arXiv subjects

Wei Li

Publications and source records attributed to Wei Li.

At least 919 records · Page 51Linked to original sources

arXiv2022

CREIME: A Convolutional Recurrent model for Earthquake Identification and Magnitude Estimation

The detection and rapid characterisation of earthquake parameters such as magnitude are of prime importance in seismology, particularly in applications such as Earthquake Early Warning (EEW). Traditionally, algorithms such as STA/LTA are used for event detection, while frequency or amplitude domain parameters calculated from 1-3 seconds of first P-arrival data are sometimes used to provide a first estimate of (body wave) magnitude. Owing to extensive involvement of human experts in parameter determination, these approaches are often found to be insufficient. Moreover, these methods are sensitive to the signal to noise ratio and may often lead to false or missed alarms depending on the choice of parameters. We, therefore, propose a multitasking deep learning model the Convolutional Recurrent model for Earthquake Identification and Magnitude Estimation (CREIME) that: (i) detects the first earthquake signal, from background seismic noise, (ii) determines first P arrival time as well as (iii) estimates the magnitude using the raw 3-component waveform data from a single station as model input. Considering, speed is of essence in EEW, we use up to two seconds of P-wave information which, to the best of our knowledge, is a significantly smaller data window (5 second window with up to of P wave data) compared to the previous studies. To examine the robustness of CREIME we test it on two independent datasets and find that it achieves an average accuracy of 98 percent for event vs noise discrimination and is able to estimate first P arrival time and local magnitude with average root mean squared errors of 0.13 seconds and 0.65 units, respectively. We also compare CREIME architecture with architectures of other baseline models, by training them on the same data, and also with traditional algorithms such as STA/LTA, and show that our architecture outperforms these methods.

physics.geo-ph↗

arXiv2022

MC-LCR: Multi-modal contrastive classification by locally correlated representations for effective face forgery detection

As the remarkable development of facial manipulation technologies is accompanied by severe security concerns, face forgery detection has become a recent research hotspot. Most existing detection methods train a binary classifier under global supervision to judge real or fake. However, advanced manipulations only perform small-scale tampering, posing challenges to comprehensively capture subtle and local forgery artifacts, especially in high compression settings and cross-dataset scenarios. To address such limitations, we propose a novel framework named Multi-modal Contrastive Classification by Locally Correlated Representations(MC-LCR), for effective face forgery detection. Instead of specific appearance features, our MC-LCR aims to amplify implicit local discrepancies between authentic and forged faces from both spatial and frequency domains. Specifically, we design the shallow style representation block that measures the pairwise correlation of shallow feature maps, which encodes local style information to extract more discriminative features in the spatial domain. Moreover, we make a key observation that subtle forgery artifacts can be further exposed in the patch-wise phase and amplitude spectrum and exhibit different clues. According to the complementarity of amplitude and phase information, we develop a patch-wise amplitude and phase dual attention module to capture locally correlated inconsistencies with each other in the frequency domain. Besides the above two modules, we further introduce the collaboration of supervised contrastive loss with cross-entropy loss. It helps the network learn more discriminative and generalized representations. Through extensive experiments and comprehensive studies, we achieve state-of-the-art performance and demonstrate the robustness and generalization of our method.

cs.CV↗

arXiv2022

MS-HLMO: Multi-scale Histogram of Local Main Orientation for Remote Sensing Image Registration

Multi-source image registration is challenging due to intensity, rotation, and scale differences among the images. Considering the characteristics and differences of multi-source remote sensing images, a feature-based registration algorithm named Multi-scale Histogram of Local Main Orientation (MS-HLMO) is proposed. Harris corner detection is first adopted to generate feature points. The HLMO feature of each Harris feature point is extracted on a Partial Main Orientation Map (PMOM) with a Generalized Gradient Location and Orientation Histogram-like (GGLOH) feature descriptor, which provides high intensity, rotation, and scale invariance. The feature points are matched through a multi-scale matching strategy. Comprehensive experiments on 17 multi-source remote sensing scenes demonstrate that the proposed MS-HLMO and its simplified version MS-HLMO$^+$ outperform other competitive registration algorithms in terms of effectiveness and generalization.

