arXiv subjects

Junjie Li

Publications and source records attributed to Junjie Li.

At least 145 records · Page 8Linked to original sources

arXiv2022

Domain Adaptive Person Search

Person search is a challenging task which aims to achieve joint pedestrian detection and person re-identification (ReID). Previous works have made significant advances under fully and weakly supervised settings. However, existing methods ignore the generalization ability of the person search models. In this paper, we take a further step and present Domain Adaptive Person Search (DAPS), which aims to generalize the model from a labeled source domain to the unlabeled target domain. Two major challenges arises under this new setting: one is how to simultaneously solve the domain misalignment issue for both detection and Re-ID tasks, and the other is how to train the ReID subtask without reliable detection results on the target domain. To address these challenges, we propose a strong baseline framework with two dedicated designs. 1) We design a domain alignment module including image-level and task-sensitive instance-level alignments, to minimize the domain discrepancy. 2) We take full advantage of the unlabeled data with a dynamic clustering strategy, and employ pseudo bounding boxes to support ReID and detection training on the target domain. With the above designs, our framework achieves 34.7% in mAP and 80.6% in top-1 on PRW dataset, surpassing the direct transferring baseline by a large margin. Surprisingly, the performance of our unsupervised DAPS model even surpasses some of the fully and weakly supervised methods. The code is available at https://github.com/caposerenity/DAPS.

cs.CV↗

arXiv2022

Revisiting the transient coarsening kinetics: a new framework in the Lifshitz-Slyozov-Wagner space

Phase coarsening is a fundamental process of microstructure evolution in multiphase materials. A thorough understanding of its kinetics is of great significance for material processing and performance. Generally, coarsening can be divided into the transient stage and the steady stage. Compared with steady coarsening kinetics, the current understanding of transient coarsening is rather limited and contradictory. In the present work, a new framework in the dimensionless Lifshitz-Slyozov-Wagner space is developed to study transient coarsening kinetics co-controlled by interface migration/reaction and matrix diffusion, where the dynamic equation for individual particles is derived from the thermodynamic extremal principle.

cond-mat.mtrl-sci↗

arXiv2022

First Experiences in Performance Benchmarking with the New SPEChpc 2021 Suites

Modern HPC systems are built with innovative system architectures and novel programming models to further push the speed limit of computing. The increased complexity poses challenges for performance portability and performance evaluation. The Standard Performance Evaluation Corporation -SPEC has a long history of producing industry standard benchmarks for modern computer systems. SPEC is a newly released SPEChpc 2021 benchmark suites, developed by the High Performance Group, are a bold attempt to provide a fair and objective benchmarking tool designed for state of the art HPC systems. With the support of multiple host and accelerator programming models, the suites are portable across both homogeneous and heterogeneous architectures. Different workloads are developed to fit system sizes ranging from a few compute nodes to a few hundred compute nodes. In this manuscript, we take a first glance at these benchmark suites and evaluate their portability and basic performance characteristics on various popular and emerging HPC architectures, including x86 CPU, NVIDIA GPU, and AMD GPU. This study provides a first-hand experience of executing the SPEChpc 2021 suites at scale on production HPC systems, discusses real-world use cases, and serves as an initial guideline for using the benchmark suites.

cs.DC↗

arXiv2022

Low-confidence Samples Matter for Domain Adaptation

Domain adaptation (DA) aims to transfer knowledge from a label-rich source domain to a related but label-scarce target domain. The conventional DA strategy is to align the feature distributions of the two domains. Recently, increasing researches have focused on self-training or other semi-supervised algorithms to explore the data structure of the target domain. However, the bulk of them depend largely on confident samples in order to build reliable pseudo labels, prototypes or cluster centers. Representing the target data structure in such a way would overlook the huge low-confidence samples, resulting in sub-optimal transferability that is biased towards the samples similar to the source domain. To overcome this issue, we propose a novel contrastive learning method by processing low-confidence samples, which encourages the model to make use of the target data structure through the instance discrimination process. To be specific, we create positive and negative pairs only using low-confidence samples, and then re-represent the original features with the classifier weights rather than directly utilizing them, which can better encode the task-specific semantic information. Furthermore, we combine cross-domain mixup to augment the proposed contrastive loss. Consequently, the domain gap can be well bridged through contrastive learning of intermediate representations across domains. We evaluate the proposed method in both unsupervised and semi-supervised DA settings, and extensive experimental results on benchmarks reveal that our method is effective and achieves state-of-the-art performance. The code can be found in https://github.com/zhyx12/MixLRCo.

