arXiv subjects

Yuan He

Publications and source records attributed to Yuan He.

At least 127 records · Page 7Linked to original sources

arXiv2022

D^2ETR: Decoder-Only DETR with Computationally Efficient Cross-Scale Attention

DETR is the first fully end-to-end detector that predicts a final set of predictions without post-processing. However, it suffers from problems such as low performance and slow convergence. A series of works aim to tackle these issues in different ways, but the computational cost is yet expensive due to the sophisticated encoder-decoder architecture. To alleviate this issue, we propose a decoder-only detector called D^2ETR. In the absence of encoder, the decoder directly attends to the fine-fused feature maps generated by the Transformer backbone with a novel computationally efficient cross-scale attention module. D^2ETR demonstrates low computational complexity and high detection accuracy in evaluations on the COCO benchmark, outperforming DETR and its variants.

cs.CV↗

arXiv2022

High precision measurement of cosmic curvature: from gravitational waves and cosmic chronometer

Although the spatial curvature has been measured with very high precision, it still suffers from the well known cosmic curvature tension. In this paper, we propose an improved method to determine the cosmic curvature, by using the simulated data of binary neutron star mergers observed by the second generation space-based DECi-hertz Interferometer Gravitational-wave Observatory (DECIGO). By applying the Hubble parameter observations of cosmic chronometers to the DECIGO standard sirens, we explore different possibilities of making measurements of the cosmic curvature referring to a distant past: one is to reconstruct the Hubble parameters through the Gaussian process without the influence of hypothetical models, and the other is deriving constraints on $Ω_K$ in the framework of non-flat $Λ$ cold dark matter model. It is shown that in the improved method DECIGO could provide a reliable and stringent constraint on the cosmic curvature ($Ω_{K} = -0.007\pm0.016$), while we could only expect the zero cosmic curvature to be established at the precision of $ΔΩ_K=0.12$ in the second model-dependent method. Therefore, our results indicate that in the framework of methodology proposed in this paper, the increasing number of well-measured standard sirens in DECIGO could significantly reduce the bias of estimations for cosmic curvature. Such constraint is also comparable to the precision of Planck 2018 results with the newest cosmic microwave background (CMB) observations ($ΔΩ_{K} \approx 0.018$), based on the concordance $Λ$CDM model.

astro-ph.CO↗

arXiv2022

$\rm ^{83}Rb$/$\rm ^{83m}Kr$ production and cross-section measurement with 3.4 MeV and 20 MeV proton beams

$\rm ^{83m}Kr$, with a short lifetime, is an ideal calibration source for liquid xenon or liquid argon detectors. The $\rm ^{83m}Kr$ isomer can be generated through the decay of $\rm ^{83} Rb$ isotope which is usually produced by proton beams bombarding natural krypton atoms. In this paper, we report a successful production of $\rm ^{83}Rb/^{83m}Kr$ with a proton beam energy of 3.4 MeV, and the first measurement of the production rate with such low energy proton beams. Another production attempt is performed using the newly available 20 MeV proton beam in China, and the measured production rate is consistent with previous measurements. The produced $\rm ^{83m}Kr$ source has been successfully injected into the PandaX-II liquid xenon detector, yielding enough statistics for detector calibration.

nucl-ex↗

arXiv2021

Monocular 3D Object Detection: An Extrinsic Parameter Free Approach

Monocular 3D object detection is an important task in autonomous driving. It can be easily intractable where there exists ego-car pose change w.r.t. ground plane. This is common due to the slight fluctuation of road smoothness and slope. Due to the lack of insight in industrial application, existing methods on open datasets neglect the camera pose information, which inevitably results in the detector being susceptible to camera extrinsic parameters. The perturbation of objects is very popular in most autonomous driving cases for industrial products. To this end, we propose a novel method to capture camera pose to formulate the detector free from extrinsic perturbation. Specifically, the proposed framework predicts camera extrinsic parameters by detecting vanishing point and horizon change. A converter is designed to rectify perturbative features in the latent space. By doing so, our 3D detector works independent of the extrinsic parameter variations and produces accurate results in realistic cases, e.g., potholed and uneven roads, where almost all existing monocular detectors fail to handle. Experiments demonstrate our method yields the best performance compared with the other state-of-the-arts by a large margin on both KITTI 3D and nuScenes datasets.

