arXiv subjects

Dong Li

Publications and source records attributed to Dong Li.

At least 307 records · Page 17Linked to original sources

arXiv2024

Optimizing Logical Execution Time Model for Both Determinism and Low Latency

The Logical Execution Time (LET) programming model has recently received considerable attention, particularly because of its timing and dataflow determinism. In LET, task computation appears always to take the same amount of time (called the task's LET interval), and the task reads (resp. writes) at the beginning (resp. end) of the interval. Compared to other communication mechanisms, such as implicit communication and Dynamic Buffer Protocol (DBP), LET performs worse on many metrics, such as end-to-end latency (including reaction time and data age) and time disparity jitter. Compared with the default LET setting, the flexible LET (fLET) model shrinks the LET interval while still guaranteeing schedulability by introducing the virtual offset to defer the read operation and using the virtual deadline to move up the write operation. Therefore, fLET has the potential to significantly improve the end-to-end timing performance while keeping the benefits of deterministic behavior on timing and dataflow. To fully realize the potential of fLET, we consider the problem of optimizing the assignments of its virtual offsets and deadlines. We propose new abstractions to describe the task communication pattern and new optimization algorithms to explore the solution space efficiently. The algorithms leverage the linearizability of communication patterns and utilize symbolic operations to achieve efficient optimization while providing a theoretical guarantee. The framework supports optimizing multiple performance metrics and guarantees bounded suboptimality when optimizing end-to-end latency. Experimental results show that our optimization algorithms improve upon the default LET and its existing extensions and significantly outperform implicit communication and DBP in terms of various metrics, such as end-to-end latency, time disparity, and its jitter.

eess.SY↗

arXiv2024

Contrastive Augmented Graph2Graph Memory Interaction for Few Shot Continual Learning

Few-Shot Class-Incremental Learning (FSCIL) has gained considerable attention in recent years for its pivotal role in addressing continuously arriving classes. However, it encounters additional challenges. The scarcity of samples in new sessions intensifies overfitting, causing incompatibility between the output features of new and old classes, thereby escalating catastrophic forgetting. A prevalent strategy involves mitigating catastrophic forgetting through the Explicit Memory (EM), which comprise of class prototypes. However, current EM-based methods retrieves memory globally by performing Vector-to-Vector (V2V) interaction between features corresponding to the input and prototypes stored in EM, neglecting the geometric structure of local features. This hinders the accurate modeling of their positional relationships. To incorporate information of local geometric structure, we extend the V2V interaction to Graph-to-Graph (G2G) interaction. For enhancing local structures for better G2G alignment and the prevention of local feature collapse, we propose the Local Graph Preservation (LGP) mechanism. Additionally, to address sample scarcity in classes from new sessions, the Contrast-Augmented G2G (CAG2G) is introduced to promote the aggregation of same class features thus helps few-shot learning. Extensive comparisons on CIFAR100, CUB200, and the challenging ImageNet-R dataset demonstrate the superiority of our method over existing methods.

cs.AI↗

arXiv2024

CO2: Efficient Distributed Training with Full Communication-Computation Overlap

The fundamental success of large language models hinges upon the efficacious implementation of large-scale distributed training techniques. Nevertheless, building a vast, high-performance cluster featuring high-speed communication interconnectivity is prohibitively costly, and accessible only to prominent entities. In this work, we aim to lower this barrier and democratize large-scale training with limited bandwidth clusters. We propose a new approach called CO2 that introduces local-updating and asynchronous communication to the distributed data-parallel training, thereby facilitating the full overlap of COmunication with COmputation. CO2 is able to attain a high scalability even on extensive multi-node clusters constrained by very limited communication bandwidth. We further propose the staleness gap penalty and outer momentum clipping techniques together with CO2 to bolster its convergence and training stability. Besides, CO2 exhibits seamless integration with well-established ZeRO-series optimizers which mitigate memory consumption of model states with large model training. We also provide a mathematical proof of convergence, accompanied by the establishment of a stringent upper bound. Furthermore, we validate our findings through an extensive set of practical experiments encompassing a wide range of tasks in the fields of computer vision and natural language processing. These experiments serve to demonstrate the capabilities of CO2 in terms of convergence, generalization, and scalability when deployed across configurations comprising up to 128 A100 GPUs. The outcomes emphasize the outstanding capacity of CO2 to hugely improve scalability, no matter on clusters with 800Gbps RDMA or 80Gbps TCP/IP inter-node connections.

