arXiv subjects

Xu Yang

Publications and source records attributed to Xu Yang.

At least 271 records · Page 15Linked to original sources

arXiv2024

Exploring Learngene via Stage-wise Weight Sharing for Initializing Variable-sized Models

In practice, we usually need to build variable-sized models adapting for diverse resource constraints in different application scenarios, where weight initialization is an important step prior to training. The Learngene framework, introduced recently, firstly learns one compact part termed as learngene from a large well-trained model, after which learngene is expanded to initialize variable-sized models. In this paper, we start from analysing the importance of guidance for the expansion of well-trained learngene layers, inspiring the design of a simple but highly effective Learngene approach termed SWS (Stage-wise Weight Sharing), where both learngene layers and their learning process critically contribute to providing knowledge and guidance for initializing models at varying scales. Specifically, to learn learngene layers, we build an auxiliary model comprising multiple stages where the layer weights in each stage are shared, after which we train it through distillation. Subsequently, we expand these learngene layers containing stage information at their corresponding stage to initialize models of variable depths. Extensive experiments on ImageNet-1K demonstrate that SWS achieves consistent better performance compared to many models trained from scratch, while reducing around 6.6x total training costs. In some cases, SWS performs better only after 1 epoch tuning. When initializing variable-sized models adapting for different resource constraints, SWS achieves better results while reducing around 20x parameters stored to initialize these models and around 10x pre-training costs, in contrast to the pre-training and fine-tuning approach.

cs.LG↗

arXiv2024

FedMPQ: Secure and Communication-Efficient Federated Learning with Multi-codebook Product Quantization

In federated learning, particularly in cross-device scenarios, secure aggregation has recently gained popularity as it effectively defends against inference attacks by malicious aggregators. However, secure aggregation often requires additional communication overhead and can impede the convergence rate of the global model, which is particularly challenging in wireless network environments with extremely limited bandwidth. Therefore, achieving efficient communication compression under the premise of secure aggregation presents a highly challenging and valuable problem. In this work, we propose a novel uplink communication compression method for federated learning, named FedMPQ, which is based on multi shared codebook product quantization.Specifically, we utilize updates from the previous round to generate sufficiently robust codebooks. Secure aggregation is then achieved through trusted execution environments (TEE) or a trusted third party (TTP).In contrast to previous works, our approach exhibits greater robustness in scenarios where data is not independently and identically distributed (non-IID) and there is a lack of sufficient public data. The experiments conducted on the LEAF dataset demonstrate that our proposed method achieves 99% of the baseline's final accuracy, while reducing uplink communications by 90-95%

cs.CR↗

arXiv2024

Analysis of the Two-Step Heterogeneous Transfer Learning for Laryngeal Blood Vessel Classification: Issue and Improvement

Accurate classification of laryngeal vascular as benign or malignant is crucial for early detection of laryngeal cancer. However, organizations with limited access to laryngeal vascular images face challenges due to the lack of large and homogeneous public datasets for effective learning. Distinguished from the most familiar works, which directly transfer the ImageNet pre-trained models to the target domain for fine-tuning, this work pioneers exploring two-step heterogeneous transfer learning (THTL) for laryngeal lesion classification with nine deep-learning models, utilizing the diabetic retinopathy color fundus images, semantically non-identical yet vascular images, as the intermediate domain. Attention visualization technique, Layer Class Activate Map (LayerCAM), reveals a novel finding that yet the intermediate and the target domain both reflect vascular structure to a certain extent, the prevalent radial vascular pattern in the intermediate domain prevents learning the features of twisted and tangled vessels that distinguish the malignant class in the target domain, summarizes a vital rule for laryngeal lesion classification using THTL. To address this, we introduce an enhanced fine-tuning strategy in THTL called Step-Wise Fine-Tuning (SWFT) and apply it to the ResNet models. SWFT progressively refines model performance by accumulating fine-tuning layers from back to front, guided by the visualization results of LayerCAM. Comparison with the original THTL approach shows significant improvements. For ResNet18, the accuracy and malignant recall increases by 26.1% and 79.8%, respectively, while for ResNet50, these indicators improve by 20.4% and 62.2%, respectively.