cs.CV↗

arXiv2021

Unrestricted Adversarial Attacks on ImageNet Competition

Many works have investigated the adversarial attacks or defenses under the settings where a bounded and imperceptible perturbation can be added to the input. However in the real-world, the attacker does not need to comply with this restriction. In fact, more threats to the deep model come from unrestricted adversarial examples, that is, the attacker makes large and visible modifications on the image, which causes the model classifying mistakenly, but does not affect the normal observation in human perspective. Unrestricted adversarial attack is a popular and practical direction but has not been studied thoroughly. We organize this competition with the purpose of exploring more effective unrestricted adversarial attack algorithm, so as to accelerate the academical research on the model robustness under stronger unbounded attacks. The competition is held on the TianChi platform (\url{https://tianchi.aliyun.com/competition/entrance/531853/introduction}) as one of the series of AI Security Challengers Program.

cs.CV↗

arXiv2021

Adversarial Attacks on ML Defense Models Competition

Due to the vulnerability of deep neural networks (DNNs) to adversarial examples, a large number of defense techniques have been proposed to alleviate this problem in recent years. However, the progress of building more robust models is usually hampered by the incomplete or incorrect robustness evaluation. To accelerate the research on reliable evaluation of adversarial robustness of the current defense models in image classification, the TSAIL group at Tsinghua University and the Alibaba Security group organized this competition along with a CVPR 2021 workshop on adversarial machine learning (https://aisecure-workshop.github.io/amlcvpr2021/). The purpose of this competition is to motivate novel attack algorithms to evaluate adversarial robustness more effectively and reliably. The participants were encouraged to develop stronger white-box attack algorithms to find the worst-case robustness of different defenses. This competition was conducted on an adversarial robustness evaluation platform -- ARES (https://github.com/thu-ml/ares), and is held on the TianChi platform (https://tianchi.aliyun.com/competition/entrance/531847/introduction) as one of the series of AI Security Challengers Program. After the competition, we summarized the results and established a new adversarial robustness benchmark at https://ml.cs.tsinghua.edu.cn/ares-bench/, which allows users to upload adversarial attack algorithms and defense models for evaluation.

cs.CV↗

arXiv2021

Fine-Grained Fashion Similarity Prediction by Attribute-Specific Embedding Learning

This paper strives to predict fine-grained fashion similarity. In this similarity paradigm, one should pay more attention to the similarity in terms of a specific design/attribute between fashion items. For example, whether the collar designs of the two clothes are similar. It has potential value in many fashion related applications, such as fashion copyright protection. To this end, we propose an Attribute-Specific Embedding Network (ASEN) to jointly learn multiple attribute-specific embeddings, thus measure the fine-grained similarity in the corresponding space. The proposed ASEN is comprised of a global branch and a local branch. The global branch takes the whole image as input to extract features from a global perspective, while the local branch takes as input the zoomed-in region-of-interest (RoI) w.r.t. the specified attribute thus able to extract more fine-grained features. As the global branch and the local branch extract the features from different perspectives, they are complementary to each other. Additionally, in each branch, two attention modules, i.e., Attribute-aware Spatial Attention and Attribute-aware Channel Attention, are integrated to make ASEN be able to locate the related regions and capture the essential patterns under the guidance of the specified attribute, thus make the learned attribute-specific embeddings better reflect the fine-grained similarity. Extensive experiments on three fashion-related datasets, i.e., FashionAI, DARN, and DeepFashion, show the effectiveness of ASEN for fine-grained fashion similarity prediction and its potential for fashion reranking. Code and data are available at https://github.com/maryeon/asenpp .