cs.CL↗

arXiv2024

Real-Time Systems Optimization with Black-box Constraints and Hybrid Variables

When optimizing real-time systems, designers often face a challenging problem where the schedulability constraints are non-convex, non-continuous, or lack an analytical form to understand their properties. Although the optimization framework NORTH proposed in previous work is general (it works with arbitrary schedulability analysis) and scalable, it can only handle problems with continuous variables, which limits its application. In this paper, we extend the applications of the framework NORTH to problems with a hybrid of continuous and discrete variables. This is achieved in a coordinate-descent method, where the continuous and discrete variables are optimized separately during iterations. The new framework, NORTH+, improves around 20% solution quality than NORTH in experiments.

eess.SY↗

arXiv2024

TransNormerLLM: A Faster and Better Large Language Model with Improved TransNormer

We present TransNormerLLM, the first linear attention-based Large Language Model (LLM) that outperforms conventional softmax attention-based models in terms of both accuracy and efficiency. TransNormerLLM evolves from the previous linear attention architecture TransNormer by making advanced modifications that include positional embedding, linear attention acceleration, gating mechanisms, tensor normalization, and inference acceleration and stabilization. Specifically, we use LRPE together with an exponential decay to avoid attention dilution issues while allowing the model to retain global interactions between tokens. Additionally, we propose Lightning Attention, a cutting-edge technique that accelerates linear attention by more than twice in runtime and reduces memory usage by a remarkable four times. To further enhance the performance of TransNormer, we leverage a gating mechanism for smooth training and a new tensor normalization scheme to accelerate the model, resulting in an impressive acceleration of over $20\%$. Furthermore, we develop a robust inference algorithm that ensures numerical stability and consistent inference speed, regardless of the sequence length, showcasing superior efficiency during both training and inference stages. We also implement an efficient model parallel schema for TransNormerLLM, enabling seamless deployment on large-scale clusters and facilitating expansion to even more extensive models, i.e., LLMs with 175B parameters. We validate our model design through a series of ablations and train models with sizes of 385M, 1B, and 7B on our self-collected corpus. Benchmark results demonstrate that our models not only match the performance of state-of-the-art LLMs with Transformer but are also significantly faster. Code is released at: https://github.com/OpenNLPLab/TransnormerLLM.

cs.CL↗

arXiv2024

Lightning Attention-2: A Free Lunch for Handling Unlimited Sequence Lengths in Large Language Models

Linear attention is an efficient attention mechanism that has recently emerged as a promising alternative to conventional softmax attention. With its ability to process tokens in linear computational complexities, linear attention, in theory, can handle sequences of unlimited length without sacrificing speed, i.e., maintaining a constant training speed for various sequence lengths with a fixed memory consumption. However, due to the issue with cumulative summation (cumsum), current linear attention algorithms cannot demonstrate their theoretical advantage in a causal setting. In this paper, we present Lightning Attention-2, the first linear attention implementation that enables linear attention to realize its theoretical computational benefits. To achieve this, we leverage the thought of tiling, separately handling the intra-block and inter-block components in linear attention calculation. Specifically, we utilize the conventional attention computation mechanism for the intra-blocks and apply linear attention kernel tricks for the inter-blocks. A tiling technique is adopted through both forward and backward procedures to take full advantage of the GPU hardware. We implement our algorithm in Triton to make it IO-aware and hardware-friendly. Various experiments are conducted on different model sizes and sequence lengths. Lightning Attention-2 retains consistent training and inference speed regardless of input sequence length and is significantly faster than other attention mechanisms. The source code is available at https://github.com/OpenNLPLab/lightning-attention.

