Search arXiv⌕ Search

arXiv subjects

Bin Li

Publications and source records attributed to Bin Li.

At least 379 records · Page 21Linked to original sources
arXiv2023

Enhanced Fine-grained Motion Diffusion for Text-driven Human Motion Synthesis

The emergence of text-driven motion synthesis technique provides animators with great potential to create efficiently. However, in most cases, textual expressions only contain general and qualitative motion descriptions, while lack fine depiction and sufficient intensity, leading to the synthesized motions that either (a) semantically compliant but uncontrollable over specific pose details, or (b) even deviates from the provided descriptions, bringing animators with undesired cases. In this paper, we propose DiffKFC, a conditional diffusion model for text-driven motion synthesis with KeyFrames Collaborated, enabling realistic generation with collaborative and efficient dual-level control: coarse guidance at semantic level, with only few keyframes for direct and fine-grained depiction down to body posture level. Unlike existing inference-editing diffusion models that incorporate conditions without training, our conditional diffusion model is explicitly trained and can fully exploit correlations among texts, keyframes and the diffused target frames. To preserve the control capability of discrete and sparse keyframes, we customize dilated mask attention modules where only partial valid tokens participate in local-to-global attention, indicated by the dilated keyframe mask. Additionally, we develop a simple yet effective smoothness prior, which steers the generated frames towards seamless keyframe transitions at inference. Extensive experiments show that our model not only achieves state-of-the-art performance in terms of semantic fidelity, but more importantly, is able to satisfy animator requirements through fine-grained guidance without tedious labor.

cs.CV↗
arXiv2023

Second-order computational homogenization of flexoelectric composites

Flexoelectricity shows promising applications for self-powered devices with its increased power density. This paper presents a second-order computational homogenization strategy for flexoelectric composite. The macro-micro scale transition, Hill-Mandel energy condition, periodic boundary conditions, and macroscopic constitutive tangents for the two-scale electromechanical coupling are investigated and considered in the homogenization formulation. The macrostructure and microstructure are discretized using $C^1$ triangular finite elements. The second-order multiscale solution scheme is implemented using ABAQUS with user subroutines. Finally, we present numerical examples including parametric analysis of a square plate with holes and the design of piezoelectric materials made of non-piezoelectric materials to demonstrate the numerical implementation and the size-dependent effects of flexoelectricity.

math.NA↗
arXiv2023

Large electrobending deformation caused by defect dipoles

Ultrahigh electrostrains (greater than 1%) in several piezoceramic systems have been reported since 2022, which attract more and more interest in the field of piezoelectricity; however, the mechanism is still unclear. Here, we have directly observed a novel electric field-induced bending (electrobending) phenomenon that visually exhibites as an alternating concave-convex deformation under an electric field of plus or minus 50 kV cm-1, in nonstoichiometric (K0.48Na0.52)0.99NbO2.995 ceramics, which causes the measured ultrahigh electrostrain. It is demonstrated that the electrobending deformation arises from the different stresses due to the stretching or compression of the defect dipoles on the upper and lower surfaces of the ceramics. As a result of the large electrobending deformation, a giant apparent electrostrain of 11.6% is obtained at room temperature, and it can even reach up to 26.0% at 210 degree Celsius, which far exceeds that of all present piezoelectric materials. Our discovery is an important addition and refinement to the field of condensed matter physics, whilst providing a new strategy and shedding light on the design of future precision actuators or intelligent devices.

physics.app-ph↗
arXiv2023

Adapter is All You Need for Tuning Visual Tasks

Pre-training & fine-tuning can enhance the transferring efficiency and performance in visual tasks. Recent delta-tuning methods provide more options for visual classification tasks. Despite their success, existing visual delta-tuning art fails to exceed the upper limit of full fine-tuning on challenging tasks like instance segmentation and semantic segmentation. To find a competitive alternative to full fine-tuning, we propose the Multi-cognitive Visual Adapter (Mona) tuning, a novel adapter-based tuning method. First, we introduce multiple vision-friendly filters into the adapter to enhance its ability to process visual signals, while previous methods mainly rely on language-friendly linear filters. Second, we add the scaled normalization layer in the adapter to regulate the distribution of input features for visual filters. To fully demonstrate the practicality and generality of Mona, we conduct experiments on multiple representative visual tasks, including instance segmentation on COCO, semantic segmentation on ADE20K, object detection on Pascal VOC, and image classification on several common datasets. Exciting results illustrate that Mona surpasses full fine-tuning on all these tasks and is the only delta-tuning method outperforming full fine-tuning on instance segmentation and semantic segmentation tasks. For example, Mona achieves a 1% performance gain on the COCO dataset compared to full fine-tuning. Comprehensive results suggest that Mona-tuning is more suitable for retaining and utilizing the capabilities of pre-trained models than full fine-tuning. The code will be released at https://github.com/Leiyi-Hu/mona.