cs.CV↗

arXiv2021

Prior-Guided Deep Interference Mitigation for FMCW Radars

A prior-guided deep learning (DL) based interference mitigation approach is proposed for frequency modulated continuous wave (FMCW) radars. In this paper, the interference mitigation problem is tackled as a regression problem. Considering the complex-valued nature of radar signals, the complex-valued convolutional neural network is utilized as an architecture for implementation, which is different from the conventional real-valued counterparts. Meanwhile, as the useful beat signals of FMCW radars and interferences exhibit different distributions in the time-frequency domain, this prior feature is exploited as a regularization term to avoid overfitting of the learned representation. The effectiveness and accuracy of our proposed complex-valued fully convolutional network (CV-FCN) based interference mitigation approach are verified and analyzed through both simulated and measured radar signals. Compared to the real-valued counterparts, the CV-FCN shows a better interference mitigation performance with a potential of half memory reduction in low Signal to Interference plus Noise Ratio (SINR) scenarios. Moreover, the CV-FCN trained using only simulated data can be directly utilized for interference mitigation in various measured radar signals and shows a superior generalization capability. Furthermore, by incorporating the prior feature, the CV-FCN trained on only 1/8 of the full data achieves comparable performance as that on the full dataset in low SINR scenarios, and the training procedure converges faster.

eess.SP↗

arXiv2021

AdvDrop: Adversarial Attack to DNNs by Dropping Information

Human can easily recognize visual objects with lost information: even losing most details with only contour reserved, e.g. cartoon. However, in terms of visual perception of Deep Neural Networks (DNNs), the ability for recognizing abstract objects (visual objects with lost information) is still a challenge. In this work, we investigate this issue from an adversarial viewpoint: will the performance of DNNs decrease even for the images only losing a little information? Towards this end, we propose a novel adversarial attack, named \textit{AdvDrop}, which crafts adversarial examples by dropping existing information of images. Previously, most adversarial attacks add extra disturbing information on clean images explicitly. Opposite to previous works, our proposed work explores the adversarial robustness of DNN models in a novel perspective by dropping imperceptible details to craft adversarial examples. We demonstrate the effectiveness of \textit{AdvDrop} by extensive experiments, and show that this new type of adversarial examples is more difficult to be defended by current defense systems.

cs.CV↗

arXiv2021

A new way to test the WIMP dark matter models

In this paper, we investigate the possibility of testing the weakly interacting massive particle (WIMP) dark matter (DM) models by applying the simplest phenomenological model which introduces an interaction term between dark energy (DE) and WIMP DM, i.e., $Q = 3γ_{DM} Hρ_{DM}$. In general, the coupling strength $γ_{DE}$ is close to $0$ as the interaction between DE and WIMP DM is very weak, thus the effect of $γ_ {DE}$ on the evolution of $Y$ associated with DM energy density can be safely neglected. Meanwhile, our numerical calculation also indicates that $x_f\approx20$ is associated with DM freeze-out temperature, which is the same as the vanishing interaction scenario. As for DM relic density, it will be magnified by $\frac{2-3γ_{DM}}{2}[{2πg_* m_{DM}^3}/{(45 s_0 x_f^3})]^{γ_{DM}}$ times, which provides a new way to test WIMP DM models. As an example, we analyze the case in which WIMP DM is a scalar DM. (SGL+SNe+Hz) and (CMB+BAO+SNe) cosmological observations will give $γ_{DM}=0.134^{+0.17}_{-0.069}$ and $γ_{DM}=-0.0008\pm0.0016$, respectively. After further considering the constraints from DM direct detection experiment, DM indirect detection experiment, and DM relic density, we find that the allowed parameter space of the scalar DM model will be completely excluded for the former cosmological observations, while it will increase for the latter ones. Those two cosmological observations lead to an almost paradoxical conclusion. Therefore, one could expect more stringent constraints on the WMIP DM models, with the accumulation of more accurate cosmological observations in the near future.

astro-ph.CO↗

arXiv2021

DRDF: Determining the Importance of Different Multimodal Information with Dual-Router Dynamic Framework

In multimodal tasks, we find that the importance of text and image modal information is different for different input cases, and for this motivation, we propose a high-performance and highly general Dual-Router Dynamic Framework (DRDF), consisting of Dual-Router, MWF-Layer, experts and expert fusion unit. The text router and image router in Dual-Router accept text modal information and image modal information, and use MWF-Layer to determine the importance of modal information. Based on the result of the determination, MWF-Layer generates fused weights for the fusion of experts. Experts are model backbones that match the current task. DRDF has high performance and high generality, and we have tested 12 backbones such as Visual BERT on multimodal dataset Hateful memes, unimodal dataset CIFAR10, CIFAR100, and TinyImagenet. Our DRDF outperforms all the baselines. We also verified the components of DRDF in detail by ablations, compared and discussed the reasons and ideas of DRDF design.