cs.CL↗

arXiv2024

A Statistical Study of Solar White-Light Flares Observed by the White-light Solar Telescope of the Lyman-alpha Solar Telescope on the Advanced Space-based Solar Observatory (ASO-S/LST/WST) at 360 nm

Solar white-light flares (WLFs) are those accompanied by brightenings in the optical continuum or integrated light. The White-light Solar Telescope (WST), as an instrument of the Lyman-alpha Solar Telescope (LST) on the Advanced Space-based Solar Observatory (ASO-S), provides continuous solar full-disk images at 360 nm, which can be used to study WLFs. We analyze 205 major flares above M1.0 from October 2022 to May 2023 and identify 49 WLFs at 360 nm from WST observations, i.e. with an occurrence rate of 23.9%. The percentages of WLFs for M1 - M4 (31 out of 180), M5 - M9 (11 out of 18), and above X1 (7 for all) flares are 17.2%, 61.1%, and 100%, respectively, namely the larger the flares, the more likely they are WLFs at 360 nm. We further analyze 39 WLFs among the identified WLFs and investigate their properties such as white-light enhancement, duration, and brightening area. It is found that the relative enhancement of the white-light emission at 360 nm is mostly (>90%) less than 30% and the mean enhancement is 19.4%. The WLFs' duration at 360 nm is mostly (>80%) less than 20 minutes and its mean is 10.3 minutes. The brightening area at 360 nm is mostly (>75%) less than 500 arcsecond2 and the median value is 225. We find that there exist good correlations between the white-light enhancement/duration/area and the peak soft X-ray (SXR) flux of the flare, with correlation coefficients of 0.68, 0.58, and 0.80, respectively. In addition, the white-light emission in most WLFs peaks around the same time as the temporal derivative of SXR flux as well as the hard X-ray emission at 20 - 50 keV, indicative of Neupert effect. It is also found that the limb WLFs are more likely to have a greater enhancement, which is consistent with numerical simulations.

astro-ph.SR↗

arXiv2024

Operational Research: Methods and Applications

Throughout its history, Operational Research has evolved to include a variety of methods, models and algorithms that have been applied to a diverse and wide range of contexts. This encyclopedic article consists of two main sections: methods and applications. The first aims to summarise the up-to-date knowledge and provide an overview of the state-of-the-art methods and key developments in the various subdomains of the field. The second offers a wide-ranging list of areas where Operational Research has been applied. The article is meant to be read in a nonlinear fashion. It should be used as a point of reference or first-port-of-call for a diverse pool of readers: academics, researchers, students, and practitioners. The entries within the methods and applications sections are presented in alphabetical order. The authors dedicate this paper to the 2023 Turkey/Syria earthquake victims. We sincerely hope that advances in OR will play a role towards minimising the pain and suffering caused by this and future catastrophes.

math.OC↗

arXiv2024

UPDP: A Unified Progressive Depth Pruner for CNN and Vision Transformer

Traditional channel-wise pruning methods by reducing network channels struggle to effectively prune efficient CNN models with depth-wise convolutional layers and certain efficient modules, such as popular inverted residual blocks. Prior depth pruning methods by reducing network depths are not suitable for pruning some efficient models due to the existence of some normalization layers. Moreover, finetuning subnet by directly removing activation layers would corrupt the original model weights, hindering the pruned model from achieving high performance. To address these issues, we propose a novel depth pruning method for efficient models. Our approach proposes a novel block pruning strategy and progressive training method for the subnet. Additionally, we extend our pruning method to vision transformer models. Experimental results demonstrate that our method consistently outperforms existing depth pruning methods across various pruning configurations. We obtained three pruned ConvNeXtV1 models with our method applying on ConvNeXtV1, which surpass most SOTA efficient models with comparable inference performance. Our method also achieves state-of-the-art pruning performance on the vision transformer model.

