arXiv subjects

Pu Wang

Publications and source records attributed to Pu Wang.

At least 91 records · Page 5Linked to original sources

arXiv2023

Skeleton-based Human Action Recognition via Convolutional Neural Networks (CNN)

Recently, there has been a remarkable increase in the interest towards skeleton-based action recognition within the research community, owing to its various advantageous features, including computational efficiency, representative features, and illumination invariance. Despite this, researchers continue to explore and investigate the most optimal way to represent human actions through skeleton representation and the extracted features. As a result, the growth and availability of human action recognition datasets have risen substantially. In addition, deep learning-based algorithms have gained widespread popularity due to the remarkable advancements in various computer vision tasks. Most state-of-the-art contributions in skeleton-based action recognition incorporate a Graph Neural Network (GCN) architecture for representing the human body and extracting features. Our research demonstrates that Convolutional Neural Networks (CNNs) can attain comparable results to GCN, provided that the proper training techniques, augmentations, and optimizers are applied. Our approach has been rigorously validated, and we have achieved a score of 95% on the NTU-60 dataset

cs.CV↗

arXiv2023

A Modular Multi-stage Lightweight Graph Transformer Network for Human Pose and Shape Estimation from 2D Human Pose

In this research, we address the challenge faced by existing deep learning-based human mesh reconstruction methods in balancing accuracy and computational efficiency. These methods typically prioritize accuracy, resulting in large network sizes and excessive computational complexity, which may hinder their practical application in real-world scenarios, such as virtual reality systems. To address this issue, we introduce a modular multi-stage lightweight graph-based transformer network for human pose and shape estimation from 2D human pose, a pose-based human mesh reconstruction approach that prioritizes computational efficiency without sacrificing reconstruction accuracy. Our method consists of a 2D-to-3D lifter module that utilizes graph transformers to analyze structured and implicit joint correlations in 2D human poses, and a mesh regression module that combines the extracted pose features with a mesh template to produce the final human mesh parameters.

cs.CV↗

arXiv2022

GaitMixer: Skeleton-based Gait Representation Learning via Wide-spectrum Multi-axial Mixer

Most existing gait recognition methods are appearance-based, which rely on the silhouettes extracted from the video data of human walking activities. The less-investigated skeleton-based gait recognition methods directly learn the gait dynamics from 2D/3D human skeleton sequences, which are theoretically more robust solutions in the presence of appearance changes caused by clothes, hairstyles, and carrying objects. However, the performance of skeleton-based solutions is still largely behind the appearance-based ones. This paper aims to close such performance gap by proposing a novel network model, GaitMixer, to learn more discriminative gait representation from skeleton sequence data. In particular, GaitMixer follows a heterogeneous multi-axial mixer architecture, which exploits the spatial self-attention mixer followed by the temporal large-kernel convolution mixer to learn rich multi-frequency signals in the gait feature maps. Experiments on the widely used gait database, CASIA-B, demonstrate that GaitMixer outperforms the previous SOTA skeleton-based methods by a large margin while achieving a competitive performance compared with the representative appearance-based solutions. Code will be available at https://github.com/exitudio/gaitmixer

cs.CV↗

arXiv2022

Versatile van der Waals Heterostructures of Gamma-GeSe with h-BN/Graphene/MoS2

Recent discovery of a novel hexagonal phase of GeSe (Gamma-GeSe) has triggered great interests in nanoelectronics applications owing to its electrical conductivity of bulk phase even higher than graphite while its monolayer is a semiconductor. For potential applications, construction of functional two-dimensional (2D) contacts is indispensable. Herein, via first-principles calculations, we propose the design of van der Waals heterostructures (vdWHs) of Gamma-GeSe contacting respectively with graphene, 2D h-BN and MoS2, as representatives of metallic, insulator, and semiconductor partners. Our work shows that the h-BN or graphene layer donates electrons to the Gamma-GeSe layer, resulting in n doping in Gamma-GeSe, while the MoS2 layer accepts electrons from the Gamma-GeSe layer leading to p doping of the latter. The Gamma-GeSe/BN heterostructure has a type-I band alignment with large band offsets, indicating that BN can be used as an effective passivating layer to protect Gamma-GeSe from its environmental disturbance while maintaining its major electronic and optical characteristics. For Gamma-GeSe/graphene heterostructure, it is prone to have a very low-Schottky barrier down to tens of meV, easily overcome by thermal excitation, which can be tunable by strain and external electric field. The Gamma-GeSe/MoS2 vdWH forms a Z-scheme interface, which is beneficial for carriers splitting and photon utilization. Our work indicates that Gamma-GeSe can be well passivated by BN, and form intimate contact with graphene for high charge injection efficiency and with MoS2 for efficient carriers splitting for redox reactions.