cs.CV↗

arXiv2021

RAMS-Trans: Recurrent Attention Multi-scale Transformer forFine-grained Image Recognition

In fine-grained image recognition (FGIR), the localization and amplification of region attention is an important factor, which has been explored a lot by convolutional neural networks (CNNs) based approaches. The recently developed vision transformer (ViT) has achieved promising results on computer vision tasks. Compared with CNNs, Image sequentialization is a brand new manner. However, ViT is limited in its receptive field size and thus lacks local attention like CNNs due to the fixed size of its patches, and is unable to generate multi-scale features to learn discriminative region attention. To facilitate the learning of discriminative region attention without box/part annotations, we use the strength of the attention weights to measure the importance of the patch tokens corresponding to the raw images. We propose the recurrent attention multi-scale transformer (RAMS-Trans), which uses the transformer's self-attention to recursively learn discriminative region attention in a multi-scale manner. Specifically, at the core of our approach lies the dynamic patch proposal module (DPPM) guided region amplification to complete the integration of multi-scale image patches. The DPPM starts with the full-size image patches and iteratively scales up the region attention to generate new patches from global to local by the intensity of the attention weights generated at each scale as an indicator. Our approach requires only the attention weights that come with ViT itself and can be easily trained end-to-end. Extensive experiments demonstrate that RAMS-Trans performs better than concurrent works, in addition to efficient CNN models, achieving state-of-the-art results on three benchmark datasets.

cs.CV↗

arXiv2021

QAIR: Practical Query-efficient Black-Box Attacks for Image Retrieval

We study the query-based attack against image retrieval to evaluate its robustness against adversarial examples under the black-box setting, where the adversary only has query access to the top-k ranked unlabeled images from the database. Compared with query attacks in image classification, which produce adversaries according to the returned labels or confidence score, the challenge becomes even more prominent due to the difficulty in quantifying the attack effectiveness on the partial retrieved list. In this paper, we make the first attempt in Query-based Attack against Image Retrieval (QAIR), to completely subvert the top-k retrieval results. Specifically, a new relevance-based loss is designed to quantify the attack effects by measuring the set similarity on the top-k retrieval results before and after attacks and guide the gradient optimization. To further boost the attack efficiency, a recursive model stealing method is proposed to acquire transferable priors on the target model and generate the prior-guided gradients. Comprehensive experiments show that the proposed attack achieves a high attack success rate with few queries against the image retrieval systems under the black-box setting. The attack evaluations on the real-world visual search engine show that it successfully deceives a commercial system such as Bing Visual Search with 98% attack success rate by only 33 queries on average.

cs.CV↗

arXiv2021

Seeking the Shape of Sound: An Adaptive Framework for Learning Voice-Face Association

Nowadays, we have witnessed the early progress on learning the association between voice and face automatically, which brings a new wave of studies to the computer vision community. However, most of the prior arts along this line (a) merely adopt local information to perform modality alignment and (b) ignore the diversity of learning difficulty across different subjects. In this paper, we propose a novel framework to jointly address the above-mentioned issues. Targeting at (a), we propose a two-level modality alignment loss where both global and local information are considered. Compared with the existing methods, we introduce a global loss into the modality alignment process. The global component of the loss is driven by the identity classification. Theoretically, we show that minimizing the loss could maximize the distance between embeddings across different identities while minimizing the distance between embeddings belonging to the same identity, in a global sense (instead of a mini-batch). Targeting at (b), we propose a dynamic reweighting scheme to better explore the hard but valuable identities while filtering out the unlearnable identities. Experiments show that the proposed method outperforms the previous methods in multiple settings, including voice-face matching, verification and retrieval.