cs.CV↗
arXiv2023

Parameter-efficient is not sufficient: Exploring Parameter, Memory, and Time Efficient Adapter Tuning for Dense Predictions

Pre-training & fine-tuning is a prevalent paradigm in computer vision (CV). Recently, parameter-efficient transfer learning (PETL) methods have shown promising performance in adapting to downstream tasks with only a few trainable parameters. Despite their success, the existing PETL methods in CV can be computationally expensive and require large amounts of memory and time cost during training, which limits low-resource users from conducting research and applications on large models. In this work, we propose Parameter, Memory, and Time Efficient Visual Adapter ($\mathrm{E^3VA}$) tuning to address this issue. We provide a gradient backpropagation highway for low-rank adapters which eliminates the need for expensive backpropagation through the frozen pre-trained model, resulting in substantial savings of training memory and training time. Furthermore, we optimise the $\mathrm{E^3VA}$ structure for CV tasks to promote model performance. Extensive experiments on COCO, ADE20K, and Pascal VOC benchmarks show that $\mathrm{E^3VA}$ can save up to 62.2% training memory and 26.2% training time on average, while achieving comparable performance to full fine-tuning and better performance than most PETL methods. Note that we can even train the Swin-Large-based Cascade Mask RCNN on GTX 1080Ti GPUs with less than 1.5% trainable parameters.

cs.CV↗
arXiv2023

Federated Transformed Learning for a Circular, Secure, and Tiny AI

Deep Learning (DL) is penetrating into a diverse range of mass mobility, smart living, and industrial applications, rapidly transforming the way we live and work. DL is at the heart of many AI implementations. A key set of challenges is to produce AI modules that are: (1) "circular" - can solve new tasks without forgetting how to solve previous ones, (2) "secure" - have immunity to adversarial data attacks, and (3) "tiny" - implementable in low power low cost embedded hardware. Clearly it is difficult to achieve all three aspects on a single horizontal layer of platforms, as the techniques require transformed deep representations that incur different computation and communication requirements. Here we set out the vision to achieve transformed DL representations across a 5G and Beyond networked architecture. We first detail the cross-sectoral motivations for each challenge area, before demonstrating recent advances in DL research that can achieve circular, secure, and tiny AI (CST-AI). Recognising the conflicting demand of each transformed deep representation, we federate their deep learning transformations and functionalities across the network to achieve connected run-time capabilities.

cs.NI↗
arXiv2023

Efficient Trigger Word Insertion

With the boom in the natural language processing (NLP) field these years, backdoor attacks pose immense threats against deep neural network models. However, previous works hardly consider the effect of the poisoning rate. In this paper, our main objective is to reduce the number of poisoned samples while still achieving a satisfactory Attack Success Rate (ASR) in text backdoor attacks. To accomplish this, we propose an efficient trigger word insertion strategy in terms of trigger word optimization and poisoned sample selection. Extensive experiments on different datasets and models demonstrate that our proposed method can significantly improve attack effectiveness in text classification tasks. Remarkably, our approach achieves an ASR of over 90% with only 10 poisoned samples in the dirty-label setting and requires merely 1.5% of the training data in the clean-label setting.