cond-mat.mtrl-sci↗

arXiv2022

Highly Modulated Dual Semimetal and Semiconducting Gamma-GeSe with Strain Engineering

Layered hexagonal Gamma--GeSe, a new polymorph of GeSe synthesized recently, shows strikingly high electronic conductivity in its bulk form (even higher than graphite) while semiconducting in the case of monolayer (1L). In this work, by using first-principles calculations, we demonstrate that, different from its orthorhombic phases of GeSe, the Gamma--GeSe shows a small spatial anisotropic dependence and a strikingly thickness-dependent behavior with transition from semimetal (bulk, 0.04 eV) to semiconductor (1L, 0.99 eV), and this dual conducting characteristic realized simply with thickness control in Gamma-GeSe has not been found in other 2D materials before. The lacking of d-orbital allows charge carrier with small effective mass (0.16 m0 for electron and 0.23 m0 for hole) which is comparable to phosphorene. Meanwhile, 1L Gamma--GeSe shows a superior flexibility with Young's modulus of 86.59 N/m, only one-quarter of that of graphene and three-quarters of that of MoS2, and Poisson's ratio of 0.26, suggesting a highly flexible lattice. Interestingly, 1L Gamma-GeSe shows an in-plane isotropic elastic modulus inherent with hexagonal symmetry while an anisotropic in-plane effective mass owing to shifted valleys around the band edges. We demonstrate the feasibility of strain engineering in inducing indirect-direct and semiconductor-metal transitions resulting from competing bands at the band edges. Our work shows that the free 1L Gamma-GeSe shows a strong light absorption (~106 cm-1) and an indirect bandgap with rich valleys at band edges, enabling high carrier concentration and a low rate of direct electron-hole recombination which would be promising for nanoelectronics and solar cell applications.

cond-mat.mtrl-sci↗

arXiv2022

Deep learning on rail profiles matching

Matching the rail cross-section profiles measured on site with the designed profile is a must to evaluate the wear of the rail, which is very important for track maintenance and rail safety. So far, the measured rail profiles to be matched usually have four features, that is, large amount of data, diverse section shapes, hardware made errors, and human experience needs to be introduced to solve the complex situation on site during matching process. However, traditional matching methods based on feature points or feature lines could no longer meet the requirements. To this end, we first establish the rail profiles matching dataset composed of 46386 pairs of professional manual matched data, then propose a general high-precision method for rail profiles matching using pre-trained convolutional neural network (CNN). This new method based on deep learning is promising to be the dominant approach for this issue. Source code is at https://github.com/Kunqi1994/Deep-learning-on-rail-profile-matching.

cs.CV↗

arXiv2022

Quantum Feature Extraction for THz Multi-Layer Imaging

A learning-based THz multi-layer imaging has been recently used for contactless three-dimensional (3D) positioning and encoding. We show a proof-of-concept demonstration of an emerging quantum machine learning (QML) framework to deal with depth variation, shadow effect, and double-sided content recognition, through an experimental validation.

quant-ph↗

arXiv2022

A Lightweight Graph Transformer Network for Human Mesh Reconstruction from 2D Human Pose

Existing deep learning-based human mesh reconstruction approaches have a tendency to build larger networks in order to achieve higher accuracy. Computational complexity and model size are often neglected, despite being key characteristics for practical use of human mesh reconstruction models (e.g. virtual try-on systems). In this paper, we present GTRS, a lightweight pose-based method that can reconstruct human mesh from 2D human pose. We propose a pose analysis module that uses graph transformers to exploit structured and implicit joint correlations, and a mesh regression module that combines the extracted pose feature with the mesh template to reconstruct the final human mesh. We demonstrate the efficiency and generalization of GTRS by extensive evaluations on the Human3.6M and 3DPW datasets. In particular, GTRS achieves better accuracy than the SOTA pose-based method Pose2Mesh while only using 10.2% of the parameters (Params) and 2.5% of the FLOPs on the challenging in-the-wild 3DPW dataset. Code will be publicly available.

cs.CV↗

arXiv2022

Bottleneck Low-rank Transformers for Low-resource Spoken Language Understanding

End-to-end spoken language understanding (SLU) systems benefit from pretraining on large corpora, followed by fine-tuning on application-specific data. The resulting models are too large for on-edge applications. For instance, BERT-based systems contain over 110M parameters. Observing the model is overparameterized, we propose lean transformer structure where the dimension of the attention mechanism is automatically reduced using group sparsity. We propose a variant where the learned attention subspace is transferred to an attention bottleneck layer. In a low-resource setting and without pre-training, the resulting compact SLU model achieves accuracies competitive with pre-trained large models.

