arXiv subjects

Xin Wen

Publications and source records attributed to Xin Wen.

At least 73 records · Page 4Linked to original sources

arXiv2022

Self-Supervised Visual Representation Learning with Semantic Grouping

In this paper, we tackle the problem of learning visual representations from unlabeled scene-centric data. Existing works have demonstrated the potential of utilizing the underlying complex structure within scene-centric data; still, they commonly rely on hand-crafted objectness priors or specialized pretext tasks to build a learning framework, which may harm generalizability. Instead, we propose contrastive learning from data-driven semantic slots, namely SlotCon, for joint semantic grouping and representation learning. The semantic grouping is performed by assigning pixels to a set of learnable prototypes, which can adapt to each sample by attentive pooling over the feature and form new slots. Based on the learned data-dependent slots, a contrastive objective is employed for representation learning, which enhances the discriminability of features, and conversely facilitates grouping semantically coherent pixels together. Compared with previous efforts, by simultaneously optimizing the two coupled objectives of semantic grouping and contrastive learning, our approach bypasses the disadvantages of hand-crafted priors and is able to learn object/group-level representations from scene-centric images. Experiments show our approach effectively decomposes complex scenes into semantic groups for feature learning and significantly benefits downstream tasks, including object detection, instance segmentation, and semantic segmentation. Code is available at: https://github.com/CVMI-Lab/SlotCon.

cs.CV↗

arXiv2022

KTNet: Knowledge Transfer for Unpaired 3D Shape Completion

Unpaired 3D object completion aims to predict a complete 3D shape from an incomplete input without knowing the correspondence between the complete and incomplete shapes. In this paper, we propose the novel KTNet to solve this task from the new perspective of knowledge transfer. KTNet elaborates a teacher-assistant-student network to establish multiple knowledge transfer processes. Specifically, the teacher network takes complete shape as input and learns the knowledge of complete shape. The student network takes the incomplete one as input and restores the corresponding complete shape. And the assistant modules not only help to transfer the knowledge of complete shape from the teacher to the student, but also judge the learning effect of the student network. As a result, KTNet makes use of a more comprehensive understanding to establish the geometric correspondence between complete and incomplete shapes in a perspective of knowledge transfer, which enables more detailed geometric inference for generating high-quality complete shapes. We conduct comprehensive experiments on several datasets, and the results show that our method outperforms previous methods of unpaired point cloud completion by a large margin.

cs.CV↗

arXiv2022

Quantum support vector machines for aerodynamic classification

Aerodynamics plays an important role in aviation industry and aircraft design. Detecting and minimizing the phenomenon of flow separation from scattered pressure data on airfoil is critical for ensuring stable and efficient aviation. However, since it is challenging to understand the mechanics of flow field separation, the aerodynamic parameters are emphasized for the identification and control of flow separation. It has been investigated extensively using traditional algorithms and machine learning methods such as the support vector machine (SVM) models. Recently, a growing interest in quantum computing and its applications among wide research communities sheds light upon the use of quantum techniques to solve aerodynamic problems. In this paper, we apply qSVM, a quantum SVM algorithm based on the quantum annealing model, to identify whether there is flow separation, with their performance in comparison to the widely-used classical SVM. We show that our approach outperforms the classical SVM with an 11.1% increase of the accuracy, from 0.818 to 0.909, for this binary classification task. We further develop multi-class qSVMs based on one-against-all algorithm. We apply it to classify multiple types of the attack angles to the wings, where the advantage over the classical multi-class counterpart is maintained with an accuracy increased from 0.67 to 0.79, by 17.9%. Our work demonstrates a useful quantum technique for classifying flow separation scenarios, and may promote rich investigations for quantum computing applications in fluid dynamics.

