arXiv subjects

Wei Ji

Publications and source records attributed to Wei Ji.

At least 181 records · Page 10Linked to original sources

arXiv2023

Two Heads Are Better Than One: Improving Fake News Video Detection by Correlating with Neighbors

The prevalence of short video platforms has spawned a lot of fake news videos, which have stronger propagation ability than textual fake news. Thus, automatically detecting fake news videos has been an important countermeasure in practice. Previous works commonly verify each news video individually with multimodal information. Nevertheless, news videos from different perspectives regarding the same event are commonly posted together, which contain complementary or contradictory information and thus can be used to evaluate each other mutually. To this end, we introduce a new and practical paradigm, i.e., cross-sample fake news video detection, and propose a novel framework, Neighbor-Enhanced fakE news video Detection (NEED), which integrates the neighborhood relationship of new videos belonging to the same event. NEED can be readily combined with existing single-sample detectors and further enhance their performances with the proposed graph aggregation (GA) and debunking rectification (DR) modules. Specifically, given the feature representations obtained from single-sample detectors, GA aggregates the neighborhood information with the dynamic graph to enrich the features of independent samples. After that, DR explicitly leverages the relationship between debunking videos and fake news videos to refute the candidate videos via textual and visual consistency. Extensive experiments on the public benchmark demonstrate that NEED greatly improves the performance of both single-modal (up to 8.34% in accuracy) and multimodal (up to 4.97% in accuracy) base detectors. Codes are available in https://github.com/ICTMCG/NEED.

cs.MM↗

arXiv2023

Generating Visual Spatial Description via Holistic 3D Scene Understanding

Visual spatial description (VSD) aims to generate texts that describe the spatial relations of the given objects within images. Existing VSD work merely models the 2D geometrical vision features, thus inevitably falling prey to the problem of skewed spatial understanding of target objects. In this work, we investigate the incorporation of 3D scene features for VSD. With an external 3D scene extractor, we obtain the 3D objects and scene features for input images, based on which we construct a target object-centered 3D spatial scene graph (Go3D-S2G), such that we model the spatial semantics of target objects within the holistic 3D scenes. Besides, we propose a scene subgraph selecting mechanism, sampling topologically-diverse subgraphs from Go3D-S2G, where the diverse local structure features are navigated to yield spatially-diversified text generation. Experimental results on two VSD datasets demonstrate that our framework outperforms the baselines significantly, especially improving on the cases with complex visual spatial relations. Meanwhile, our method can produce more spatially-diversified generation. Code is available at https://github.com/zhaoyucs/VSD.

cs.CV↗

arXiv2023

Cross2StrA: Unpaired Cross-lingual Image Captioning with Cross-lingual Cross-modal Structure-pivoted Alignment

Unpaired cross-lingual image captioning has long suffered from irrelevancy and disfluency issues, due to the inconsistencies of the semantic scene and syntax attributes during transfer. In this work, we propose to address the above problems by incorporating the scene graph (SG) structures and the syntactic constituency (SC) trees. Our captioner contains the semantic structure-guided image-to-pivot captioning and the syntactic structure-guided pivot-to-target translation, two of which are joined via pivot language. We then take the SG and SC structures as pivoting, performing cross-modal semantic structure alignment and cross-lingual syntactic structure alignment learning. We further introduce cross-lingual&cross-modal back-translation training to fully align the captioning and translation stages. Experiments on English-Chinese transfers show that our model shows great superiority in improving captioning relevancy and fluency.

cs.CV↗

arXiv2023

Segment Anything Is Not Always Perfect: An Investigation of SAM on Different Real-world Applications

Recently, Meta AI Research approaches a general, promptable Segment Anything Model (SAM) pre-trained on an unprecedentedly large segmentation dataset (SA-1B). Without a doubt, the emergence of SAM will yield significant benefits for a wide array of practical image segmentation applications. In this study, we conduct a series of intriguing investigations into the performance of SAM across various applications, particularly in the fields of natural images, agriculture, manufacturing, remote sensing, and healthcare. We analyze and discuss the benefits and limitations of SAM, while also presenting an outlook on its future development in segmentation tasks. By doing so, we aim to give a comprehensive understanding of SAM's practical applications. This work is expected to provide insights that facilitate future research activities toward generic segmentation. Source code is publicly available.