cs.CR↗
arXiv2023

The first Ka-band (26.1-35 GHz) blind line survey towards Orion KL

We conducted a Ka-band (26.1--35 GHz) line survey towards Orion KL using the TianMa 65-m Radio Telescope (TMRT). It is the first blind line survey in the Ka band, and achieves a sensitivity of mK level (1--3 mK at a spectral resolution of $\sim$1 km s$^{-1}$). In total, 592 Gaussian features are extracted. Among them, 257 radio recombination lines (RRLs) are identified. The maximum $Δn$ of RRLs of H, He and C are 20, 15, and 5, respectively. Through stacking, we have detected the $β$ lines of ion RRLs (RRLs of C$^+$ with possible contribution of other ions like O$^+$) for the first time, and tentative signal of the $γ$ lines of ion RRLs can also be seen on the stacked spectrum. Besides, 318 other line features were assigned to 37 molecular species, and ten of these species were not detected in the Q-band survey of TMRT. The vibrationally excited states of nine species were also detected. Emission of most species can be modeled under LTE. A number of transitions of E-CH3OH ($J_2-J_1$) display maser effects, which are confirmed by our modeling, and besides the bumping peak at $J\sim 6$ there is another peak at $J\sim 13$. Methylcyanoacetylene (CH$_3$C$_3$N) is detected in Orion KL for the first time. This work emphasizes that the Ka band, which was long-ignored for spectral line surveys, is very useful for surveying RRLs and molecular lines simultaneously.

astro-ph.GA↗
arXiv2023

NAS-ASDet: An Adaptive Design Method for Surface Defect Detection Network using Neural Architecture Search

Deep convolutional neural networks (CNNs) have been widely used in surface defect detection. However, no CNN architecture is suitable for all detection tasks and designing effective task-specific requires considerable effort. The neural architecture search (NAS) technology makes it possible to automatically generate adaptive data-driven networks. Here, we propose a new method called NAS-ASDet to adaptively design network for surface defect detection. First, a refined and industry-appropriate search space that can adaptively adjust the feature distribution is designed, which consists of repeatedly stacked basic novel cells with searchable attention operations. Then, a progressive search strategy with a deep supervision mechanism is used to explore the search space faster and better. This method can design high-performance and lightweight defect detection networks with data scarcity in industrial scenarios. The experimental results on four datasets demonstrate that the proposed method achieves superior performance and a relatively lighter model size compared to other competitive methods, including both manual and NAS-based approaches.

cs.CV↗
arXiv2023

Peer is Your Pillar: A Data-unbalanced Conditional GANs for Few-shot Image Generation

Few-shot image generation aims to train generative models using a small number of training images. When there are few images available for training (e.g. 10 images), Learning From Scratch (LFS) methods often generate images that closely resemble the training data while Transfer Learning (TL) methods try to improve performance by leveraging prior knowledge from GANs pre-trained on large-scale datasets. However, current TL methods may not allow for sufficient control over the degree of knowledge preservation from the source model, making them unsuitable for setups where the source and target domains are not closely related. To address this, we propose a novel pipeline called Peer is your Pillar (PIP), which combines a target few-shot dataset with a peer dataset to create a data-unbalanced conditional generation. Our approach includes a class embedding method that separates the class space from the latent space, and we use a direction loss based on pre-trained CLIP to improve image diversity. Experiments on various few-shot datasets demonstrate the advancement of the proposed PIP, especially reduces the training requirements of few-shot image generation.

cs.CV↗
arXiv2023

Converse Flexoelectricity of Low-Dimensional Bismuth Selenite (Bi2Se3) Revealed by Piezoresponse Force Microscopy (PFM)

Many kinds of two-dimensional (2D) van der Waals (vdW) have been demonstrated to exhibit electromechanical coupling effects, which makes them promising candidates for next-generation devices, such as piezotronics and nanogenerators. Recently, flexoelectricity was found to account for the out-of-plane electromechanical coupling in many 2D transition metal dichalcogenides (TMDs) who only exhibit in-plane piezoelectricity. However, low dimensional vdW three-dimensional (3D) topological insulators (TIs) have been overlooked regarding their electromechanical properties. In this study, for the first time, we experimentally investigate the electromechanical coupling of low dimensional 3D TIs with a centrosymmetric crystal structure, where a binary compound, bismuth selenite (Bi2Se3), is taken as an example. The results of piezoresponse force microscope (PFM) tests on the Bi2Se3 nanoflakes show that the material exhibits both out-of-plane and in-plane electromechanical responses. The Bi2Se3 nanoflake with a thickness of 37 nm possesses an effective out-of-plane piezoelectric coefficient of ~0.65 pm V-1. With careful analyses, the electromechanical responses are verified to arise from the converse flexoelectricity. The measured effective out-of-plane piezoelectric coefficient is mainly contributed by flexoelectric coefficient, μ_39, which is estimated to be approximately 0.13 nC m-1. However, it is rather difficult to obtain the in-plane component of the flexoelectric tensor from the in-plane PFM measurements since the direction of the in-plane stress is always not normal to the AFM cantilever axis. The results provide useful guidance for understanding the flexoelectric effect of low dimensional vdW materials with centrosymmetric crystal structures. Moreover, the work can pave to way to explore the electromechanical devices based on the flexoelectricity of vdW TIs.