quant-ph↗

arXiv2022

Learning Deep Implicit Functions for 3D Shapes with Dynamic Code Clouds

Deep Implicit Function (DIF) has gained popularity as an efficient 3D shape representation. To capture geometry details, current methods usually learn DIF using local latent codes, which discretize the space into a regular 3D grid (or octree) and store local codes in grid points (or octree nodes). Given a query point, the local feature is computed by interpolating its neighboring local codes with their positions. However, the local codes are constrained at discrete and regular positions like grid points, which makes the code positions difficult to be optimized and limits their representation ability. To solve this problem, we propose to learn DIF with Dynamic Code Cloud, named DCC-DIF. Our method explicitly associates local codes with learnable position vectors, and the position vectors are continuous and can be dynamically optimized, which improves the representation ability. In addition, we propose a novel code position loss to optimize the code positions, which heuristically guides more local codes to be distributed around complex geometric details. In contrast to previous methods, our DCC-DIF represents 3D shapes more efficiently with a small amount of local codes, and improves the reconstruction quality. Experiments demonstrate that DCC-DIF achieves better performance over previous methods. Code and data are available at https://github.com/lity20/DCCDIF.

cs.CV↗

arXiv2022

3D Shape Reconstruction from 2D Images with Disentangled Attribute Flow

Reconstructing 3D shape from a single 2D image is a challenging task, which needs to estimate the detailed 3D structures based on the semantic attributes from 2D image. So far, most of the previous methods still struggle to extract semantic attributes for 3D reconstruction task. Since the semantic attributes of a single image are usually implicit and entangled with each other, it is still challenging to reconstruct 3D shape with detailed semantic structures represented by the input image. To address this problem, we propose 3DAttriFlow to disentangle and extract semantic attributes through different semantic levels in the input images. These disentangled semantic attributes will be integrated into the 3D shape reconstruction process, which can provide definite guidance to the reconstruction of specific attribute on 3D shape. As a result, the 3D decoder can explicitly capture high-level semantic features at the bottom of the network, and utilize low-level features at the top of the network, which allows to reconstruct more accurate 3D shapes. Note that the explicit disentangling is learned without extra labels, where the only supervision used in our training is the input image and its corresponding 3D shape. Our comprehensive experiments on ShapeNet dataset demonstrate that 3DAttriFlow outperforms the state-of-the-art shape reconstruction methods, and we also validate its generalization ability on shape completion task.

cs.CV↗

arXiv2022

PMP-Net++: Point Cloud Completion by Transformer-Enhanced Multi-step Point Moving Paths

Point cloud completion concerns to predict missing part for incomplete 3D shapes. A common strategy is to generate complete shape according to incomplete input. However, unordered nature of point clouds will degrade generation of high-quality 3D shapes, as detailed topology and structure of unordered points are hard to be captured during the generative process using an extracted latent code. We address this problem by formulating completion as point cloud deformation process. Specifically, we design a novel neural network, named PMP-Net++, to mimic behavior of an earth mover. It moves each point of incomplete input to obtain a complete point cloud, where total distance of point moving paths (PMPs) should be the shortest. Therefore, PMP-Net++ predicts unique PMP for each point according to constraint of point moving distances. The network learns a strict and unique correspondence on point-level, and thus improves quality of predicted complete shape. Moreover, since moving points heavily relies on per-point features learned by network, we further introduce a transformer-enhanced representation learning network, which significantly improves completion performance of PMP-Net++. We conduct comprehensive experiments in shape completion, and further explore application on point cloud up-sampling, which demonstrate non-trivial improvement of PMP-Net++ over state-of-the-art point cloud completion/up-sampling methods.

cs.CV↗

arXiv2021

Multi-View Partial (MVP) Point Cloud Challenge 2021 on Completion and Registration: Methods and Results

As real-scanned point clouds are mostly partial due to occlusions and viewpoints, reconstructing complete 3D shapes based on incomplete observations becomes a fundamental problem for computer vision. With a single incomplete point cloud, it becomes the partial point cloud completion problem. Given multiple different observations, 3D reconstruction can be addressed by performing partial-to-partial point cloud registration. Recently, a large-scale Multi-View Partial (MVP) point cloud dataset has been released, which consists of over 100,000 high-quality virtual-scanned partial point clouds. Based on the MVP dataset, this paper reports methods and results in the Multi-View Partial Point Cloud Challenge 2021 on Completion and Registration. In total, 128 participants registered for the competition, and 31 teams made valid submissions. The top-ranked solutions will be analyzed, and then we will discuss future research directions.

