arXiv subjects

Ying Zhang

Publications and source records attributed to Ying Zhang.

At least 271 records · Page 15Linked to original sources

arXiv2024

Observation of optical chaotic solitons and modulated subharmonic route to chaos in mode-locked laser

We reveal a new scenario for the transition of solitons to chaos in a mode-locked fiber laser: the modulated subharmonic route. Its universality is confirmed in two different laser configurations, namely, a figure-of-eight and a ring laser. Numerical simulations of the laser models agree well with the experiments. The modulated subharmonic route to chaos could stimulate parallel research in many nonlinear physical systems.

physics.optics↗

arXiv2024

Domain Generalization for Cross-Receiver Radio Frequency Fingerprint Identification

Radio Frequency Fingerprint Identification (RFFI) technology uniquely identifies emitters by analyzing unique distortions in the transmitted signal caused by non-ideal hardware. Recently, RFFI based on deep learning methods has gained popularity and is seen as a promising way to address the device authentication problem for Internet of Things (IoT) systems. However, in cross-receiver scenarios, where the RFFI model is trained over RF signals from some receivers but deployed at a new receiver, the alteration of receivers' characteristics would lead to data distribution shift and cause significant performance degradation at the new receiver. To address this problem, we first perform a theoretical analysis of the cross-receiver generalization error bound and propose a sufficient condition, named Separable Condition (SC), to minimize the classification error probability on the new receiver. Guided by the SC, a Receiver-Independent Emitter Identification (RIEI)model is devised to decouple the received signals into emitter-related features and receiver-related features and only the emitter-related features are used for identification. Furthermore, by leveraging federated learning, we also develop a FedRIEI model to eliminate the need for centralized collection of raw data from multiple receivers. Experiments on two real-world datasets demonstrate the superiority of our proposed methods over some baseline methods.

eess.SP↗

arXiv2024

One-dimension Periodic Potentials in Schrödinger Equation Solved by the Finite Difference Method

The one-dimensional Kronig-Penney potential in the Schrödinger equation, a standard periodic potential in quantum mechanics textbooks known for generating band structures, is solved by using the finite difference method with periodic boundary conditions. This method significantly improves the eigenvalue accuracy compared to existing approaches such as the filter method. The effects of the width and height of the Kronig-Penney potential on the eigenvalues and wave functions are then analyzed. As the potential height increases, the variation of eigenvalues with the wave vector slows down. Additionally, for higher-order band structures, the magnitude of the eigenvalue significantly decreases with increasing potential width. Finally, the Dirac comb potential, a periodic $δ$ potential, is examined using the present framework. This potential corresponds to the Kronig-Penney potential's width and height approaching zero and infinity, respectively. The numerical results obtained by the finite difference method for the Dirac comb potential are also perfectly consistent with the analytical solution.

quant-ph↗

arXiv2024

Terahertz semiconductor laser chaos

Chaos characterized by its irregularity and high sensitivity to initial conditions finds various applications in secure optical communications, random number generations, light detection and ranging systems, etc. Semiconductor lasers serve as ideal light platforms for chaos generations owing to the advantages in on-chip integration and complex nonlinear effects. In near-infrared wavelengths, semiconductor laser based chaotic light sources have been extensively studied and experimentally demonstrated. However, in the terahertz (THz) spectral range, due to the lack of effective THz light sources and high-speed detectors, chaos generation in THz semiconductor lasers, e.g., quantum cascade lasers (QCLs), is particularly challenging. Due to the fast intersubband carrier transitions, single mode THz QCLs resemble Class A lasers, where chaos can be hardly excited, even with external perturbations. In this work, we experimentally show a THz chaos source based on a sole multimode THz QCL without any external perturbations. Such a dynamical regime is characterized by the largest Lyapunov exponent associated to the temporal traces of the measured radio frequency (intermode beatnote) signal of the laser. The experimental results and chaos validation are confirmed by simulations of our model based on effective semiconductor Maxwell-Bloch Equations. To further understand the physical mechanism of the chaos generation in THz QCLs, a reduced model based on two coupled complex Ginzburg-Landau equations is derived from the full model cited above to systematically investigate the effects of the linewidth enhancement factor and group velocity dispersion on the chaotic regime. This model allows us to show that the chaos generation in the THz QCL can be ascribed to the system attaining the defect mediated turbulence regime.