cs.CV↗

arXiv2024

Breaking the Interference and Fading Gridlock in Backscatter Communications: State-of-the-Art, Design Challenges, and Future Directions

As the Internet of Things (IoT) advances by leaps and bounds, a multitude of devices are becoming interconnected, marking the onset of an era where all things are connected. While this growth opens up opportunities for novel products and applications, it also leads to increased energy demand and battery reliance for IoT devices, creating a significant bottleneck that hinders sustainable progress. At this juncture, backscatter communication (BackCom), as a low-power and passive communication method, emerges as one of the promising solutions to this energy impasse by reducing the manufacturing costs and energy consumption of IoT devices. However, BackCom systems face challenges such as complex interference environments, including direct link interference (DLI) and mutual interference (MI) between tags, which can severely disrupt the efficiency of BackCom networks. Moreover, double-path fading is another major issue that leads to the degraded system performance. To fully unleash the potential of BackComs, the purpose of this paper is to furnish a comprehensive review of existing solutions with a focus on combatting these specific interference challenges and overcoming dual-path fading, offering an insightful analysis and comparison of various strategies for effectively mitigating these issues. Specifically, we begin by introducing the preliminaries for the BackCom, including its history, operating mechanisms, main architectures, etc, providing a foundational understanding of the field. Then, we delve into fundamental issues related to BackCom systems, such as solutions for the DLI, the MI, and the double-path fading. This paper thoroughly provides state-of-the-art advances for each case, particularly highlighting how the latest innovations in theoretical approaches and system design can strategically address these challenges.

cs.IT↗

arXiv2023

PG-LBO: Enhancing High-Dimensional Bayesian Optimization with Pseudo-Label and Gaussian Process Guidance

Variational Autoencoder based Bayesian Optimization (VAE-BO) has demonstrated its excellent performance in addressing high-dimensional structured optimization problems. However, current mainstream methods overlook the potential of utilizing a pool of unlabeled data to construct the latent space, while only concentrating on designing sophisticated models to leverage the labeled data. Despite their effective usage of labeled data, these methods often require extra network structures, additional procedure, resulting in computational inefficiency. To address this issue, we propose a novel method to effectively utilize unlabeled data with the guidance of labeled data. Specifically, we tailor the pseudo-labeling technique from semi-supervised learning to explicitly reveal the relative magnitudes of optimization objective values hidden within the unlabeled data. Based on this technique, we assign appropriate training weights to unlabeled data to enhance the construction of a discriminative latent space. Furthermore, we treat the VAE encoder and the Gaussian Process (GP) in Bayesian optimization as a unified deep kernel learning process, allowing the direct utilization of labeled data, which we term as Gaussian Process guidance. This directly and effectively integrates the goal of improving GP accuracy into the VAE training, thereby guiding the construction of the latent space. The extensive experiments demonstrate that our proposed method outperforms existing VAE-BO algorithms in various optimization scenarios. Our code will be published at https://github.com/TaicaiChen/PG-LBO.

cs.LG↗

arXiv2023

Secure-ISAC: Secure V2X Communication: An Integrated Sensing and Communication Perspective