cs.CV↗

arXiv2022

Securing Reconfigurable Intelligent Surface-Aided Cell-Free Networks

In this paper, we investigate the physical layer security in the reconfigurable intelligent surface (RIS)-aided cell-free networks. A maximum weighted sum secrecy rate problem is formulated by jointly optimizing the active beamforming (BF) at the base stations and passive BF at the RISs. To handle this non-trivial problem, we adopt the alternating optimization to decouple the original problem into two sub-ones, which are solved using the semidefinite relaxation and continuous convex approximation theory. To decrease the complexity for obtaining overall channel state information (CSI), we extend the proposed framework to the case that only requires part of the RIS' CSI. This is achieved via deliberately discarding the RIS that has a small contribution to the user's secrecy rate. Based on this, we formulate a mixed integer non-linear programming problem, and the linear conic relaxation is used to obtained the solutions. Finally, the simulation results show that the proposed schemes can obtain a higher secrecy rate than the existing ones.

cs.IT↗

arXiv2022

Research on Dual Channel News Headline Classification Based on ERNIE Pre-training Model

The classification of news headlines is an important direction in the field of NLP, and its data has the characteristics of compactness, uniqueness and various forms. Aiming at the problem that the traditional neural network model cannot adequately capture the underlying feature information of the data and cannot jointly extract key global features and deep local features, a dual-channel network model DC-EBAD based on the ERNIE pre-training model is proposed. Use ERNIE to extract the lexical, semantic and contextual feature information at the bottom of the text, generate dynamic word vector representations fused with context, and then use the BiLSTM-AT network channel to secondary extract the global features of the data and use the attention mechanism to give key parts higher The weight of the DPCNN channel is used to overcome the long-distance text dependence problem and obtain deep local features. The local and global feature vectors are spliced, and finally passed to the fully connected layer, and the final classification result is output through Softmax. The experimental results show that the proposed model improves the accuracy, precision and F1-score of news headline classification compared with the traditional neural network model and the single-channel model under the same conditions. It can be seen that it can perform well in the multi-classification application of news headline text under large data volume.

cs.CL↗

arXiv2022

In-situ comparison of interface instability of basal and edge planes during unidirectional growth of sea ice

The unique anisotropy of ice has endowed sea ice growth a peculiar and attractive subject from both fundamental and applied viewpoints. The distinct growth behaviors between edge and basal plane of ice are one of the central topics in ice growth. And the unidirectional freezing pattern stems from perturbations of both basal and edge planes. To date there is no direct comparison of unidirectional freezing behavior between basal and edge plane ice. Here, we in-situ investigate the planar instability as well as the unidirectional freezing pattern of basal and edge planes of ice by a design of parallel freezing samples with specified ice orientations in a NaCl solution as a modeled sea water. The planar instability is discussed via neutral stability curves with surface tension anisotropy for both basal and edge plane ice. For the first time, we realize the simultaneous observation of solid/liquid interfaces of basal and edge plane ice under the same set of freezing conditions. The results show that planar instability occurs faster for edge plane ice than basal plane ice. The time-lapse observations confirm a transient competitive interaction of perturbations between the basal and edge planes ice, which is explained by the anisotropic growth of perturbations in basal and edge planes of ice. These experimental results provide a link between morphology evolution of unidirectional grown sea ice and different ice orientations and are suggested to enrich our understanding of sea ice growth as well as crystallization pattern of other anisotropic materials.

cond-mat.mtrl-sci↗

arXiv2021

Yielding and jerky plasticity of tilt grain boundaries in high-temperature graphene

Graphene is well known for its extraordinary mechanical properties combining brittleness and ductility. While most mechanical studies of graphene focused on the strength and brittle fracture behavior, its ductility, plastic deformation, and the possible brittle-to-ductile transition, which are important for high-temperature mechanical performance and applications, still remain much less understood. Here the mechanical response and deformation dynamics of graphene grain boundaries are investigated through a phase field crystal modeling, showing the pivotal effects of temperature and local dislocation structure. Our results indicate that even at relatively high temperature (around 3350 K), the system is still governed by a brittle fracture and cracking dynamics as found in previous low-temperature experimental and atomistic studies. We also identify another type of failure dynamics with low-angle grain boundary disintegration. When temperature increases a transition to plastic deformation is predicted. The appearance of plastic flow at ultrahigh temperature, particularly the phenomenon of jerky plasticity, is attributed to the stick and climb-glide motion of dislocations around the grain boundary. The corresponding mechanism is intrinsic to two-dimensional systems, and governed by the competition between the driving force of accumulated local stress and the defect pinning effect, without the traditional pathways of dislocation generation needed in three-dimensional materials.

cond-mat.mtrl-sci↗

arXiv2021

Planar chiral metasurfaces with maximal tunable chiroptical response driven by bound states in the continuum