cs.CV↗

arXiv2024

Texture-Preserving Diffusion Models for High-Fidelity Virtual Try-On

Image-based virtual try-on is an increasingly important task for online shopping. It aims to synthesize images of a specific person wearing a specified garment. Diffusion model-based approaches have recently become popular, as they are excellent at image synthesis tasks. However, these approaches usually employ additional image encoders and rely on the cross-attention mechanism for texture transfer from the garment to the person image, which affects the try-on's efficiency and fidelity. To address these issues, we propose an Texture-Preserving Diffusion (TPD) model for virtual try-on, which enhances the fidelity of the results and introduces no additional image encoders. Accordingly, we make contributions from two aspects. First, we propose to concatenate the masked person and reference garment images along the spatial dimension and utilize the resulting image as the input for the diffusion model's denoising UNet. This enables the original self-attention layers contained in the diffusion model to achieve efficient and accurate texture transfer. Second, we propose a novel diffusion-based method that predicts a precise inpainting mask based on the person and reference garment images, further enhancing the reliability of the try-on results. In addition, we integrate mask prediction and image synthesis into a single compact model. The experimental results show that our approach can be applied to various try-on tasks, e.g., garment-to-person and person-to-person try-ons, and significantly outperforms state-of-the-art methods on popular VITON, VITON-HD databases.

cs.CV↗

arXiv2024

Temporal Difference Learning for High-Dimensional PIDEs with Jumps

In this paper, we propose a deep learning framework for solving high-dimensional partial integro-differential equations (PIDEs) based on the temporal difference learning. We introduce a set of Levy processes and construct a corresponding reinforcement learning model. To simulate the entire process, we use deep neural networks to represent the solutions and non-local terms of the equations. Subsequently, we train the networks using the temporal difference error, termination condition, and properties of the non-local terms as the loss function. The relative error of the method reaches O(10^{-3}) in 100-dimensional experiments and O(10^{-4}) in one-dimensional pure jump problems. Additionally, our method demonstrates the advantages of low computational cost and robustness, making it well-suited for addressing problems with different forms and intensities of jumps.

math.NA↗

arXiv2024

MemoNav: Working Memory Model for Visual Navigation

Image-goal navigation is a challenging task that requires an agent to navigate to a goal indicated by an image in unfamiliar environments. Existing methods utilizing diverse scene memories suffer from inefficient exploration since they use all historical observations for decision-making without considering the goal-relevant fraction. To address this limitation, we present MemoNav, a novel memory model for image-goal navigation, which utilizes a working memory-inspired pipeline to improve navigation performance. Specifically, we employ three types of navigation memory. The node features on a map are stored in the short-term memory (STM), as these features are dynamically updated. A forgetting module then retains the informative STM fraction to increase efficiency. We also introduce long-term memory (LTM) to learn global scene representations by progressively aggregating STM features. Subsequently, a graph attention module encodes the retained STM and the LTM to generate working memory (WM) which contains the scene features essential for efficient navigation. The synergy among these three memory types boosts navigation performance by enabling the agent to learn and leverage goal-relevant scene features within a topological map. Our evaluation on multi-goal tasks demonstrates that MemoNav significantly outperforms previous methods across all difficulty levels in both Gibson and Matterport3D scenes. Qualitative results further illustrate that MemoNav plans more efficient routes.

cs.CV↗

arXiv2024

SecGraph: Towards SGX-based Efficient and Confidentiality-Preserving Graph Search

Graphs have more expressive power and are widely researched in various search demand scenarios, compared with traditional relational and XML models. Today, many graph search services have been deployed on a third-party server, which can alleviate users from the burdens of maintaining large-scale graphs and huge computation costs. Nevertheless, outsourcing graph search services to the third-party server may invade users' privacy. PeGraph was recently proposed to achieve the encrypted search over the social graph. The main idea of PeGraph is to maintain two data structures XSet and TSet motivated by the OXT technology to support encrypted conductive search. However, PeGraph still has some limitations. First, PeGraph suffers from high communication and computation costs in search operations. Second, PeGraph cannot support encrypted search over dynamic graphs. In this paper, we propose an SGX-based efficient and confidentiality-preserving graph search scheme SecGraph that can support insertion and deletion operations. We first design a new proxy-token generation method to reduce the communication cost. Then, we design an LDCF-encoded XSet based on the Logarithmic Dynamic Cuckoo Filter to reduce the computation cost. Finally, we design a new dynamic version of TSet named Twin-TSet to enable encrypted search over dynamic graphs. We have demonstrated the confidentiality preservation property of SecGraph through rigorous security analysis. Experiment results show that SecGraph yields up to 208x improvement in search time compared with PeGraph and the communication cost in PeGraph is up to 540x larger than that in SecGraph.