cs.CV↗

arXiv2021

Quantum-enhanced rubidium atomic magnetometer based on Faraday rotation via 795-nm Stokes operator squeezed light

With the help of Stokes operator S2 squeezed state (also called polarization squeezed state (PSS)) of 795-nm light, rubidium-87 (87Rb) atomic magnetometer based on Faraday rotation has been implemented and characterized.The PSS of Stokes operator S2 of 795-nm light has been prepared by means of coherently combining the polarization coherent state (PCS) of a linearly p-polarized bright 795-nm light beam and a linearly s-polarized squeezed vacuum state (SVS) generated by a 397.5-nm ultraviolet laser pumped sub-threshold optical parametric oscillator (OPO) with a PPKTP bulk crystal inside the OPO cavity.PSS with a squeezing level of -3.7 has been achieved around the analysis frequency of 10 kHz. At different transitions of D1 line, various frequency detuning, and reasonable atomic vapor cells temperature, Faraday rotation has been measured and compared.To decrease absorption (scattering) losses and the back-action from atomic spin noise to the probe beams polarization noise for maintaining the quantum properties of PSS of Stokes operator S2 of 795-nm light, we had to run our magnetometer with 87Rb vapor cells temperature below 60, at which the PSS was almost destroyed.The sensitivities of magnetic field measurement were characterized via measuring signal-to-noise ratio of the alternating current (AC) calibrated magnetic field signal with a balanced polarimeter. Under the conditions of the atomic number density of 5.8*1010 /cm3 and the probe beam with a detuning of - 400 MHz relative to the 5S1/2 (Fg=2) - 5P1/2 (Fe=1) transition of 87Rb D1 line, a typical sensitivity of 19.5 pT/Hz1/2 has been achieved employing PSS of Stokes operator S2 as the probe, compared with a sensitivity of 28.3 pT/Hz1/2 using PCS as the probe.We preliminarily demonstrated that the quantum-enhanced sensitivity in a Faraday-rotation-based 87Rb atomic magnetometer with the help of PSS of 795-nm light.

quant-ph↗

arXiv2021

Enhancement of spin noise spectroscopy of rubidium atomic ensemble by using of the polarization squeezed light

We measured the spin noise spectroscopy (SNS) of rubidium atomic ensemble with two different atomic vapor cells (filled with the buffer gases or coated with paraffin film on the inner wall), and demonstrated the enhancement of signal to noise ratio (SNR) by using of the polarization squeezed state (PSS) of 795 nm light field with Stokes operator S2 squeezed. PSS is prepared by locking the relative phase between the squeezed vacuum state of light obtained by a sub-threshold optical parametric oscillator and the orthogonal polarized local oscillator beam by means of the quantum noise lock. Under the same conditions, PSS can be employed not only to improve SNR, but also to keep the full width at half maximum (FWHM) of SNS unchanged, compared with the case of using polarization coherent state (PCS), and the enhancement of SNR is positively correlated with the squeezing level of PSS. With the increase of probe laser power and atomic number density, the SNR and FWHM of SNS will increase correspondingly. With the help of PSS of Stokes operator S2, quantum enhancement of both SNR and FWHM of SNS signal has been demonstrated by controlling optical power of the S2 polarization squeezed light beam or atomic number density in our experiments.

quant-ph↗

arXiv2021

SnowflakeNet: Point Cloud Completion by Snowflake Point Deconvolution with Skip-Transformer

Point cloud completion aims to predict a complete shape in high accuracy from its partial observation. However, previous methods usually suffered from discrete nature of point cloud and unstructured prediction of points in local regions, which makes it hard to reveal fine local geometric details on the complete shape. To resolve this issue, we propose SnowflakeNet with Snowflake Point Deconvolution (SPD) to generate the complete point clouds. The SnowflakeNet models the generation of complete point clouds as the snowflake-like growth of points in 3D space, where the child points are progressively generated by splitting their parent points after each SPD. Our insight of revealing detailed geometry is to introduce skip-transformer in SPD to learn point splitting patterns which can fit local regions the best. Skip-transformer leverages attention mechanism to summarize the splitting patterns used in the previous SPD layer to produce the splitting in the current SPD layer. The locally compact and structured point cloud generated by SPD is able to precisely capture the structure characteristic of 3D shape in local patches, which enables the network to predict highly detailed geometries, such as smooth regions, sharp edges and corners. Our experimental results outperform the state-of-the-art point cloud completion methods under widely used benchmarks. Code will be available at https://github.com/AllenXiangX/SnowflakeNet.