cond-mat.mes-hall↗
arXiv2023

A protein network refinement method based on module discovery and biological information

The identification of essential proteins can help in understanding the minimum requirements for cell survival and development. Network-based centrality approaches are commonly used to identify essential proteins from protein-protein interaction networks (PINs). Unfortunately, these approaches are limited by the poor quality of the underlying PIN data. To overcome this problem, researchers have focused on the prediction of essential proteins by combining PINs with other biological data. In this paper, we proposed a network refinement method based on module discovery and biological information to obtain a higher quality PIN. First, to extract the maximal connected subgraph in the PIN and to divide it into different modules by using Fast-unfolding algorithm; then, to detect critical modules based on the homology information, subcellular localization information and topology information within each module, and to construct a more refined network (CM-PIN). To evaluate the effectiveness of the proposed method, we used 10 typical network-based centrality methods (LAC, DC, DMNC, NC, TP, LID, CC, BC, PR, LR) to compare the overall performance of the CM-PIN with those the refined dynamic protein network (RD-PIN). The experimental results showed that the CM-PIN was optimal in terms of precision-recall curve, jackknife curve and other criteria, and can help to identify essential proteins more accurately.

cs.CE↗
arXiv2023

Time-Conditioned Generative Modeling of Object-Centric Representations for Video Decomposition and Prediction

When perceiving the world from multiple viewpoints, humans have the ability to reason about the complete objects in a compositional manner even when an object is completely occluded from certain viewpoints. Meanwhile, humans are able to imagine novel views after observing multiple viewpoints. Recent remarkable advances in multi-view object-centric learning still leaves some unresolved problems: 1) The shapes of partially or completely occluded objects can not be well reconstructed. 2) The novel viewpoint prediction depends on expensive viewpoint annotations rather than implicit rules in view representations. In this paper, we introduce a time-conditioned generative model for videos. To reconstruct the complete shape of an object accurately, we enhance the disentanglement between the latent representations of objects and views, where the latent representations of time-conditioned views are jointly inferred with a Transformer and then are input to a sequential extension of Slot Attention to learn object-centric representations. In addition, Gaussian processes are employed as priors of view latent variables for video generation and novel-view prediction without viewpoint annotations. Experiments on multiple datasets demonstrate that the proposed model can make object-centric video decomposition, reconstruct the complete shapes of occluded objects, and make novel-view predictions.

cs.CV↗
arXiv2023

Training-free Diffusion Model Adaptation for Variable-Sized Text-to-Image Synthesis

Diffusion models (DMs) have recently gained attention with state-of-the-art performance in text-to-image synthesis. Abiding by the tradition in deep learning, DMs are trained and evaluated on the images with fixed sizes. However, users are demanding for various images with specific sizes and various aspect ratio. This paper focuses on adapting text-to-image diffusion models to handle such variety while maintaining visual fidelity. First we observe that, during the synthesis, lower resolution images suffer from incomplete object portrayal, while higher resolution images exhibit repetitively disordered presentation. Next, we establish a statistical relationship indicating that attention entropy changes with token quantity, suggesting that models aggregate spatial information in proportion to image resolution. The subsequent interpretation on our observations is that objects are incompletely depicted due to limited spatial information for low resolutions, while repetitively disorganized presentation arises from redundant spatial information for high resolutions. From this perspective, we propose a scaling factor to alleviate the change of attention entropy and mitigate the defective pattern observed. Extensive experimental results validate the efficacy of the proposed scaling factor, enabling models to achieve better visual effects, image quality, and text alignment. Notably, these improvements are achieved without additional training or fine-tuning techniques.

cs.CV↗
arXiv2023

Adaptive Digital Twin for UAV-Assisted Integrated Sensing, Communication, and Computation Networks