In Vehicle-to-Everything (V2X) systems, reliable and secure information exchange plays a pivotal role in road safety and traffic management. Due to the open nature of the wireless medium and the constant or intermittent mobility of vehicles, the security of transmissions in V2X is more challenging compared to traditional wireless networks. Physical layer security (PLS) leverages the inherent randomness of wireless communication channels to ensure the confidentiality and security of information transmission. Current PLS schemes in integrated communications and sensing (ISAC) enabled V2X systems is to utilize communication interference to significantly impact the eavesdropping channel more than the legitimate channel. However, in an ISAC-enabled V2X system, it is crucial to prioritize and address the issue of interference coupling as it significantly impacts the confidentiality and security of information exchange. This goes beyond simply relying on the communication interference. Until now, no discussions or studies on integrating security with ISAC (Seucue-ISAC) in ISAC-enabled V2X systems, specifically regarding the exploitation of sensing interference or coupling interference. In this article, we provide a comprehensive review on PLS metrics and security threats encountered in V2X communication. And then, we discuss and analyze four popular PLS techniques and the main challenges associated with their implementation in ISAC-enabled V2X systems. Finally, we share our vision for PLS studies in ISAC-based V2X systems to promote Secure-ISAC.

cs.IT↗

arXiv2023

Room-temperature orbit-transfer torque enabling van der Waals magnetoresistive memories

The nonvolatile magnetoresistive random access memory (MRAM) is believed to facilitate emerging applications, such as in memory computing, neuromorphic computing and stochastic computing. Two dimensional (2D) materials and their van der Waals heterostructures promote the development of MRAM technology, due to their atomically smooth interfaces and tunable physical properties. Here we report the all-2D magnetoresistive memories featuring all electrical data reading and writing at room temperature based on WTe2/Fe3GaTe2/BN/Fe3GaTe2 heterostructures. The data reading process relies on the tunnel magnetoresistance of Fe3GaTe2/BN/Fe3GaTe2. The data writing is achieved through current induced polarization of orbital magnetic moments in WTe2, which exert torques on Fe3GaTe2, known as the orbit transfer torque (OTT) effect. In contrast to the conventional reliance on spin moments in spin transfer torque and spin orbit torque, the OTT effect leverages the natural out of plane orbital moments, facilitating field-free perpendicular magnetization switching through interface currents. Our results indicate that the emerging OTT MRAM is promising for low power, high performance memory applications.

cond-mat.mes-hall↗

arXiv2023

Semantic-Aware Frame-Event Fusion based Pattern Recognition via Large Vision-Language Models

Pattern recognition through the fusion of RGB frames and Event streams has emerged as a novel research area in recent years. Current methods typically employ backbone networks to individually extract the features of RGB frames and event streams, and subsequently fuse these features for pattern recognition. However, we posit that these methods may suffer from key issues like sematic gaps and small-scale backbone networks. In this study, we introduce a novel pattern recognition framework that consolidates the semantic labels, RGB frames, and event streams, leveraging pre-trained large-scale vision-language models. Specifically, given the input RGB frames, event streams, and all the predefined semantic labels, we employ a pre-trained large-scale vision model (CLIP vision encoder) to extract the RGB and event features. To handle the semantic labels, we initially convert them into language descriptions through prompt engineering, and then obtain the semantic features using the pre-trained large-scale language model (CLIP text encoder). Subsequently, we integrate the RGB/Event features and semantic features using multimodal Transformer networks. The resulting frame and event tokens are further amplified using self-attention layers. Concurrently, we propose to enhance the interactions between text tokens and RGB/Event tokens via cross-attention. Finally, we consolidate all three modalities using self-attention and feed-forward layers for recognition. Comprehensive experiments on the HARDVS and PokerEvent datasets fully substantiate the efficacy of our proposed SAFE model. The source code will be made available at https://github.com/Event-AHU/SAFE_LargeVLM.

cs.CV↗

arXiv2023

Observational signature of continuously operating drivers of decayless kink oscillation