physics.optics↗

arXiv2024

FastAttention: Extend FlashAttention2 to NPUs and Low-resource GPUs

FlashAttention series has been widely applied in the inference of large language models (LLMs). However, FlashAttention series only supports the high-level GPU architectures, e.g., Ampere and Hopper. At present, FlashAttention series is not easily transferrable to NPUs and low-resource GPUs. Moreover, FlashAttention series is inefficient for multi- NPUs or GPUs inference scenarios. In this work, we propose FastAttention which pioneers the adaptation of FlashAttention series for NPUs and low-resource GPUs to boost LLM inference efficiency. Specifically, we take Ascend NPUs and Volta-based GPUs as representatives for designing our FastAttention. We migrate FlashAttention series to Ascend NPUs by proposing a novel two-level tiling strategy for runtime speedup, tiling-mask strategy for memory saving and the tiling-AllReduce strategy for reducing communication overhead, respectively. Besides, we adapt FlashAttention for Volta-based GPUs by redesigning the operands layout in shared memory and introducing a simple yet effective CPU-GPU cooperative strategy for efficient memory utilization. On Ascend NPUs, our FastAttention can achieve a 10.7$\times$ speedup compared to the standard attention implementation. Llama-7B within FastAttention reaches up to 5.16$\times$ higher throughput than within the standard attention. On Volta architecture GPUs, FastAttention yields 1.43$\times$ speedup compared to its equivalents in \texttt{xformers}. Pangu-38B within FastAttention brings 1.46$\times$ end-to-end speedup using FasterTransformer. Coupled with the propose CPU-GPU cooperative strategy, FastAttention supports a maximal input length of 256K on 8 V100 GPUs. All the codes will be made available soon.

cs.LG↗

arXiv2024

GSSF: Generalized Structural Sparse Function for Deep Cross-modal Metric Learning

Cross-modal metric learning is a prominent research topic that bridges the semantic heterogeneity between vision and language. Existing methods frequently utilize simple cosine or complex distance metrics to transform the pairwise features into a similarity score, which suffers from an inadequate or inefficient capability for distance measurements. Consequently, we propose a Generalized Structural Sparse Function to dynamically capture thorough and powerful relationships across modalities for pair-wise similarity learning while remaining concise but efficient. Specifically, the distance metric delicately encapsulates two formats of diagonal and block-diagonal terms, automatically distinguishing and highlighting the cross-channel relevancy and dependency inside a structured and organized topology. Hence, it thereby empowers itself to adapt to the optimal matching patterns between the paired features and reaches a sweet spot between model complexity and capability. Extensive experiments on cross-modal and two extra uni-modal retrieval tasks (image-text retrieval, person re-identification, fine-grained image retrieval) have validated its superiority and flexibility over various popular retrieval frameworks. More importantly, we further discover that it can be seamlessly incorporated into multiple application scenarios, and demonstrates promising prospects from Attention Mechanism to Knowledge Distillation in a plug-and-play manner. Our code is publicly available at: https://github.com/Paranioar/GSSF.

cs.CV↗

arXiv2024

Bias Behind the Wheel: Fairness Testing of Autonomous Driving Systems

This paper conducts fairness testing of automated pedestrian detection, a crucial but under-explored issue in autonomous driving systems. We evaluate eight state-of-the-art deep learning-based pedestrian detectors across demographic groups on large-scale real-world datasets. To enable thorough fairness testing, we provide extensive annotations for the datasets, resulting in 8,311 images with 16,070 gender labels, 20,115 age labels, and 3,513 skin tone labels. Our findings reveal significant fairness issues, particularly related to age. The proportion of undetected children is 20.14% higher compared to adults. Furthermore, we explore how various driving scenarios affect the fairness of pedestrian detectors. We find that pedestrian detectors demonstrate significant gender biases during night time, potentially exacerbating the prevalent societal issue of female safety concerns during nighttime out. Moreover, we observe that pedestrian detectors can demonstrate both enhanced fairness and superior performance under specific driving conditions, which challenges the fairness-performance trade-off theory widely acknowledged in the fairness literature. We publicly release the code, data, and results to support future research on fairness in autonomous driving.

cs.CY↗

arXiv2024

Reconsidering Degeneration of Token Embeddings with Definitions for Encoder-based Pre-trained Language Models