cs.CV↗

arXiv2021

Irreducible modules of $\frak {sl}_{mp}$ in characteristic $p$ with regular or subregular nilpotent $p$-character

Let $\bbk$ be an algebraically closed field of prime characteristic $p$. If $p$ does not divide $n$, irreducible modules over $\frak {sl}_n$ for regular and subregular nilpotent representations have already known(see \cite{Jan2} and \cite{Jan3}). In this article, we investigate the question when $p$ divides $n$, and precisely describe simple modules of $\frak {sl}_n$ for regular and subregular nilpotent representations.

math.RT↗

arXiv2021

Performance Prediction of InP/GaAsSb Double Heterojunction Bipolar Transistors for THz applications

The intrinsic performance of "type-II" InP/GaAsSb double heterojunction bipolar transistors (DHBTs) towards and beyond THz is predicted and analyzed based on a multi-scale technology computer aided design (TCAD) modeling platform calibrated against experimental measurements. Two-dimensional hydrodynamic simulations are combined with 1-D full-band, atomistic quantum transport calculations to shed light on future DHBT generations whose dimensions are decreased step-by-step, starting from the current device configuration. Simulations predict that a peak transit frequency $f_{T,peak}$ of around 1.6 THz could be reached in aggressively scaled type-II DHBTs with a total thickness of 256 nm and an emitter width $W_E$ of 37.5 nm. The corresponding breakdown voltage $BV_{CEO}$ is estimated to be 2.2 V. The investigations are put in perspective with two DHBT performance limiting factors, self-heating and breakdown characteristics.

physics.app-ph↗

arXiv2021

PMP-Net: Point Cloud Completion by Learning Multi-step Point Moving Paths

The task of point cloud completion aims to predict the missing part for an incomplete 3D shape. A widely used strategy is to generate a complete point cloud from the incomplete one. However, the unordered nature of point clouds will degrade the generation of high-quality 3D shapes, as the detailed topology and structure of discrete points are hard to be captured by the generative process only using a latent code. In this paper, we address the above problem by reconsidering the completion task from a new perspective, where we formulate the prediction as a point cloud deformation process. Specifically, we design a novel neural network, named PMP-Net, to mimic the behavior of an earth mover. It moves each point of the incomplete input to complete the point cloud, where the total distance of point moving paths (PMP) should be shortest. Therefore, PMP-Net predicts a unique point moving path for each point according to the constraint of total point moving distances. As a result, the network learns a strict and unique correspondence on point-level, which can capture the detailed topology and structure relationships between the incomplete shape and the complete target, and thus improves the quality of the predicted complete shape. We conduct comprehensive experiments on Completion3D and PCN datasets, which demonstrate our advantages over the state-of-the-art point cloud completion methods.

cs.CV↗

arXiv2021

Cycle4Completion: Unpaired Point Cloud Completion using Cycle Transformation with Missing Region Coding

In this paper, we present a novel unpaired point cloud completion network, named Cycle4Completion, to infer the complete geometries from a partial 3D object. Previous unpaired completion methods merely focus on the learning of geometric correspondence from incomplete shapes to complete shapes, and ignore the learning in the reverse direction, which makes them suffer from low completion accuracy due to the limited 3D shape understanding ability. To address this problem, we propose two simultaneous cycle transformations between the latent spaces of complete shapes and incomplete ones. The insight of cycle transformation is to promote networks to understand 3D shapes by learning to generate complete or incomplete shapes from their complementary ones. Specifically, the first cycle transforms shapes from incomplete domain to complete domain, and then projects them back to the incomplete domain. This process learns the geometric characteristic of complete shapes, and maintains the shape consistency between the complete prediction and the incomplete input. Similarly, the inverse cycle transformation starts from complete domain to incomplete domain, and goes back to complete domain to learn the characteristic of incomplete shapes. We provide a comprehensive evaluation in experiments, which shows that our model with the learned bidirectional geometry correspondence outperforms state-of-the-art unpaired completion methods.