cs.CR↗

arXiv2024

Unfitted Spectral Element Method for interfacial models

In this paper, we propose the unfitted spectral element method for solving elliptic interface and corresponding eigenvalue problems. The novelty of the proposed method lies in its combination of the spectral accuracy of the spectral element method and the flexibility of the unfitted Nitsche's method. We also use tailored ghost penalty terms to enhance its robustness. We establish optimal $hp$ convergence rates for both elliptic interface problems and interface eigenvalue problems. Additionally, we demonstrate spectral accuracy for model problems in terms of polynomial degree.

math.NA↗

arXiv2024

DA-PFL: Dynamic Affinity Aggregation for Personalized Federated Learning

Personalized federated learning becomes a hot research topic that can learn a personalized learning model for each client. Existing personalized federated learning models prefer to aggregate similar clients with similar data distribution to improve the performance of learning models. However, similaritybased personalized federated learning methods may exacerbate the class imbalanced problem. In this paper, we propose a novel Dynamic Affinity-based Personalized Federated Learning model (DA-PFL) to alleviate the class imbalanced problem during federated learning. Specifically, we build an affinity metric from a complementary perspective to guide which clients should be aggregated. Then we design a dynamic aggregation strategy to dynamically aggregate clients based on the affinity metric in each round to reduce the class imbalanced risk. Extensive experiments show that the proposed DA-PFL model can significantly improve the accuracy of each client in three real-world datasets with state-of-the-art comparison methods.

cs.LG↗

arXiv2024

A Lightweight Inception Boosted U-Net Neural Network for Routability Prediction

As the modern CPU, GPU, and NPU chip design complexity and transistor counts keep increasing, and with the relentless shrinking of semiconductor technology nodes to nearly 1 nanometer, the placement and routing have gradually become the two most pivotal processes in modern very-large-scale-integrated (VLSI) circuit back-end design. How to evaluate routability efficiently and accurately in advance (at the placement and global routing stages) has grown into a crucial research area in the field of artificial intelligence (AI) assisted electronic design automation (EDA). In this paper, we propose a novel U-Net variant model boosted by an Inception embedded module to predict Routing Congestion (RC) and Design Rule Checking (DRC) hotspots. Experimental results on the recently published CircuitNet dataset benchmark show that our proposed method achieves up to 5% (RC) and 20% (DRC) rate reduction in terms of Avg-NRMSE (Average Normalized Root Mean Square Error) compared to the classic architecture. Furthermore, our approach consistently outperforms the prior model on the SSIM (Structural Similarity Index Measure) metric.

cs.AR↗

arXiv2024

Do You Guys Want to Dance: Zero-Shot Compositional Human Dance Generation with Multiple Persons

Human dance generation (HDG) aims to synthesize realistic videos from images and sequences of driving poses. Despite great success, existing methods are limited to generating videos of a single person with specific backgrounds, while the generalizability for real-world scenarios with multiple persons and complex backgrounds remains unclear. To systematically measure the generalizability of HDG models, we introduce a new task, dataset, and evaluation protocol of compositional human dance generation (cHDG). Evaluating the state-of-the-art methods on cHDG, we empirically find that they fail to generalize to real-world scenarios. To tackle the issue, we propose a novel zero-shot framework, dubbed MultiDance-Zero, that can synthesize videos consistent with arbitrary multiple persons and background while precisely following the driving poses. Specifically, in contrast to straightforward DDIM or null-text inversion, we first present a pose-aware inversion method to obtain the noisy latent code and initialization text embeddings, which can accurately reconstruct the composed reference image. Since directly generating videos from them will lead to severe appearance inconsistency, we propose a compositional augmentation strategy to generate augmented images and utilize them to optimize a set of generalizable text embeddings. In addition, consistency-guided sampling is elaborated to encourage the background and keypoints of the estimated clean image at each reverse step to be close to those of the reference image, further improving the temporal consistency of generated videos. Extensive qualitative and quantitative results demonstrate the effectiveness and superiority of our approach.