Learning token embeddings based on token co-occurrence statistics has proven effective for both pre-training and fine-tuning in natural language processing. However, recent studies have pointed out that the distribution of learned embeddings degenerates into anisotropy (i.e., non-uniform distribution), and even pre-trained language models (PLMs) suffer from a loss of semantics-related information in embeddings for low-frequency tokens. This study first analyzes the fine-tuning dynamics of encoder-based PLMs and demonstrates their robustness against degeneration. On the basis of this analysis, we propose DefinitionEMB, a method that utilizes definitions to re-construct isotropically distributed and semantics-related token embeddings for encoder-based PLMs while maintaining original robustness during fine-tuning. Our experiments demonstrate the effectiveness of leveraging definitions from Wiktionary to re-construct such embeddings for two encoder-based PLMs: RoBERTa-base and BART-large. Furthermore, the re-constructed embeddings for low-frequency tokens improve the performance of these models across various GLUE and four text summarization datasets.

cs.CL↗

arXiv2024

Bridging Large Language Models and Graph Structure Learning Models for Robust Representation Learning

Graph representation learning, involving both node features and graph structures, is crucial for real-world applications but often encounters pervasive noise. State-of-the-art methods typically address noise by focusing separately on node features with large language models (LLMs) and on graph structures with graph structure learning models (GSLMs). In this paper, we introduce LangGSL, a robust framework that integrates the complementary strengths of pre-trained language models and GSLMs to jointly enhance both node feature and graph structure learning. In LangGSL, we first leverage LLMs to filter noise in the raw data and extract valuable cleaned information as features, enhancing the synergy of downstream models. During the mutual learning phase in LangGSL, the core idea is to leverage the relatively small language model (LM) to process local attributes and generate reliable pseudo-labels and informative node embeddings, which are then integrated into the GSLM's prediction phase. This approach enriches the global context and enhances overall performance. Meanwhile, GSLM refines the evolving graph structure constructed from the LM's output, offering updated labels back to the LM as additional guidance, thus facilitating a more effective mutual learning process. The LM and GSLM work synergistically, complementing each other's strengths and offsetting weaknesses within a variational information-maximizing framework, resulting in enhanced node features and a more robust graph structure. Extensive experiments on diverse graph datasets of varying scales and across different task scenarios demonstrate the scalability and effectiveness of the proposed approach.

cs.LG↗

arXiv2024

CLIP-SCGI: Synthesized Caption-Guided Inversion for Person Re-Identification

Person re-identification (ReID) has recently benefited from large pretrained vision-language models such as Contrastive Language-Image Pre-Training (CLIP). However, the absence of concrete descriptions necessitates the use of implicit text embeddings, which demand complicated and inefficient training strategies. To address this issue, we first propose one straightforward solution by leveraging existing image captioning models to generate pseudo captions for person images, and thereby boost person re-identification with large vision language models. Using models like the Large Language and Vision Assistant (LLAVA), we generate high-quality captions based on fixed templates that capture key semantic attributes such as gender, clothing, and age. By augmenting ReID training sets from uni-modality (image) to bi-modality (image and text), we introduce CLIP-SCGI, a simple yet effective framework that leverages synthesized captions to guide the learning of discriminative and robust representations. Built on CLIP, CLIP-SCGI fuses image and text embeddings through two modules to enhance the training process. To address quality issues in generated captions, we introduce a caption-guided inversion module that captures semantic attributes from images by converting relevant visual information into pseudo-word tokens based on the descriptions. This approach helps the model better capture key information and focus on relevant regions. The extracted features are then utilized in a cross-modal fusion module, guiding the model to focus on regions semantically consistent with the caption, thereby facilitating the optimization of the visual encoder to extract discriminative and robust representations. Extensive experiments on four popular ReID benchmarks demonstrate that CLIP-SCGI outperforms the state-of-the-art by a significant margin.

cs.CV↗

arXiv2024

TCGU: Data-centric Graph Unlearning based on Transferable Condensation

With growing demands for data privacy and model robustness, graph unlearning (GU), which erases the influence of specific data on trained GNN models, has gained significant attention. However, existing exact unlearning methods suffer from either low efficiency or poor model performance. While being more utility-preserving and efficient, current approximate unlearning methods are not applicable in the zero-glance privacy setting, where the deleted samples cannot be accessed during unlearning due to immediate deletion requested by regulations. Besides, these approximate methods, which try to directly perturb model parameters still involve high privacy concerns in practice. To fill the gap, we propose Transferable Condensation Graph Unlearning (TCGU), a data-centric solution to zero-glance graph unlearning. Specifically, we first design a two-level alignment strategy to pre-condense the original graph into a small yet utility-preserving dataset. Upon receiving an unlearning request, we fine-tune the pre-condensed data with a low-rank plugin, to directly align its distribution with the remaining graph, thus efficiently revoking the information of deleted data without accessing them. A novel similarity distribution matching approach and a discrimination regularizer are proposed to effectively transfer condensed data and preserve its utility in GNN training, respectively. Finally, we retrain the GNN on the transferred condensed data. Extensive experiments on 6 benchmark datasets demonstrate that TCGU can achieve superior performance in terms of model utility, unlearning efficiency, and unlearning efficacy than existing GU methods.