Optical metasurfaces with high-Q chiral resonances can boost light-matter interaction for various applications of chiral response for ultrathin, active, and nonlinear metadevices. Usually, such metasurfaces require sophisticated depth-resolved nanofabrication to realize subwavelength stereo-nanostructures, posing overwhelming challenges, especially in the short-wavelength range. Here, we suggest a novel planar design for chiral metasurfaces supporting bound states in the continuum (BICs) and demonstrate experimentally chiroptical responses with record-high Q-factors (Q=390) and near-perfect circular dichroism (CD=0.93) at optical frequencies. The symmetry-reduced meta-atoms are highly birefringent and support winding elliptical eigen-polarizations with opposite helicity surrounding the BIC polarization singularity, providing a convenient way for achieving maximal planar chirality tuned by either breaking in-plane symmetry or changing illumination direction. Such sharply resonant chirality realized in planar metasurfaces promises various practical applications in classical and quantum optics including chiral sensing, enantiomer selection, and chiral quantum emitters.

physics.optics↗

arXiv2021

5th Place Solution for VSPW 2021 Challenge

In this article, we introduce the solution we used in the VSPW 2021 Challenge. Our experiments are based on two baseline models, Swin Transformer and MaskFormer. To further boost performance, we adopt stochastic weight averaging technique and design hierarchical ensemble strategy. Without using any external semantic segmentation dataset, our solution ranked the 5th place in the private leaderboard. Besides, we have some interesting attempts to tackle long-tail recognition and overfitting issues, which achieves improvement on val subset. Maybe due to distribution difference, these attempts don't work on test subset. We will also introduce these attempts and hope to inspire other researchers.

cs.CV↗

arXiv2021

Revealing curvature and stochastic effects on grain growth kinetics: A thermodynamic perspective from the extremal principle

In the theoretical development of normal grain growth, the roles of "drift" (curvature effect) and "diffusion" (stochastic effect) have been an open question for many years. By coupling contributions of microstructure entropy and grain topological interactions with thermodynamic extremal principle (TEP), this letter extends the existing thermodynamic framework of grain growth. It not only explains the curvature and stochastic effects by thermodynamics but provides a new mean-field model for grain growth. Through thermodynamic modification, this new model can yield a high accuracy representation of existing ultra-large phase-field simulated results.

cond-mat.mtrl-sci↗

arXiv2021

TAL: Two-stream Adaptive Learning for Generalizable Person Re-identification

Domain generalizable person re-identification aims to apply a trained model to unseen domains. Prior works either combine the data in all the training domains to capture domain-invariant features, or adopt a mixture of experts to investigate domain-specific information. In this work, we argue that both domain-specific and domain-invariant features are crucial for improving the generalization ability of re-id models. To this end, we design a novel framework, which we name two-stream adaptive learning (TAL), to simultaneously model these two kinds of information. Specifically, a domain-specific stream is proposed to capture training domain statistics with batch normalization (BN) parameters, while an adaptive matching layer is designed to dynamically aggregate domain-level information. In the meantime, we design an adaptive BN layer in the domain-invariant stream, to approximate the statistics of various unseen domains. These two streams work adaptively and collaboratively to learn generalizable re-id features. Our framework can be applied to both single-source and multi-source domain generalization tasks, where experimental results show that our framework notably outperforms the state-of-the-art methods.

cs.CV↗

arXiv2021

Photoinduced anisotropic lattice dynamic response and domain formation in thermoelectric SnSe

Identifying and understanding the mechanisms behind strong phonon-phonon scattering in condensed matter systems is critical to maximizing the efficiency of thermoelectric devices. To date, the leading method to address this has been to meticulously survey the full phonon dispersion of the material in order to isolate modes with anomalously large linewidth and temperature-dependence. Here we combine quantitative MeV ultrafast electron diffraction (UED) analysis with Monte Carlo based dynamic diffraction simulation and first-principles calculations to directly unveil the soft, anharmonic lattice distortions of model thermoelectric material SnSe. A small single-crystal sample is photoexcited with ultrafast optical pulses and the soft, anharmonic lattice distortions are isolated using MeV-UED as those associated with long relaxation time and large displacements. We reveal that these modes have interlayer shear strain character, induced mainly by c-axis atomic displacements, resulting in domain formation in the transient state. These findings provide an innovative approach to identify mechanisms for ultralow and anisotropic thermal conductivity and a promising route to optimizing thermoelectric devices.

cond-mat.mtrl-sci↗

arXiv2021

Evolutionary Generative Adversarial Networks with Crossover Based Knowledge Distillation