cs.CV↗

arXiv2024

Exploring Diverse In-Context Configurations for Image Captioning

After discovering that Language Models (LMs) can be good in-context few-shot learners, numerous strategies have been proposed to optimize in-context sequence configurations. Recently, researchers in Vision-Language (VL) domains also develop their few-shot learners, while they only use the simplest way, ie., randomly sampling, to configure in-context image-text pairs. In order to explore the effects of varying configurations on VL in-context learning, we devised four strategies for image selection and four for caption assignment to configure in-context image-text pairs for image captioning. Here Image Captioning is used as the case study since it can be seen as the visually-conditioned LM. Our comprehensive experiments yield two counter-intuitive but valuable insights, highlighting the distinct characteristics of VL in-context learning due to multi-modal synergy, as compared to the NLP case. Furthermore, in our exploration of optimal combination strategies, we observed an average performance enhancement of 20.9 of CIDEr scores compared to the baseline. The code is given in https://github.com/yongliang-wu/ExploreCfg.

cs.CV↗

arXiv2024

FedDCSR: Federated Cross-domain Sequential Recommendation via Disentangled Representation Learning

Cross-domain Sequential Recommendation (CSR) which leverages user sequence data from multiple domains has received extensive attention in recent years. However, the existing CSR methods require sharing origin user data across domains, which violates the General Data Protection Regulation (GDPR). Thus, it is necessary to combine federated learning (FL) and CSR to fully utilize knowledge from different domains while preserving data privacy. Nonetheless, the sequence feature heterogeneity across different domains significantly impacts the overall performance of FL. In this paper, we propose FedDCSR, a novel federated cross-domain sequential recommendation framework via disentangled representation learning. Specifically, to address the sequence feature heterogeneity across domains, we introduce an approach called inter-intra domain sequence representation disentanglement (SRD) to disentangle the user sequence features into domain-shared and domain-exclusive features. In addition, we design an intra domain contrastive infomax (CIM) strategy to learn richer domain-exclusive features of users by performing data augmentation on user sequences. Extensive experiments on three real-world scenarios demonstrate that FedDCSR achieves significant improvements over existing baselines.

cs.LG↗

arXiv2024

Capacity Results for Multiple-Input Multiple-Output Optical Wireless Communication With Per-Antenna Intensity Constraints

In this paper, we investigate the capacity of a multiple-input multiple-output (MIMO) optical intensity channel (OIC) under per-antenna peak- and average-intensity constraints. We first consider the case where the average intensities of input are required to be equal to preassigned constants due to the requirement of illumination quality and color temperature. When the channel graph of the MIMO OIC is strongly connected, we prove that the strongest eigen-subchannel must have positive channel gains, which simplifies the capacity analysis. Then we derive various capacity bounds by utilizing linear precoding, generalized entropy power inequality, and QR decomposition, etc. These bounds are numerically verified to approach the capacity in the low or high signal-to-noise ratio regime. Specifically, when the channel rank is one less than the number of transmit antennas, we derive an equivalent capacity expression from the perspective of convex geometry, and new lower bounds are derived based on this equivalent expression. Finally, the developed results are extended to the more general case where the average intensities of input are required to be no larger than preassigned constants.

cs.IT↗

arXiv2023

Pairing Symmetry and Fermion Projective Symmetry Groups

The Ginzburg-Landau (GL) theory is very successful in describing the pairing symmetry, a fundamental characterization of the broken symmetries in a paired superfluid or superconductor. However, GL theory does not describe fermionic excitations such as Bogoliubov quasiparticles or Andreev bound states that are directly related to topological properties of the superconductor. In this work, we show that the symmetries of the fermionic excitations are captured by a Projective Symmetry Group (PSG), which is a group extension of the bosonic symmetry group in the superconducting state. We further establish a correspondence between the pairing symmetry and the fermion PSG. When the normal and superconducting states share the same spin rotational symmetry, there is a simpler correspondence between the pairing symmetry and the fermion PSG, which we enumerate for all 32 crystalline point groups. We also discuss the general framework for computing PSGs when the spin rotational symmetry is spontaneously broken in the superconducting state. This PSG formalism leads to experimental consequences, and as an example, we show how a given pairing symmetry dictates the classification of topological superconductivity.