cs.CL↗

arXiv2022

EdgeML: Towards Network-Accelerated Federated Learning over Wireless Edge

Federated learning (FL) is a distributed machine learning technology for next-generation AI systems that allows a number of workers, i.e., edge devices, collaboratively learn a shared global model while keeping their data locally to prevent privacy leakage. Enabling FL over wireless multi-hop networks can democratize AI and make it accessible in a cost-effective manner. However, the noisy bandwidth-limited multi-hop wireless connections can lead to delayed and nomadic model updates, which significantly slows down the FL convergence speed. To address such challenges, this paper aims to accelerate FL convergence over wireless edge by optimizing the multi-hop federated networking performance. In particular, the FL convergence optimization problem is formulated as a Markov decision process (MDP). To solve such MDP, multi-agent reinforcement learning (MA-RL) algorithms along with domain-specific action space refining schemes are developed, which online learn the delay-minimum forwarding paths to minimize the model exchange latency between the edge devices (i.e., workers) and the remote server. To validate the proposed solutions, FedEdge is developed and implemented, which is the first experimental framework in the literature for FL over multi-hop wireless edge computing networks. FedEdge allows us to fast prototype, deploy, and evaluate novel FL algorithms along with RL-based system optimization methods in real wireless devices. Moreover, a physical experimental testbed is implemented by customizing the widely adopted Linux wireless routers and ML computing nodes.Finally, our experimentation results on the testbed show that the proposed network-accelerated FL system can practically and significantly improve FL convergence speed, compared to the FL system empowered by the production-grade commercially available wireless networking protocol, BATMAN-Adv.

cs.NI↗

arXiv2022

Quantum Transfer Learning for Wi-Fi Sensing

Beyond data communications, commercial-off-the-shelf Wi-Fi devices can be used to monitor human activities, track device locomotion, and sense the ambient environment. In particular, spatial beam attributes that are inherently available in the 60-GHz IEEE 802.11ad/ay standards have shown to be effective in terms of overhead and channel measurement granularity for these indoor sensing tasks. In this paper, we investigate transfer learning to mitigate domain shift in human monitoring tasks when Wi-Fi settings and environments change over time. As a proof-of-concept study, we consider quantum neural networks (QNN) as well as classical deep neural networks (DNN) for the future quantum-ready society. The effectiveness of both DNN and QNN is validated by an in-house experiment for human pose recognition, achieving greater than 90% accuracy with a limited data size.

cs.LG↗

arXiv2022

AutoQML: Automated Quantum Machine Learning for Wi-Fi Integrated Sensing and Communications

Commercial Wi-Fi devices can be used for integrated sensing and communications (ISAC) to jointly exchange data and monitor indoor environment. In this paper, we investigate a proof-of-concept approach using automated quantum machine learning (AutoQML) framework called AutoAnsatz to recognize human gesture. We address how to efficiently design quantum circuits to configure quantum neural networks (QNN). The effectiveness of AutoQML is validated by an in-house experiment for human pose recognition, achieving state-of-the-art performance greater than 80% accuracy for a limited data size with a significantly small number of trainable parameters.

cs.LG↗

arXiv2022

Local Learning Matters: Rethinking Data Heterogeneity in Federated Learning

Federated learning (FL) is a promising strategy for performing privacy-preserving, distributed learning with a network of clients (i.e., edge devices). However, the data distribution among clients is often non-IID in nature, making efficient optimization difficult. To alleviate this issue, many FL algorithms focus on mitigating the effects of data heterogeneity across clients by introducing a variety of proximal terms, some incurring considerable compute and/or memory overheads, to restrain local updates with respect to the global model. Instead, we consider rethinking solutions to data heterogeneity in FL with a focus on local learning generality rather than proximal restriction. To this end, we first present a systematic study informed by second-order indicators to better understand algorithm effectiveness in FL. Interestingly, we find that standard regularization methods are surprisingly strong performers in mitigating data heterogeneity effects. Based on our findings, we further propose a simple and effective method, FedAlign, to overcome data heterogeneity and the pitfalls of previous methods. FedAlign achieves competitive accuracy with state-of-the-art FL methods across a variety of settings while minimizing computation and memory overhead. Code is available at https://github.com/mmendiet/FedAlign

cs.LG↗

arXiv2022

Exploiting Temporal Relations on Radar Perception for Autonomous Driving

We consider the object recognition problem in autonomous driving using automotive radar sensors. Comparing to Lidar sensors, radar is cost-effective and robust in all-weather conditions for perception in autonomous driving. However, radar signals suffer from low angular resolution and precision in recognizing surrounding objects. To enhance the capacity of automotive radar, in this work, we exploit the temporal information from successive ego-centric bird-eye-view radar image frames for radar object recognition. We leverage the consistency of an object's existence and attributes (size, orientation, etc.), and propose a temporal relational layer to explicitly model the relations between objects within successive radar images. In both object detection and multiple object tracking, we show the superiority of our method compared to several baseline approaches.