eess.IV↗

arXiv2022

Stochastic Backpropagation: A Memory Efficient Strategy for Training Video Models

We propose a memory efficient method, named Stochastic Backpropagation (SBP), for training deep neural networks on videos. It is based on the finding that gradients from incomplete execution for backpropagation can still effectively train the models with minimal accuracy loss, which attributes to the high redundancy of video. SBP keeps all forward paths but randomly and independently removes the backward paths for each network layer in each training step. It reduces the GPU memory cost by eliminating the need to cache activation values corresponding to the dropped backward paths, whose amount can be controlled by an adjustable keep-ratio. Experiments show that SBP can be applied to a wide range of models for video tasks, leading to up to 80.0% GPU memory saving and 10% training speedup with less than 1% accuracy drop on action recognition and temporal action detection.

cs.CV↗

arXiv2022

MeMOT: Multi-Object Tracking with Memory

We propose an online tracking algorithm that performs the object detection and data association under a common framework, capable of linking objects after a long time span. This is realized by preserving a large spatio-temporal memory to store the identity embeddings of the tracked objects, and by adaptively referencing and aggregating useful information from the memory as needed. Our model, called MeMOT, consists of three main modules that are all Transformer-based: 1) Hypothesis Generation that produce object proposals in the current video frame; 2) Memory Encoding that extracts the core information from the memory for each tracked object; and 3) Memory Decoding that solves the object detection and data association tasks simultaneously for multi-object tracking. When evaluated on widely adopted MOT benchmark datasets, MeMOT observes very competitive performance.

cs.CV↗

arXiv2022

Implication of two-baryon azimuthal correlations in $pp$ collisions at LHC energies on the QGP

The near-side depression in two-proton or two-antiproton azimuthal correlations in $pp$ collisions at $\sqrt{s}=$7 TeV has been observed experimentally and then qualitatively reproduced in our earlier studies with a multi-phase transport model. In this study, we further investigate the origin of the depression feature in two-baryon correlations in small collision systems. We find that the initial parton-level spatial correlation, a finite expansion in the parton stage, and quark coalescence are important ingredients leading to the near-side depression. In particular, we find that a finite expansion of the parton system leads to a finite space-momentum correlation at hadronization, which then converts the near-side depression in the coordinate space to that in the momentum space. These results suggest that a partonic matter with a finite lifetime is formed in the pp collisions. Further studies are needed to determine whether the partonic matter is near local equilibrium and can thus be called a QGP or far away from local equilibrium.

hep-ph↗

arXiv2022

The newly observed state $D_{s0}(2590)^{+}$

We choose the Reduction Formula, PCAC and Low Energy Theory to reduce the $S$ matrix of a OZI allowed two-body strong decay involving a light pseudoscalar, the covariant transition amplitude formula with relativistic wave functions as input is derived. After confirm this method by the decay $D^*(2010)\to Dπ$, we study the newly observed $D_{s0}(2590)^{+}$ with supposing it to be the state $D_s(2^1S_0)^+$, we find its decay width $Γ$ is highly sensitive to the $D_{s0}(2590)^{+}$ mass, which result in the meaningless comparison of widths by different models with various input masses. Instead of width, we studied the overlap integral over the wave functions of initial and final states, here we parameterized it as $X$ which is model-independent, and the ratio $Γ/{|{\vec P_f}|^3}$, both are almost mass independent, to give us useful information. The results show that, all the existing theoretical predictions $X_{D_s(2S) \to D^*K}=0.25\sim 0.41$ and $Γ/{|{\vec P_f}|^3}=0.81\sim1.77$ MeV$^{-2}$ are smaller than experimental data $0.585^{+0.015}_{-0.035}$ and $4.54^{+0.25}_{-0.52}$ MeV$^{-2}$. Further compared with $X^{ex}_{D^*(2010) \to Dπ}=0.540\pm0.009$, the current data $X^{ex}_{D_s(2S) \to D^*K}=0.585^{+0.015}_{-0.035}$ is too big to be an reasonable value, so it is early to say $D_{s0}(2590)^{+}$ is the conventional $D_s(2^1S_0)^+$ meson.