cs.CV↗

arXiv2023

Discovery and construction of surface kagome electronic states induced by p-d electronic hybridization

Kagome-lattice materials possess attractive properties for quantum computing applications, but their synthesis remains challenging. Herein, we show surface kagome electronic states (SKESs) on a Sn-terminated triangular Co3Sn2S2 surface, which are imprinted by vertical p-d electronic hybridization between the surface Sn (subsurface S) atoms and the buried Co kagome lattice network in the Co3Sn layer under the surface. Owing to the subsequent lateral hybridization of the Sn and S atoms in a corner-sharing manner, the kagome symmetry and topological electronic properties of the Co3Sn layer is proximate to the Sn surface. The SKESs and both hybridizations were verified via qPlus non-contact atomic force microscopy (nc-AFM) and density functional theory calculations. The construction of SKESs with tunable properties can be achieved by the atomic substitution of surface Sn (subsurface S) with other group III-V elements (Se or Te), which was demonstrated theoretically. This work exhibits the powerful capacity of nc-AFM in characterizing localized topological states and reveals the strategy for synthesis of large-area transition-metal-based kagome lattice materials using conventional surface deposition techniques.

cond-mat.mtrl-sci↗

arXiv2023

Controllable dimensionality conversion between 1D and 2D CrCl3 magnetic nanostructures

The fabrication of one-dimensional (1D) magnetic systems on solid surfaces, although of high fundamental interest, has yet to be achieved for a crossover between two-dimensional (2D) magnetic layers and their associated 1D spin chain systems. In this study, we report the fabrication of 1D single-unit-cell-width CrCl3 atomic wires and their stacked few-wire arrays on the surface of a van der Waals (vdW) superconductor NbSe2. Scanning tunneling microscopy/spectroscopy and first-principles calculations jointly revealed that the single wire shows an antiferromagnetic large-bandgap semiconducting state in an unexplored structure different from the well-known 2D CrCl3 phase. Competition among the total energies and nanostructure-substrate interfacial interactions of these two phases result in the appearance of the 1D phase. This phase was transformable to the 2D phase either prior to or after the growth for in situ or ex situ manipulations, in which the electronic interactions at the vdW interface play a nontrivial role that could regulate the dimensionality conversion and structural transformation between the 1D-2D CrCl3 phases.

cond-mat.mtrl-sci↗

arXiv2023

The Stellar Spectra Factory (SSF) Based On SLAM

In this work, we present Stellar Spectra Factory (SSF), a tool to generate empirical-based stellar spectra from arbitrary stellar atmospheric parameters. The relative flux-calibrated empirical spectra can be predicted by SSF given arbitrary effective temperature, surface gravity, and metallicity. SSF constructs the interpolation approach based on the SLAM, using ATLAS-A library as the training dataset. SSF is composed of 4 data-driven sub-models to predict empirical stellar spectra. SSF-N can generate spectra from A to K type and some M giant stars, covering 3700 < Teff < 8700 K, 0 < logg < 6 dex, and -1.5 < [M/H] < 0.5 dex. SSF-gM is mainly used to predict M giant spectra with 3520 < Teff < 4000K and -1.5 < [M/H] < 0.4 dex. SSF-dM is for generating M dwarf spectra with 3295 < Teff < 4040K, -1.0 < [M/H] < 0.1 dex. And SSF-B can predict B-type spectra with 9000 < Teff < 24000K and -5.2< MG < 1.5 mag. The accuracy of the predicted spectra is validated by comparing the flux of predicted spectra to those with same stellar parameters selected from the known spectral libraries, MILES and MaStar. The averaged difference of flux over optical wavelength between the predicted spectra and the corresponding ones in MILES and MaStar is less than 5%. More verification is conducted between the magnitudes calculated from the integration of the predicted spectra and the observations in PS1 and APASS bands with the same stellar parameters. No significant systematic difference is found between the predicted spectra and the photomatric observations. The uncertainty is 0.08mag in r band for SSF-gM when comparing with the stars with the same stellar parameters selected from PS1. And the uncertainty becomes 0.31mag in i band for SSF-dM when comparing with the stars with the same stellar parameters selected from APASS.