cond-mat.supr-con↗

arXiv2023

SeisT: A foundational deep learning model for earthquake monitoring tasks

Seismograms, the fundamental seismic records, have revolutionized earthquake research and monitoring. Recent advancements in deep learning have further enhanced seismic signal processing, leading to even more precise and effective earthquake monitoring capabilities. This paper introduces a foundational deep learning model, the Seismogram Transformer (SeisT), designed for a variety of earthquake monitoring tasks. SeisT combines multiple modules tailored to different tasks and exhibits impressive out-of-distribution generalization performance, outperforming or matching state-of-the-art models in tasks like earthquake detection, seismic phase picking, first-motion polarity classification, magnitude estimation, back-azimuth estimation, and epicentral distance estimation. The performance scores on the tasks are 0.96, 0.96, 0.68, 0.95, 0.86, 0.55, and 0.81, respectively. The most significant improvements, in comparison to existing models, are observed in phase-P picking, phase-S picking, and magnitude estimation, with gains of 1.7%, 9.5%, and 8.0%, respectively. Our study, through rigorous experiments and evaluations, suggests that SeisT has the potential to contribute to the advancement of seismic signal processing and earthquake research.

physics.geo-ph↗

arXiv2023

Transformer as Linear Expansion of Learngene

We propose expanding the shared Transformer module to produce and initialize Transformers of varying depths, enabling adaptation to diverse resource constraints. Drawing an analogy to genetic expansibility, we term such module as learngene. To identify the expansion mechanism, we delve into the relationship between the layer's position and its corresponding weight value, and find that linear function appropriately approximates this relationship. Building on this insight, we present Transformer as Linear Expansion of learnGene (TLEG), a novel approach for flexibly producing and initializing Transformers of diverse depths. Specifically, to learn learngene, we firstly construct an auxiliary Transformer linearly expanded from learngene, after which we train it through employing soft distillation. Subsequently, we can produce and initialize Transformers of varying depths via linearly expanding the well-trained learngene, thereby supporting diverse downstream scenarios. Extensive experiments on ImageNet-1K demonstrate that TLEG achieves comparable or better performance in contrast to many individual models trained from scratch, while reducing around 2x training cost. When transferring to several downstream classification datasets, TLEG surpasses existing initialization methods by a large margin (e.g., +6.87% on iNat 2019 and +7.66% on CIFAR-100). Under the situation where we need to produce models of varying depths adapting for different resource constraints, TLEG achieves comparable results while reducing around 19x parameters stored to initialize these models and around 5x pre-training costs, in contrast to the pre-training and fine-tuning approach. When transferring a fixed set of parameters to initialize different models, TLEG presents better flexibility and competitive performance while reducing around 2.9x parameters stored to initialize, compared to the pre-training approach.

cs.AI↗

arXiv2023

Building Variable-sized Models via Learngene Pool

Recently, Stitchable Neural Networks (SN-Net) is proposed to stitch some pre-trained networks for quickly building numerous networks with different complexity and performance trade-offs. In this way, the burdens of designing or training the variable-sized networks, which can be used in application scenarios with diverse resource constraints, are alleviated. However, SN-Net still faces a few challenges. 1) Stitching from multiple independently pre-trained anchors introduces high storage resource consumption. 2) SN-Net faces challenges to build smaller models for low resource constraints. 3). SN-Net uses an unlearned initialization method for stitch layers, limiting the final performance. To overcome these challenges, motivated by the recently proposed Learngene framework, we propose a novel method called Learngene Pool. Briefly, Learngene distills the critical knowledge from a large pre-trained model into a small part (termed as learngene) and then expands this small part into a few variable-sized models. In our proposed method, we distill one pretrained large model into multiple small models whose network blocks are used as learngene instances to construct the learngene pool. Since only one large model is used, we do not need to store more large models as SN-Net and after distilling, smaller learngene instances can be created to build small models to satisfy low resource constraints. We also insert learnable transformation matrices between the instances to stitch them into variable-sized models to improve the performance of these models. Exhaustive experiments have been implemented and the results validate the effectiveness of the proposed Learngene Pool compared with SN-Net.

cs.LG↗