hep-ph↗

arXiv2022

kDecay: Just adding k-decay items on Learning-Rate Schedule to improve Neural Networks

Recent work has shown that optimizing the Learning Rate (LR) schedule can be a very accurate and efficient way to train deep neural networks. We observe that the rate of change (ROC) of LR has correlation with the training process, but how to use this relationship to control the training to achieve the purpose of improving accuracy? We propose a new method, k-decay, just add an extra item to the commonly used and easy LR schedule(exp, cosine and polynomial), is effectively improves the performance of these schedule, also better than the state-of-the-art algorithms of LR shcedule such as SGDR, CLR and AutoLRS. In the k-decay, by adjusting the hyper-parameter $k$, to generate different LR schedule, when k increases, the performance is improved. We evaluate the k-decay method on CIFAR And ImageNet datasets with different neural networks (ResNet, Wide ResNet). Our experiments show that this method can improve on most of them. The accuracy has been improved by 1.08\% on the CIFAR-10 dataset and by 2.07 \% on the CIFAR-100 dataset. On the ImageNet, accuracy is improved by 1.25\%. Our method is not only a general method to be applied other LR Shcedule, but also has no additional computational cost.

cs.LG↗

arXiv2022

Complex Evolutional Pattern Learning for Temporal Knowledge Graph Reasoning

A Temporal Knowledge Graph (TKG) is a sequence of KGs corresponding to different timestamps. TKG reasoning aims to predict potential facts in the future given the historical KG sequences. One key of this task is to mine and understand evolutional patterns of facts from these sequences. The evolutional patterns are complex in two aspects, length-diversity and time-variability. Existing models for TKG reasoning focus on modeling fact sequences of a fixed length, which cannot discover complex evolutional patterns that vary in length. Furthermore, these models are all trained offline, which cannot well adapt to the changes of evolutional patterns from then on. Thus, we propose a new model, called Complex Evolutional Network (CEN), which uses a length-aware Convolutional Neural Network (CNN) to handle evolutional patterns of different lengths via an easy-to-difficult curriculum learning strategy. Besides, we propose to learn the model under the online setting so that it can adapt to the changes of evolutional patterns over time. Extensive experiments demonstrate that CEN obtains substantial performance improvement under both the traditional offline and the proposed online settings.

cs.AI↗

arXiv2022

Towards 3D Scene Understanding by Referring Synthetic Models

Promising performance has been achieved for visual perception on the point cloud. However, the current methods typically rely on labour-extensive annotations on the scene scans. In this paper, we explore how synthetic models alleviate the real scene annotation burden, i.e., taking the labelled 3D synthetic models as reference for supervision, the neural network aims to recognize specific categories of objects on a real scene scan (without scene annotation for supervision). The problem studies how to transfer knowledge from synthetic 3D models to real 3D scenes and is named Referring Transfer Learning (RTL). The main challenge is solving the model-to-scene (from a single model to the scene) and synthetic-to-real (from synthetic model to real scene's object) gap between the synthetic model and the real scene. To this end, we propose a simple yet effective framework to perform two alignment operations. First, physical data alignment aims to make the synthetic models cover the diversity of the scene's objects with data processing techniques. Then a novel \textbf{convex-hull regularized feature alignment} introduces learnable prototypes to project the point features of both synthetic models and real scenes to a unified feature space, which alleviates the domain gap. These operations ease the model-to-scene and synthetic-to-real difficulty for a network to recognize the target objects on a real unseen scene. Experiments show that our method achieves the average mAP of 46.08\% and 55.49\% on the ScanNet and S3DIS datasets by learning the synthetic models from the ModelNet dataset. Code will be publicly available.