astro-ph.SR↗

arXiv2023

3D Magic Mirror: Clothing Reconstruction from a Single Image via a Causal Perspective

This research aims to study a self-supervised 3D clothing reconstruction method, which recovers the geometry shape and texture of human clothing from a single image. Compared with existing methods, we observe that three primary challenges remain: (1) 3D ground-truth meshes of clothing are usually inaccessible due to annotation difficulties and time costs; (2) Conventional template-based methods are limited to modeling non-rigid objects, e.g., handbags and dresses, which are common in fashion images; (3) The inherent ambiguity compromises the model training, such as the dilemma between a large shape with a remote camera or a small shape with a close camera. In an attempt to address the above limitations, we propose a causality-aware self-supervised learning method to adaptively reconstruct 3D non-rigid objects from 2D images without 3D annotations. In particular, to solve the inherent ambiguity among four implicit variables, i.e., camera position, shape, texture, and illumination, we introduce an explainable structural causal map (SCM) to build our model. The proposed model structure follows the spirit of the causal map, which explicitly considers the prior template in the camera estimation and shape prediction. When optimization, the causality intervention tool, i.e., two expectation-maximization loops, is deeply embedded in our algorithm to (1) disentangle four encoders and (2) facilitate the prior template. Extensive experiments on two 2D fashion benchmarks (ATR and Market-HQ) show that the proposed method could yield high-fidelity 3D reconstruction. Furthermore, we also verify the scalability of the proposed method on a fine-grained bird dataset, i.e., CUB. The code is available at https://github.com/layumi/ 3D-Magic-Mirror .

cs.CV↗

arXiv2023

Scalable Attribution of Adversarial Attacks via Multi-Task Learning

Deep neural networks (DNNs) can be easily fooled by adversarial attacks during inference phase when attackers add imperceptible perturbations to original examples, i.e., adversarial examples. Many works focus on adversarial detection and adversarial training to defend against adversarial attacks. However, few works explore the tool-chains behind adversarial examples, which can help defenders to seize the clues about the originator of the attack, their goals, and provide insight into the most effective defense algorithm against corresponding attacks. With such a gap, it is necessary to develop techniques that can recognize tool-chains that are leveraged to generate the adversarial examples, which is called Adversarial Attribution Problem (AAP). In this paper, AAP is defined as the recognition of three signatures, i.e., {\em attack algorithm}, {\em victim model} and {\em hyperparameter}. Current works transfer AAP into single label classification task and ignore the relationship between these signatures. The former will meet combination explosion problem as the number of signatures is increasing. The latter dictates that we cannot treat AAP simply as a single task problem. We first conduct some experiments to validate the attributability of adversarial examples. Furthermore, we propose a multi-task learning framework named Multi-Task Adversarial Attribution (MTAA) to recognize the three signatures simultaneously. MTAA contains perturbation extraction module, adversarial-only extraction module and classification and regression module. It takes the relationship between attack algorithm and corresponding hyperparameter into account and uses the uncertainty weighted loss to adjust the weights of three recognition tasks. The experimental results on MNIST and ImageNet show the feasibility and scalability of the proposed framework as well as its effectiveness in dealing with false alarms.

cs.LG↗

arXiv2023

Controllable CVD-Growth of 2D Cr5Te8 Nanosheets with Thickness-Dependent Magnetic Domains

As a typical 2D magnetic material with self-intercalated structure, Cr5Te8 exhibits many fascinating magnetic properties. Although the ferromagnetism of 2D Cr5Te8 has been reported, the study of its magnetic domain is still blank. Herein, we successfully fabricated 2D Cr5Te8 nanosheets with controlled thickness and lateral size by chemical vapor deposition (CVD). Then magnetic property measurement system suggested Cr5Te8 nanosheets possessing intense out-of-plane ferromagnetism with the Curie temperature of 179 K. Most importantly, we found magnetic bubbles and thickness-dependent maze-like magnetic domains by cryogenic magnetic force microscopy (MFM) for the first time. The domain width of the maze-like magnetic domains increases rapidly with decreasing sample thickness, while domain contrast decreases. This indicates dipolar interaction is the dominant role over exchange interaction and magnetic anisotropy. Our work not only paves a way for the controllable growth of 2D magnetic materials, but also suggests new directions for controlling magnetic phases and systematically adjusting domain properties.