cs.CV↗

arXiv2022

Line-field Coherent Sensing with LED Illumination

We describe a method of low-coherence interferometry based optical profilometry using standard light-emitting diode (LED) illumination and complementary metal-oxide-semiconductor (CMOS) image sensors. A line-field illumination strategy allows for the simultaneous measurement of many points in space. Micron scale accuracy and resolution are achieved and demonstrated using a variety of targets.

physics.optics↗

arXiv2022

Oxygen Deficient α-MoO3 with Promoted Adsorption and State-Quenching of H2O for Gas Sensor: A DFT Study

Semiconducting oxides with reducible cations are ideal platforms for various functional applications in nanoelectronics and catalysts. Here we report an ultrathin monolayer alpha-MoO3 where tunable electronic properties and different gas adsorbing behaviors upon introducing the oxygen vacancies (VO). The unique property of alpha-MoO3 is that it contains three different types of oxygen atoms occupying three Wyckoff sites that are absent in other low-dimensional oxides and provides rich electronic hybridized states. The presence of VO triggers intermediate state in the gap at ~0.59 eV below the conduction band minimum and reduces the work function dramatically, together with new excitations at near infrared. The realigned Fermi level associated with the dangling state of VO reduces the neighboring Mo atoms and affects the gas adsorption thereafter. The binding energy of H2O molecules above VO is 2.5 times up to -0.75 eV compared with that of perfect lattice site and trends of transfer of electrons also reverse. The latter is related with the shallow localized state in the band gap due to H2O adsorbed above perfect MoO3 which becomes quenched upon adsorbing at the VO site. Those rich in-gap defective states in oxygen deficient MoO3, broadening the light absorption and promoting the uptake of water, are conductive to the application of alpha-MoO3 for optoelectronics, photothermal therapy, and sensor of moisture.

cond-mat.mtrl-sci↗

arXiv2022

KST-GCN: A Knowledge-Driven Spatial-Temporal Graph Convolutional Network for Traffic Forecasting

While considering the spatial and temporal features of traffic, capturing the impacts of various external factors on travel is an essential step towards achieving accurate traffic forecasting. However, existing studies seldom consider external factors or neglect the effect of the complex correlations among external factors on traffic. Intuitively, knowledge graphs can naturally describe these correlations. Since knowledge graphs and traffic networks are essentially heterogeneous networks, it is challenging to integrate the information in both networks. On this background, this study presents a knowledge representation-driven traffic forecasting method based on spatial-temporal graph convolutional networks. We first construct a knowledge graph for traffic forecasting and derive knowledge representations by a knowledge representation learning method named KR-EAR. Then, we propose the Knowledge Fusion Cell (KF-Cell) to combine the knowledge and traffic features as the input of a spatial-temporal graph convolutional backbone network. Experimental results on the real-world dataset show that our strategy enhances the forecasting performances of backbones at various prediction horizons. The ablation and perturbation analysis further verify the effectiveness and robustness of the proposed method. To the best of our knowledge, this is the first study that constructs and utilizes a knowledge graph to facilitate traffic forecasting; it also offers a promising direction to integrate external information and spatial-temporal information for traffic forecasting. The source code is available at https://github.com/lehaifeng/T-GCN/tree/master/KST-GCN.

cs.LG↗

arXiv2022

From Raman frequency combs to supercontinuum generation in nitrogen-filled hollow-core anti-resonant fiber

We demonstrate a route to supercontinuum generation in gas-filled hollow-core anti-resonant fibers through the creation of a broad vibrational Raman frequency comb followed by continuous broadening and merging of the comb lines through either rotational Raman scattering or the optical Kerr effect. Our demonstration experiments, utilizing a single pump pulse with 20 ps duration at 532 nm in a nitrogen-filled fiber, produce a supercontinuum spanning from 440 nm to 1200 nm, with an additional deep ultraviolet continuum from 250 nm to 360 nm. Numerical results suggest that this approach can produce even broader supercontinuum spectra extending from the ultraviolet to mid-infrared.

physics.optics↗