cs.CV↗

arXiv2022

UNIMO-2: End-to-End Unified Vision-Language Grounded Learning

Vision-Language Pre-training (VLP) has achieved impressive performance on various cross-modal downstream tasks. However, most existing methods can only learn from aligned image-caption data and rely heavily on expensive regional features, which greatly limits their scalability and performance. In this paper, we propose an end-to-end unified-modal pre-training framework, namely UNIMO-2, for joint learning on both aligned image-caption data and unaligned image-only and text-only corpus. We build a unified Transformer model to jointly learn visual representations, textual representations and semantic alignment between images and texts. In particular, we propose to conduct grounded learning on both images and texts via a sharing grounded space, which helps bridge unaligned images and texts, and align the visual and textual semantic spaces on different types of corpora. The experiments show that our grounded learning method can improve textual and visual semantic alignment for improving performance on various cross-modal tasks. Moreover, benefiting from effective joint modeling of different types of corpora, our model also achieves impressive performance on single-modal visual and textual tasks. Our code and models are public at the UNIMO project page https://unimo-ptm.github.io/.

cs.CV↗

arXiv2022

Bulk locality from the celestial amplitude

In this paper, we study the implications of bulk locality on the celestial amplitude. In the context of the four-point amplitude, the fact that the bulk S-matrix factorizes locally in poles of Mandelstam variables is reflected in the imaginary part of the celestial amplitude. In particular, on the real axis in the complex plane of the boost weight, the imaginary part of the celestial amplitude can be given as a positive expansion on the Poincaré partial waves, which are nothing but the projection of flat-space spinning polynomials onto the celestial sphere. Furthermore, we derive the celestial dispersion relation, which relates the imaginary part to the residue of the celestial amplitude for negative even integer boost weight. The latter is precisely the projection of low energy EFT coefficients onto the celestial sphere. We demonstrate these properties explicitly on the open and closed string celestial amplitudes. Finally, we give an explicit expansion of the Poincaré partial waves in terms of 2D conformal partial waves.

hep-th↗

arXiv2022

UNIMO: Towards Unified-Modal Understanding and Generation via Cross-Modal Contrastive Learning

Existed pre-training methods either focus on single-modal tasks or multi-modal tasks, and cannot effectively adapt to each other. They can only utilize single-modal data (i.e. text or image) or limited multi-modal data (i.e. image-text pairs). In this work, we propose a unified-modal pre-training architecture, namely UNIMO, which can effectively adapt to both single-modal and multi-modal understanding and generation tasks. Large scale of free text corpus and image collections can be utilized to improve the capability of visual and textual understanding, and cross-modal contrastive learning (CMCL) is leveraged to align the textual and visual information into a unified semantic space over a corpus of image-text pairs. As the non-paired single-modal data is very rich, our model can utilize much larger scale of data to learn more generalizable representations. Moreover, the textual knowledge and visual knowledge can enhance each other in the unified semantic space. The experimental results show that UNIMO significantly improves the performance of several single-modal and multi-modal downstream tasks. Our code and pre-trained models are public at the UNIMO project page https://unimo-ptm.github.io/

cs.CL↗

arXiv2022

Semi-Supervised Semantic Segmentation Using Unreliable Pseudo-Labels

The crux of semi-supervised semantic segmentation is to assign adequate pseudo-labels to the pixels of unlabeled images. A common practice is to select the highly confident predictions as the pseudo ground-truth, but it leads to a problem that most pixels may be left unused due to their unreliability. We argue that every pixel matters to the model training, even its prediction is ambiguous. Intuitively, an unreliable prediction may get confused among the top classes (i.e., those with the highest probabilities), however, it should be confident about the pixel not belonging to the remaining classes. Hence, such a pixel can be convincingly treated as a negative sample to those most unlikely categories. Based on this insight, we develop an effective pipeline to make sufficient use of unlabeled data. Concretely, we separate reliable and unreliable pixels via the entropy of predictions, push each unreliable pixel to a category-wise queue that consists of negative samples, and manage to train the model with all candidate pixels. Considering the training evolution, where the prediction becomes more and more accurate, we adaptively adjust the threshold for the reliable-unreliable partition. Experimental results on various benchmarks and training settings demonstrate the superiority of our approach over the state-of-the-art alternatives.