cs.LG↗

arXiv2024

Exploring LLM-based Data Annotation Strategies for Medical Dialogue Preference Alignment

This research examines the use of Reinforcement Learning from AI Feedback (RLAIF) techniques to improve healthcare dialogue models, with the aim of tackling the challenges of preference-aligned data annotation while reducing the reliance on medical experts. We argue that the primary challenges in current RLAIF research for healthcare are the limitations of automated evaluation methods and the difficulties in accurately representing physician preferences. To address these challenges, we present a new evaluation framework based on standardized patient examinations. This framework is designed to objectively assess the effectiveness of large language models (LLMs) in guiding users and following instructions, enabling a comprehensive comparison across different models. Furthermore, our investigation of effective ways to express physician preferences using Constitutional AI algorithms highlighted the particular effectiveness of flowcharts. Utilizing this finding, we introduce an innovative agent-based approach for annotating preference data. This approach autonomously creates medical dialogue flows tailored to the patient's condition, demonstrates strong generalization abilities, and reduces the need for expert involvement. Our results show that the agent-based approach outperforms existing RLAIF annotation methods in standardized patient examinations and surpasses current open source medical dialogue LLMs in various test scenarios.

cs.CL↗

arXiv2024

The comparison of the state-of-the-art nucleon-nucleon potentials from phase shift to nuclear matter

The nucleon-nucleon ($NN$) potential is the residual interaction of the strong interaction in the low-energy region and is also the fundamental input to the study of atomic nuclei. Based on the non-perturbative properties of the quantum chromodynamics (QCD), $NN$ potential is not yet directly accessible from QCD theory. Therefore, various models of $NN$ interactions have been constructed based on Yukawa's meson exchange pictures since the 1930s, including one-boson-exchange models, coordinate operator models and chiral effective field models. Analysis of extensive $NN$ scattering data has shown that the two-body nuclear force exhibits a short-range repulsion and intermediate-range attraction, and decays rapidly with increasing distance. A series of charge-dependent high-precision $NN$ interactions have been further developed in the past thirty years, such as the AV18 potential, CD-Bonn potential, pvCD-Bonn potentials, and the chiral effective nuclear potentials with momentum expansion up to the fifth order. In this work, the phase shifts at different channels, the cross sections, the entanglement entropy in spin space, and the equations of state of symmetric nuclear matter and pure neutron matter from these high-precision $NN$ interactions are calculated and systematically compared. It can be found that they have significant differences in the cases with high angular momentum, high laboratory energy, and high-density regions.

nucl-th↗

arXiv2024

AlterMOMA: Fusion Redundancy Pruning for Camera-LiDAR Fusion Models with Alternative Modality Masking

Camera-LiDAR fusion models significantly enhance perception performance in autonomous driving. The fusion mechanism leverages the strengths of each modality while minimizing their weaknesses. Moreover, in practice, camera-LiDAR fusion models utilize pre-trained backbones for efficient training. However, we argue that directly loading single-modal pre-trained camera and LiDAR backbones into camera-LiDAR fusion models introduces similar feature redundancy across modalities due to the nature of the fusion mechanism. Unfortunately, existing pruning methods are developed explicitly for single-modal models, and thus, they struggle to effectively identify these specific redundant parameters in camera-LiDAR fusion models. In this paper, to address the issue above on camera-LiDAR fusion models, we propose a novelty pruning framework Alternative Modality Masking Pruning (AlterMOMA), which employs alternative masking on each modality and identifies the redundant parameters. Specifically, when one modality parameters are masked (deactivated), the absence of features from the masked backbone compels the model to reactivate previous redundant features of the other modality backbone. Therefore, these redundant features and relevant redundant parameters can be identified via the reactivation process. The redundant parameters can be pruned by our proposed importance score evaluation function, Alternative Evaluation (AlterEva), which is based on the observation of the loss changes when certain modality parameters are activated and deactivated. Extensive experiments on the nuScene and KITTI datasets encompassing diverse tasks, baseline models, and pruning algorithms showcase that AlterMOMA outperforms existing pruning methods, attaining state-of-the-art performance.