In this paper, we study a digital twin (DT)-empowered integrated sensing, communication, and computation network. Specifically, the users perform radar sensing and computation offloading on the same spectrum, while unmanned aerial vehicles (UAVs) are deployed to provide edge computing service. We first formulate a multi-objective optimization problem to minimize the beampattern performance of multi-input multi-output (MIMO) radars and the computation offloading energy consumption simultaneously. Then, we explore the prediction capability of DT to provide intelligent offloading decision, where the DT estimation deviation is considered. To track this challenge, we reformulate the original problem as a multi-agent Markov decision process and design a multi-agent proximal policy optimization (MAPPO) framework to achieve a flexible learning policy. Furthermore, the Beta-policy and attention mechanism are used to improve the training performance. Numerical results show that the proposed method is able to balance the performance tradeoff between sensing and computation functions, while reducing the energy consumption compared with the existing studies.

eess.SP↗
arXiv2023

Large Language Models are Better Reasoners with Self-Verification

Recently, with the chain of thought (CoT) prompting, large language models (LLMs), e.g., GPT-3, have shown strong reasoning ability in several natural language processing tasks such as arithmetic, commonsense, and logical reasoning. However, LLMs with CoT require multi-step prompting and multi-token prediction, which is highly sensitive to individual mistakes and vulnerable to error accumulation. The above issues make the LLMs need the ability to verify the answers. In fact, after inferring conclusions in some thinking decision tasks, people often check them by re-verifying steps to avoid some mistakes. In this paper, we propose and prove that LLMs also have similar self-verification abilities. We take the conclusion obtained by CoT as one of the conditions for solving the original problem. By performing a backward verification of the answers that LLM deduced for itself, we can obtain interpretable answer validation scores to select the candidate answer with the highest score. Experimental results demonstrate that the proposed method can improve the reasoning performance on various arithmetic, commonsense, and logical reasoning datasets. Our code is publicly available at: https://github.com/WENGSYX/Self-Verification.

cs.AI↗
arXiv2023

Domain Re-Modulation for Few-Shot Generative Domain Adaptation

In this study, we delve into the task of few-shot Generative Domain Adaptation (GDA), which involves transferring a pre-trained generator from one domain to a new domain using only a few reference images. Inspired by the way human brains acquire knowledge in new domains, we present an innovative generator structure called Domain Re-Modulation (DoRM). DoRM not only meets the criteria of high quality, large synthesis diversity, and cross-domain consistency, which were achieved by previous research in GDA, but also incorporates memory and domain association, akin to how human brains operate. Specifically, DoRM freezes the source generator and introduces new mapping and affine modules (M&A modules) to capture the attributes of the target domain during GDA. This process resembles the formation of new synapses in human brains. Consequently, a linearly combinable domain shift occurs in the style space. By incorporating multiple new M&A modules, the generator gains the capability to perform high-fidelity multi-domain and hybrid-domain generation. Moreover, to maintain cross-domain consistency more effectively, we introduce a similarity-based structure loss. This loss aligns the auto-correlation map of the target image with its corresponding auto-correlation map of the source image during training. Through extensive experiments, we demonstrate the superior performance of our DoRM and similarity-based structure loss in few-shot GDA, both quantitatively and qualitatively. The code will be available at https://github.com/wuyi2020/DoRM.

cs.CV↗
arXiv2023

Evading Detection Actively: Toward Anti-Forensics against Forgery Localization

Anti-forensics seeks to eliminate or conceal traces of tampering artifacts. Typically, anti-forensic methods are designed to deceive binary detectors and persuade them to misjudge the authenticity of an image. However, to the best of our knowledge, no attempts have been made to deceive forgery detectors at the pixel level and mis-locate forged regions. Traditional adversarial attack methods cannot be directly used against forgery localization due to the following defects: 1) they tend to just naively induce the target forensic models to flip their pixel-level pristine or forged decisions; 2) their anti-forensics performance tends to be severely degraded when faced with the unseen forensic models; 3) they lose validity once the target forensic models are retrained with the anti-forensics images generated by them. To tackle the three defects, we propose SEAR (Self-supErvised Anti-foRensics), a novel self-supervised and adversarial training algorithm that effectively trains deep-learning anti-forensic models against forgery localization. SEAR sets a pretext task to reconstruct perturbation for self-supervised learning. In adversarial training, SEAR employs a forgery localization model as a supervisor to explore tampering features and constructs a deep-learning concealer to erase corresponding traces. We have conducted largescale experiments across diverse datasets. The experimental results demonstrate that, through the combination of self-supervised learning and adversarial learning, SEAR successfully deceives the state-of-the-art forgery localization methods, as well as tackle the three defects regarding traditional adversarial attack methods mentioned above.

cs.CV↗