cs.CV↗

arXiv2022

Efficient universal shuffle attack for visual object tracking

Recently, adversarial attacks have been applied in visual object tracking to deceive deep trackers by injecting imperceptible perturbations into video frames. However, previous work only generates the video-specific perturbations, which restricts its application scenarios. In addition, existing attacks are difficult to implement in reality due to the real-time of tracking and the re-initialization mechanism. To address these issues, we propose an offline universal adversarial attack called Efficient Universal Shuffle Attack. It takes only one perturbation to cause the tracker malfunction on all videos. To improve the computational efficiency and attack performance, we propose a greedy gradient strategy and a triple loss to efficiently capture and attack model-specific feature representations through the gradients. Experimental results show that EUSA can significantly reduce the performance of state-of-the-art trackers on OTB2015 and VOT2018.

cs.CV↗

arXiv2022

UniVIP: A Unified Framework for Self-Supervised Visual Pre-training

Self-supervised learning (SSL) holds promise in leveraging large amounts of unlabeled data. However, the success of popular SSL methods has limited on single-centric-object images like those in ImageNet and ignores the correlation among the scene and instances, as well as the semantic difference of instances in the scene. To address the above problems, we propose a Unified Self-supervised Visual Pre-training (UniVIP), a novel self-supervised framework to learn versatile visual representations on either single-centric-object or non-iconic dataset. The framework takes into account the representation learning at three levels: 1) the similarity of scene-scene, 2) the correlation of scene-instance, 3) the discrimination of instance-instance. During the learning, we adopt the optimal transport algorithm to automatically measure the discrimination of instances. Massive experiments show that UniVIP pre-trained on non-iconic COCO achieves state-of-the-art transfer performance on a variety of downstream tasks, such as image classification, semi-supervised learning, object detection and segmentation. Furthermore, our method can also exploit single-centric-object dataset such as ImageNet and outperforms BYOL by 2.5% with the same pre-training epochs in linear probing, and surpass current self-supervised object detection methods on COCO dataset, demonstrating its universality and potential.

cs.CV↗

arXiv2022

Lepton flavor violating decays $τ\rightarrow{Pl}$ in the $U(1)_X$SSM model

$U(1)_X$SSM is the extension of the minimal supersymmetric standard model(MSSM) and its local gauge group is $SU(3)_C\times SU(2)_L \times U(1)_Y \times U(1)_X$. To obtain this model, three singlet new Higgs superfields and right-handed neutrinos are added to MSSM. In the framework of $U(1)_X$SSM, we study lepton flavor violating decays $τ\rightarrow{Pl}(P=π,~η,~η^\prime;~l=μ,~e)$. According to the latest experimental data of $τ\rightarrow{Pe}$ and $τ\rightarrow{Pμ}$, the influence of different sensitive parameters on the results is analyzed to make a reasonable prediction for future experiments. From the numerical analysis, the non-diagonal elements which correspond to the generations of the initial lepton and final lepton are main sensitive parameters and lepton flavor violation(LFV) sources. This work can provide a basis for finding the existence of new physics(NP).

hep-ph↗

arXiv2022

Faithfulness in Natural Language Generation: A Systematic Survey of Analysis, Evaluation and Optimization Methods

Natural Language Generation (NLG) has made great progress in recent years due to the development of deep learning techniques such as pre-trained language models. This advancement has resulted in more fluent, coherent and even properties controllable (e.g. stylistic, sentiment, length etc.) generation, naturally leading to development in downstream tasks such as abstractive summarization, dialogue generation, machine translation, and data-to-text generation. However, the faithfulness problem that the generated text usually contains unfaithful or non-factual information has become the biggest challenge, which makes the performance of text generation unsatisfactory for practical applications in many real-world scenarios. Many studies on analysis, evaluation, and optimization methods for faithfulness problems have been proposed for various tasks, but have not been organized, compared and discussed in a combined manner. In this survey, we provide a systematic overview of the research progress on the faithfulness problem of NLG, including problem analysis, evaluation metrics and optimization methods. We organize the evaluation and optimization methods for different tasks into a unified taxonomy to facilitate comparison and learning across tasks. Several research trends are discussed further.

cs.CL↗