Generative Adversarial Networks (GAN) is an adversarial model, and it has been demonstrated to be effective for various generative tasks. However, GAN and its variants also suffer from many training problems, such as mode collapse and gradient vanish. In this paper, we firstly propose a general crossover operator, which can be widely applied to GANs using evolutionary strategies. Then we design an evolutionary GAN framework C-GAN based on it. And we combine the crossover operator with evolutionary generative adversarial networks (EGAN) to implement the evolutionary generative adversarial networks with crossover (CE-GAN). Under the premise that a variety of loss functions are used as mutation operators to generate mutation individuals, we evaluate the generated samples and allow the mutation individuals to learn experiences from the output in a knowledge distillation manner, imitating the best output outcome, resulting in better offspring. Then, we greedily selected the best offspring as parents for subsequent training using discriminator as evaluator. Experiments on real datasets demonstrate the effectiveness of CE-GAN and show that our method is competitive in terms of generated images quality and time efficiency.

cs.LG↗

arXiv2021

Enhanced superconductivity in bilayer PtTe$_2$ by alkali-metal intercalations

Layered platinum tellurium (PtTe2) was recently synthesized with controllable layer numbers down to a monolayer limit. Using ab initio calculations based on anisotropic Midgal-Eliashberg formalism, we show that by rubidium (Rb) intercalation, weak superconductivity in bilayer PtTe2 can be significantly boosted with superconducting Tc = 8 K in the presence of spin-orbit coupling (SOC). The intercalant on one hand mediates the interlayer coupling and serves as an electron donor, leading to large density of states at Fermi energy. On the other hand, it increases the mass-enhancement parameter with electron-phonon coupling strength comparable to that of Pt. The potassium intercalated bilayer PtTe2 has a comparable Tc to the case of Rb intercalation. The relatively high Tc with SOC combined with experimental accessible crystal structures suggest that these superconductors are promising platforms to study the novel quantum physics associated with two-dimensional superconductivity, such as the recently proposed type-II Ising superconductivity.

cond-mat.mtrl-sci↗

arXiv2021

Tilting behavior of lamellar ice tip during unidirectional freezing of aqueous solutions

Freezing of ice has been largely reported from many aspects, especially its complex pattern formation. Ice grown from liquid phase is usually characteristic of lamellar morphology which plays a significant role in various domains. However, tilted growth of ice via transition from coplanar to non-coplanar growth in directional solidification has been paid little attention in previous studies and there is misleading explanation of the formation of tilted lamellar ice. Here, we in-situ investigated the variations of tilting behavior of lamellar ice tip under different conditions within a single ice crystal with manipulated orientation via unidirectional freezing of aqueous solutions. It is found that tilted growth of ice tips is sensitive to pulling velocity and solute type. These experimental results reveal intrinsic tilted growth behavior of lamellar ice and enrich our understanding in pattern formation of ice.

cond-mat.soft↗

arXiv2021

A Better Approach to Track the Evolution of Static Code Warnings

Static bug detection tools help developers detect code problems. However, it is known that they remain underutilized due to various reasons. Recent advances to incorporate static bug detectors in modern software development workflows can better motivate developers to fix the reported warnings on the fly. In this paper, we study the effectiveness of the state-of-the-art (SOA) solution in tracking warnings by static bug detectors and propose a better solution based on our analysis of the insufficiencies of the SOA solution. In particular, we examined four large-scale open-source systems and crafted a data set of 3,452 static code warnings by two static bug detectors. We manually uncover the ground-truth evolution status of the selected warnings: persistent, resolved, or newly-introduced. Moreover, upon manual analysis, we identified the critical reasons behind the insufficiencies of the SOA matching algorithm. Finally, we propose a better approach to improve the tracking of static warnings over software development history. Our evaluation shows that our proposed approach provides a significant improvement in the precision of the tracking, i.e., from 66.9% to 90.0%.

cs.SE↗

arXiv2021

Information Bottleneck Theory on Convolutional Neural Networks

Recent years, many researches attempt to open the black box of deep neural networks and propose a various of theories to understand it. Among them, Information Bottleneck (IB) theory claims that there are two distinct phases consisting of fitting phase and compression phase in the course of training. This statement attracts many attentions since its success in explaining the inner behavior of feedforward neural networks. In this paper, we employ IB theory to understand the dynamic behavior of convolutional neural networks (CNNs) and investigate how the fundamental features such as convolutional layer width, kernel size, network depth, pooling layers and multi-fully connected layer have impact on the performance of CNNs. In particular, through a series of experimental analysis on benchmark of MNIST and Fashion-MNIST, we demonstrate that the compression phase is not observed in all these cases. This shows us the CNNs have a rather complicated behavior than feedforward neural networks.

stat.ML↗