, where $ϕ$ is the azimuthal angle between the scattered lepton transverse momentum and the recoiled proton's transverse momentum. We also present a numerical estimate of the asymmetry for the kinematics accessible at EIC and EicC.

hep-ph↗

arXiv2024

Ultrafast Photocurrent Hysteresis in Photoferroelectric α-In2Se3

The photon-electron interactions are generally volatile and the intricate multiphysics details of photoexcited carrier dynamics are not yet distinguished. How to nonvolatile control the physical state through all-optical means and clarify the intricate physical processes has been a long-term goal pursued in polar materials. Photoferroelectric α-In2Se3 holds the great potential for capturing multimodal nonvolatile states due to the spontaneous reversible in-plane and out-of-plane polarizations and its tunable light-matter interactions arising from the electronic degree of freedom. Here we uncover a nonvolatile zero-bias ultrafast photocurrent hysteresis response with an all-optical scheme, diagnosed by in-plane and out-of-plane terahertz waves emitted from the photoferroelectric α-In2Se3. The mechanism of such ultrafast photocurrent hysteresis emerges as a result of anomalous bulk linear and circular photovoltaic effect synchronously driven by local polarization rearrangement. Utilizing anisotropic ferroelectric kinetics-induced relative phase between the in-plane and out-of-plane directions, we further show flexibly selective chirality, tunable rotational angle, and optimizable ellipticity of terahertz wave polarizations. Our finding offers a promising avenue towards direct ultrafast nonvolatile processing of photocurrent signals through an all-optical scheme.

physics.optics↗

arXiv2024

Efficient Surgical Tool Recognition via HMM-Stabilized Deep Learning

Recognizing various surgical tools, actions and phases from surgery videos is an important problem in computer vision with exciting clinical applications. Existing deep-learning-based methods for this problem either process each surgical video as a series of independent images without considering their dependence, or rely on complicated deep learning models to count for dependence of video frames. In this study, we revealed from exploratory data analysis that surgical videos enjoy relatively simple semantic structure, where the presence of surgical phases and tools can be well modeled by a compact hidden Markov model (HMM). Based on this observation, we propose an HMM-stabilized deep learning method for tool presence detection. A wide range of experiments confirm that the proposed approaches achieve better performance with lower training and running costs, and support more flexible ways to construct and utilize training data in scenarios where not all surgery videos of interest are extensively labelled. These results suggest that popular deep learning approaches with over-complicated model structures may suffer from inefficient utilization of data, and integrating ingredients of deep learning and statistical learning wisely may lead to more powerful algorithms that enjoy competitive performance, transparent interpretation and convenient model training simultaneously.

cs.CV↗

arXiv2024

Medical Visual Prompting (MVP): A Unified Framework for Versatile and High-Quality Medical Image Segmentation

Accurate segmentation of lesion regions is crucial for clinical diagnosis and treatment across various diseases. While deep convolutional networks have achieved satisfactory results in medical image segmentation, they face challenges such as loss of lesion shape information due to continuous convolution and downsampling, as well as the high cost of manually labeling lesions with varying shapes and sizes. To address these issues, we propose a novel medical visual prompting (MVP) framework that leverages pre-training and prompting concepts from natural language processing (NLP). The framework utilizes three key components: Super-Pixel Guided Prompting (SPGP) for superpixelating the input image, Image Embedding Guided Prompting (IEGP) for freezing patch embedding and merging with superpixels to provide visual prompts, and Adaptive Attention Mechanism Guided Prompting (AAGP) for pinpointing prompt content and efficiently adapting all layers. By integrating SPGP, IEGP, and AAGP, the MVP enables the segmentation network to better learn shape prompting information and facilitates mutual learning across different tasks. Extensive experiments conducted on five datasets demonstrate superior performance of this method in various challenging medical image tasks, while simplifying single-task medical segmentation models. This novel framework offers improved performance with fewer parameters and holds significant potential for accurate segmentation of lesion regions in various medical tasks, making it clinically valuable.

cs.CV↗

arXiv2024

Existence of solutions for a class of Kirchhoff-type equations with indefinite potential

In this paper, we consider the existence of solutions of the following Kirchhoff-type problem \[ \left\{ \begin{array} [c]{ll} -\left(a+b\int_{\mathbb{R}^3}|\nabla u|^2dx\right)Δu+ V(x)u=f(x,u),~{\rm{in}}~ \mathbb{R}^{3},\\ u\in H^1(\mathbb{R}^3), \end{array} \right. \] where $a,b$ are postive constants, and the potential $V(x)$ is continuous and indefinite in sign. Under some suitable assumptions on $V(x)$ and $f$, we obtain the existence of solutions by the Symmetric Mountain Pass Theorem.

math.AP↗