cs.CV↗

arXiv2020

Temporal Context Aggregation for Video Retrieval with Contrastive Learning

The current research focus on Content-Based Video Retrieval requires higher-level video representation describing the long-range semantic dependencies of relevant incidents, events, etc. However, existing methods commonly process the frames of a video as individual images or short clips, making the modeling of long-range semantic dependencies difficult. In this paper, we propose TCA (Temporal Context Aggregation for Video Retrieval), a video representation learning framework that incorporates long-range temporal information between frame-level features using the self-attention mechanism. To train it on video retrieval datasets, we propose a supervised contrastive learning method that performs automatic hard negative mining and utilizes the memory bank mechanism to increase the capacity of negative samples. Extensive experiments are conducted on multiple video retrieval tasks, such as CC_WEB_VIDEO, FIVR-200K, and EVVE. The proposed method shows a significant performance advantage (~17% mAP on FIVR-200K) over state-of-the-art methods with video-level features, and deliver competitive results with 22x faster inference time comparing with frame-level features.

cs.CV↗

arXiv2020

Distilling Visual Priors from Self-Supervised Learning

Convolutional Neural Networks (CNNs) are prone to overfit small training datasets. We present a novel two-phase pipeline that leverages self-supervised learning and knowledge distillation to improve the generalization ability of CNN models for image classification under the data-deficient setting. The first phase is to learn a teacher model which possesses rich and generalizable visual representations via self-supervised learning, and the second phase is to distill the representations into a student model in a self-distillation manner, and meanwhile fine-tune the student model for the image classification task. We also propose a novel margin loss for the self-supervised contrastive learning proxy task to better learn the representation under the data-deficient scenario. Together with other tricks, we achieve competitive performance in the VIPriors image classification challenge.

cs.CV↗

arXiv2020

Stochastic switching in Rydberg atomic ensemble

We demonstrated stochastic switching in a bistable system implemented with Rydberg atomic ensemble. The transition between the two states of the bistable system is driven by intensity noise of the laser beams. Rydberg atomic ensemble accumulates energy in an equilibrium situation and brings the nonlinear system across the threshold, where stochastic switching occurs between the two states. Measurement of Rydberg state's population by means of the ladder-type electromagnetically-induced transparency approaching relevant Rydberg state allows us to investigate the nonlinear behavior in Rydberg atomic ensemble based bistable system experimentally.

physics.atom-ph↗

arXiv2020

Point Cloud Completion by Skip-attention Network with Hierarchical Folding

Point cloud completion aims to infer the complete geometries for missing regions of 3D objects from incomplete ones. Previous methods usually predict the complete point cloud based on the global shape representation extracted from the incomplete input. However, the global representation often suffers from the information loss of structure details on local regions of incomplete point cloud. To address this problem, we propose Skip-Attention Network (SA-Net) for 3D point cloud completion. Our main contributions lie in the following two-folds. First, we propose a skip-attention mechanism to effectively exploit the local structure details of incomplete point clouds during the inference of missing parts. The skip-attention mechanism selectively conveys geometric information from the local regions of incomplete point clouds for the generation of complete ones at different resolutions, where the skip-attention reveals the completion process in an interpretable way. Second, in order to fully utilize the selected geometric information encoded by skip-attention mechanism at different resolutions, we propose a novel structure-preserving decoder with hierarchical folding for complete shape generation. The hierarchical folding preserves the structure of complete point cloud generated in upper layer by progressively detailing the local regions, using the skip-attentioned geometry at the same resolution. We conduct comprehensive experiments on ShapeNet and KITTI datasets, which demonstrate that the proposed SA-Net outperforms the state-of-the-art point cloud completion methods.

cs.CV↗