cs.CV↗

arXiv2021

Adversarial Laser Beam: Effective Physical-World Attack to DNNs in a Blink

Though it is well known that the performance of deep neural networks (DNNs) degrades under certain light conditions, there exists no study on the threats of light beams emitted from some physical source as adversarial attacker on DNNs in a real-world scenario. In this work, we show by simply using a laser beam that DNNs are easily fooled. To this end, we propose a novel attack method called Adversarial Laser Beam ($AdvLB$), which enables manipulation of laser beam's physical parameters to perform adversarial attack. Experiments demonstrate the effectiveness of our proposed approach in both digital- and physical-settings. We further empirically analyze the evaluation results and reveal that the proposed laser beam attack may lead to some interesting prediction errors of the state-of-the-art DNNs. We envisage that the proposed $AdvLB$ method enriches the current family of adversarial attacks and builds the foundation for future robustness studies for light.

cs.LG↗

arXiv2021

Spatial-Phase Shallow Learning: Rethinking Face Forgery Detection in Frequency Domain

The remarkable success in face forgery techniques has received considerable attention in computer vision due to security concerns. We observe that up-sampling is a necessary step of most face forgery techniques, and cumulative up-sampling will result in obvious changes in the frequency domain, especially in the phase spectrum. According to the property of natural images, the phase spectrum preserves abundant frequency components that provide extra information and complement the loss of the amplitude spectrum. To this end, we present a novel Spatial-Phase Shallow Learning (SPSL) method, which combines spatial image and phase spectrum to capture the up-sampling artifacts of face forgery to improve the transferability, for face forgery detection. And we also theoretically analyze the validity of utilizing the phase spectrum. Moreover, we notice that local texture information is more crucial than high-level semantic information for the face forgery detection task. So we reduce the receptive fields by shallowing the network to suppress high-level features and focus on the local region. Extensive experiments show that SPSL can achieve the state-of-the-art performance on cross-datasets evaluation as well as multi-class classification and obtain comparable results on single dataset evaluation.

cs.CV↗

arXiv2021

Testing f(R) gravity with the simulated data of gravitational waves from the Einstein Telescope

In this paper we analyze the implications of gravitational waves (GWs) as standard sirens on the modified gravity models by using the third-generation gravitational wave detector, i.e., the Einstein Telescope. Two viable models in $f(R)$ theories within the Palatini formalism are considered in our analysis ($f_{1}(\mathcal{R})=\mathcal{R}-\fracβ{\mathcal{R}^{n}}$ and $f_{2}(\mathcal{R})=\mathcal{R}+α\ln{\mathcal{R}}-β$), with the combination of simulated GW data and the latest electromagnetic (EM) observational data (including the recently released Pantheon type Ia supernovae sample, the cosmic chronometer data, and baryon acoustic oscillation distance measurements). Our analysis reveals that the standard sirens GWs, which provide an independent and complementary alternative to current experiments, could effectively eliminate the degeneracies among parameters in the two modified gravity models. In addition, we thoroughly investigate the nature of geometrical dark energy in the modified gravity theories with the assistance of $Om(z)$ and statefinder diagnostic analysis. The present analysis makes it clear-cut that the simplest cosmological constant model is still the most preferred by the current data. However, the combination of future naturally improved GW data most recent EM observations will reveal the consistency or acknowledge the tension between the $Λ$CDM model and modified gravity theories.

astro-ph.CO↗

arXiv2021

Self-Supervised Learning For Few-Shot Image Classification

Few-shot image classification aims to classify unseen classes with limited labelled samples. Recent works benefit from the meta-learning process with episodic tasks and can fast adapt to class from training to testing. Due to the limited number of samples for each task, the initial embedding network for meta-learning becomes an essential component and can largely affect the performance in practice. To this end, most of the existing methods highly rely on the efficient embedding network. Due to the limited labelled data, the scale of embedding network is constrained under a supervised learning(SL) manner which becomes a bottleneck of the few-shot learning methods. In this paper, we proposed to train a more generalized embedding network with self-supervised learning (SSL) which can provide robust representation for downstream tasks by learning from the data itself. We evaluate our work by extensive comparisons with previous baseline methods on two few-shot classification datasets ({\em i.e.,} MiniImageNet and CUB) and achieve better performance over baselines. Tests on four datasets in cross-domain few-shot learning classification show that the proposed method achieves state-of-the-art results and further prove the robustness of the proposed model. Our code is available at \hyperref[https://github.com/phecy/SSL-FEW-SHOT.]{https://github.com/phecy/SSL-FEW-SHOT.}

cs.CV↗