cond-mat.mes-hall↗

arXiv2023

Electronic Janus lattice and kagome-like bands in coloring-triangular MoTe2 monolayers

Polymorphic structures of transition-metal dichalcogenides (TMDs) host exotic electronic states, like charge density wave and superconductivity. However, the number of these structures is limited by crystal symmetries, which poses a challenge to achieve tailored lattices and properties both theoretically and experimentally. Here, we report a coloring-triangle (CT) latticed MoTe2 monolayer, termed CT-MoTe2, constructed by controllably introducing uniform and ordered mirror-twin-boundaries into a pristine monolayer in molecular beam epitaxy. Low-temperature scanning tunneling microscopy and spectroscopy (STM/STS) together with theoretical calculations reveal that the monolayer has an electronic Janus lattice, i.e., an energy-dependent atomic-lattice and a pseudo-Te sublattice, and shares the identical geometry with the Mo5Te8 layer. Dirac-like and flat electronic bands inherently existing in the CT lattice are identified by two broad and two prominent peaks in STS spectra, respectively, and verified with density-functional-theory calculations. Two types of intrinsic domain boundaries were observed, in one of which the electronic-Janus-lattice feature maintains, implying potential applications as an energy-tunable electron-tunneling barrier in future functional devices.

cond-mat.mtrl-sci↗

arXiv2023

Layer sliding and twisting induced electronic transitions in correlated magnetic 1T-NbSe2 bilayers

Correlated two-dimensional (2D) layers, like 1T-phases of TaS2, TaSe2 and NbSe2, exhibit rich tunability through varying interlayer couplings, which promotes the understanding of electron-correlation in the 2D limit. However, the coupling mechanism is, so far, poorly understood and was tentatively ascribed to interactions among the d_(z^2 ) orbitals of Ta or Nb atoms. Here, we theoretically show that the interlayer hybridization and localization strength of interfacial Se pz orbitals, rather than Nb d_(z^2 ) orbitals, govern the variation of electron-correlated properties upon interlayer sliding or twisting in correlated magnetic 1T-NbSe2 bilayers. Each of the both layers is in a star-of-David (SOD) charge-density-wave phase. Geometric and electronic structures, and magnetic properties of 28 different stacking configurations were examined and analyzed using density-functional-theory calculations. We found that the SOD contains a localized region (Reg-L), in which interlayer Se pz hybridization plays a paramount role in varying the energy levels of the two Hubbard bands. These variations lead to three electronic transitions among four insulating states, which demonstrated the effectiveness of interlayer interactions to modulate correlated magnetic properties in a prototypical correlated magnetic insulator.

cond-mat.mtrl-sci↗

arXiv2023

Emergent Electronic Kagome Lattice in Correlated Charge-Density-Wave State of 1T-TaS$_2$

Quantum materials with tunable correlated and/or topological electronic states, such as the electronic Kagome lattice, provide an ideal platform to study the exotic quantum properties. However, the real-space investigations on the correlated electronic Kagome lattice have been rarely reported. Herein, we report on the electronic Kagome lattice emerging in the correlated charge-density-wave (CDW) state of 1T-TaS$_2$ at ~200 K via variable-temperature scanning tunneling microscopy (VT-STM). This emergent Kagome lattice can be considered a fractional electron-filling superstructure with reduced translational and rotational symmetries, confirmed by STM measurements and density functional theory simulations. The characteristic band structure and density of states of this electronic Kagome lattice are further explored based on theoretical calculations. Our results demonstrate a self-organized electronic Kagome lattice from the correlated CDW state via the effective tuning parameter of temperature and provide a platform to directly explore the interplay of correlated electrons and topological physics.

cond-mat.mtrl-sci↗

arXiv2023

Optical Switching Data Center Networks: Understanding Techniques and Challenges

Relying on the flexible-access interconnects to the scalable storage and compute resources, data centers deliver critical communications connectivity among numerous servers to support the housed applications and services. To provide the high-speeds and long-distance communications, the data centers have turned to fiber interconnections. With the stringently increased traffic volume, the data centers are then expected to further deploy the optical switches into the systems infrastructure to implement the full optical switching. This paper first summarizes the topologies and traffic characteristics in data centers and analyzes the reasons and importance of moving to optical switching. Recent techniques related to the optical switching, and main challenges limiting the practical deployments of optical switches in data centers are also summarized and reported.