cs.CV↗

arXiv2024

Rail-only: A Low-Cost High-Performance Network for Training LLMs with Trillion Parameters

This paper presents a low-cost network architecture for training large language models (LLMs) at hyperscale. We study the optimal parallelization strategy of LLMs and propose a novel datacenter network design tailored to LLM's unique communication pattern. We show that LLM training generates sparse communication patterns in the network and, therefore, does not require any-to-any full-bisection network to complete efficiently. As a result, our design eliminates the spine layer in traditional GPU clusters. We name this design a Rail-only network and demonstrate that it achieves the same training performance while reducing the network cost by 38% to 77% and network power consumption by 37% to 75% compared to a conventional GPU datacenter. Our architecture also supports Mixture-of-Expert (MoE) models with all-to-all communication through forwarding, with only 8.2% to 11.2% completion time overhead for all-to-all traffic. We study the failure robustness of Rail-only networks and provide insights into the performance impact of different network and training parameters.

cs.NI↗

arXiv2024

Edge Computing for IoT: Novel Insights from a Comparative Analysis of Access Control Models

IoT edge computing positions computing resources closer to the data sources to reduce the latency, relieve the bandwidth pressure on the cloud, and enhance data security. Nevertheless, data security in IoT edge computing still faces critical threats (e.g., data breaches). Access control is fundamental for mitigating these threats. However, IoT edge computing introduces notable challenges for achieving resource-conserving, low-latency, flexible, and scalable access control. To review recent access control measures, we novelly organize them according to different data lifecycles--data collection, storage, and usage--and, meanwhile, review blockchain technology in this novel organization. In this way, we provide novel insights and envisage several potential research directions. This survey can help readers find gaps systematically and prompt the development of access control techniques in IoT edge computing under the intricacy of innovations in access control.

eess.SY↗

arXiv2024

Can We Leave Deepfake Data Behind in Training Deepfake Detector?

The generalization ability of deepfake detectors is vital for their applications in real-world scenarios. One effective solution to enhance this ability is to train the models with manually-blended data, which we termed "blendfake", encouraging models to learn generic forgery artifacts like blending boundary. Interestingly, current SoTA methods utilize blendfake without incorporating any deepfake data in their training process. This is likely because previous empirical observations suggest that vanilla hybrid training (VHT), which combines deepfake and blendfake data, results in inferior performance to methods using only blendfake data (so-called "1+1<2"). Therefore, a critical question arises: Can we leave deepfake behind and rely solely on blendfake data to train an effective deepfake detector? Intuitively, as deepfakes also contain additional informative forgery clues (e.g., deep generative artifacts), excluding all deepfake data in training deepfake detectors seems counter-intuitive. In this paper, we rethink the role of blendfake in detecting deepfakes and formulate the process from "real to blendfake to deepfake" to be a progressive transition. Specifically, blendfake and deepfake can be explicitly delineated as the oriented pivot anchors between "real-to-fake" transitions. The accumulation of forgery information should be oriented and progressively increasing during this transition process. To this end, we propose an Oriented Progressive Regularizor (OPR) to establish the constraints that compel the distribution of anchors to be discretely arranged. Furthermore, we introduce feature bridging to facilitate the smooth transition between adjacent anchors. Extensive experiments confirm that our design allows leveraging forgery information from both blendfake and deepfake effectively and comprehensively.

cs.CV↗

arXiv2024

Mumpy: Multilateral Temporal-view Pyramid Transformer for Video Inpainting Detection

The task of video inpainting detection is to expose the pixel-level inpainted regions within a video sequence. Existing methods usually focus on leveraging spatial and temporal inconsistencies. However, these methods typically employ fixed operations to combine spatial and temporal clues, limiting their applicability in different scenarios. In this paper, we introduce a novel Multilateral Temporal-view Pyramid Transformer ({\em MumPy}) that collaborates spatial-temporal clues flexibly. Our method utilizes a newly designed multilateral temporal-view encoder to extract various collaborations of spatial-temporal clues and introduces a deformable window-based temporal-view interaction module to enhance the diversity of these collaborations. Subsequently, we develop a multi-pyramid decoder to aggregate the various types of features and generate detection maps. By adjusting the contribution strength of spatial and temporal clues, our method can effectively identify inpainted regions. We validate our method on existing datasets and also introduce a new challenging and large-scale Video Inpainting dataset based on the YouTube-VOS dataset, which employs several more recent inpainting methods. The results demonstrate the superiority of our method in both in-domain and cross-domain evaluation scenarios.

cs.CV↗