cs.NI↗

arXiv2022

MRTNet: Multi-Resolution Temporal Network for Video Sentence Grounding

Given an untrimmed video and natural language query, video sentence grounding aims to localize the target temporal moment in the video. Existing methods mainly tackle this task by matching and aligning semantics of the descriptive sentence and video segments on a single temporal resolution, while neglecting the temporal consistency of video content in different resolutions. In this work, we propose a novel multi-resolution temporal video sentence grounding network: MRTNet, which consists of a multi-modal feature encoder, a Multi-Resolution Temporal (MRT) module, and a predictor module. MRT module is an encoder-decoder network, and output features in the decoder part are in conjunction with Transformers to predict the final start and end timestamps. Particularly, our MRT module is hot-pluggable, which means it can be seamlessly incorporated into any anchor-free models. Besides, we utilize a hybrid loss to supervise cross-modal features in MRT module for more accurate grounding in three scales: frame-level, clip-level and sequence-level. Extensive experiments on three prevalent datasets have shown the effectiveness of MRTNet.

cs.CV↗

arXiv2022

Multi-queue Momentum Contrast for Microvideo-Product Retrieval

The booming development and huge market of micro-videos bring new e-commerce channels for merchants. Currently, more micro-video publishers prefer to embed relevant ads into their micro-videos, which not only provides them with business income but helps the audiences to discover their interesting products. However, due to the micro-video recording by unprofessional equipment, involving various topics and including multiple modalities, it is challenging to locate the products related to micro-videos efficiently, appropriately, and accurately. We formulate the microvideo-product retrieval task, which is the first attempt to explore the retrieval between the multi-modal and multi-modal instances. A novel approach named Multi-Queue Momentum Contrast (MQMC) network is proposed for bidirectional retrieval, consisting of the uni-modal feature and multi-modal instance representation learning. Moreover, a discriminative selection strategy with a multi-queue is used to distinguish the importance of different negatives based on their categories. We collect two large-scale microvideo-product datasets (MVS and MVS-large) for evaluation and manually construct the hierarchical category ontology, which covers sundry products in daily life. Extensive experiments show that MQMC outperforms the state-of-the-art baselines. Our replication package (including code, dataset, etc.) is publicly available at https://github.com/duyali2000/MQMC.

cs.CV↗

arXiv2022

Dynamic Speed Guidance for CAV Ramp Merging in Non-Cooperative Environment: An On-Site Experiment

Ramp merging is a typical application of cooperative intelligent transportation system (C-ITS). Vehicle trajectories perceived by roadside sensors are importation complement to the limited visual field of on-board perception. Vehicle tracking and trajectory denoising algorithm is proposed in this paper to take full advantage of roadside cameras for vehicle trajectory and speed profile estimation. Dynamic speed guidance algorithm is proposed to help on-ramp vehicles to merge into mainline smoothly, even in non-cooperative environment where mainline vehicles are not expected to slow down to accommodate on-ramp vehicles. On-site experiments were taken out in a merging area of Hangzhou Belt Highway to testify our prototype system, and simulation analysis shows our proposed algorithm can achieve significant fuel savings during the ramp merging process.

cs.RO↗

arXiv2022

Driving Style Recognition at First Impression for Online Trajectory Prediction

This paper proposes a new driving style recognition approach that allows autonomous vehicles (AVs) to perform trajectory predictions for surrounding vehicles with minimal data. Toward that end, we use a hybrid of offline and online methods in the proposed approach. We first learn typical driving styles with PCA and K-means algorithms in the offline part. After that, local Maximum-Likelihood techniques are used to perform online driving style recognition. We benchmarked our method on a real driving dataset against other methods in terms of the RMSE value of the predicted trajectory and the observed trajectory over a 5s duration. The proposed approach can reduce trajectory prediction error by up to 37.7\% compared to using the parameters from other literature and up to 24.4\% compared to not performing driving style recognition.

eess.SY↗