Decayless kink oscillations, which are nearly omnipresent in the solar corona, are believed to be driven by continuously operating energy supply. In this letter, we investigate an external continuous excitation of an apparent decayless oscillation during an X1.1 flare on June 20, 2023 (SOL2023-06-20T16:42).The decayless kink oscillation was identified in the coronal loop at extreme ultraviolet (EUV) wavelengths and the associated flare quasi-periodic pulsations (QPPs) were simultaneously observed in passbands of hard X-ray (HXR), microwave, and ultraviolet (UV) emissions. The kink oscillation is detected as a transverse oscillation of the coronal loop, which reveals five apparent cycles with an average period of about 130-10 s. The oscillation amplitude does not show any significantly decay, suggesting a decayless oscillation. At the same time, the solar flare occurs in the vicinity of the oscillating loop and exhibits five main pulses in HXR, microwave, and UV emissions, which could be regarded as flare QPPs. They have similar periods of about 100-130 s, which may indicate successive and repetitive energy releases during the flare impulsive phase. The peak of each loop oscillation cycle appears to follow the pulse of the QPPs, suggesting that the transverse oscillation is closely associated with flare QPPs. Our observations support the scenario where the repetitive energy released following flare QPPs could be invoked as external, continuously operating drivers of the apparent decayless kink oscillation.

astro-ph.SR↗

arXiv2023

How Much Time is Required for Phase Shift Delivery in RIS-Aided Wireless Systems?

Reconfigurable intelligent surface (RIS) has become a focal point of extensive research due to its remarkable "squared gain". However, achieving a substantial beamforming gain typically requires a significant number of elements, which leads to a non-negligible overhead that forwards the coherent phase shift to the RIS. Different from previous works, which primarily focus on the information transmission phase, we consider the phase delivery overhead during the phase-shift delivery phase to explore the trade-off between performance and overhead. To reduce the phase delivery overhead via the control link, we introduce a hybrid phase shift mechanism, encompassing both the coherent and fixed phase shifts. Specifically, a beamforming problem is formulated for maximizing the throughput. In light of the intractability of the problem, we develop an alternating optimization-based iterative algorithm by combining quadratic transformation and successive convex approximation. To gain more insights, we derive the closed-form expression of the number of elements adopting the coherent phase shift in the large signal-to-noise ratio region. This expression serves as a valuable guide for the practical implementation of the RIS technology. Our simulation results conclusively demonstrate the effectiveness of the proposed algorithm in achieving a favorable trade-off between throughput and overhead. Furthermore, the introduction of the hybrid phase shift approach significantly reduces phase delivery overhead while concurrently enhancing the system throughput.

cs.IT↗

arXiv2023

Simultaneous detection of flare-associated kink oscillations and extreme-ultraviolet waves

Kink oscillations, which are frequently observed in coronal loops and prominences, are often accompanied by extreme-ultraviolet (EUV) waves. However, much more needs to be explored regarding the causal relationships between kink oscillations and EUV waves. In this article, we report the simultaneous detection of kink oscillations and EUV waves that are both associated with an X2.1 flare on 2023 March 03 (SOL2023-03-03T17:39). The kink oscillations, which are almost perpendicular to the axes of loop-like structures, are observed in three coronal loops and one prominence. One short loop shows in-phase oscillation within the same period of 5.2 minutes at three positions. This oscillation could be triggered by the pushing of an expanding loop and interpreted as the standing kink wave. Time lags are found between the kink oscillations of the short loop and two long loops, suggesting that the kink wave travels in different loops. The kink oscillations of one long loop and the prominence are possibly driven by the disturbance of the CME, and that of another long loop might be attributed to the interaction of the EUV wave. The onset time of the kink oscillation of the short loop is nearly same as the beginning of an EUV wave. This fact demonstrates that they are almost simultaneous. The EUV wave is most likely excited by the expanding loop structure and shows two components. The leading component is a fast coronal wave, and the trailing one could be due to the stretching magnetic field lines.

astro-ph.SR↗

arXiv2023

Enhancing Spectrum Sensing via Reconfigurable Intelligent Surfaces: Passive or Active Sensing and How Many Reflecting Elements are Needed?

Cognitive radio has been proposed to alleviate the scarcity of available spectrum caused by the significant demand for wideband services and the fragmentation of spectrum resources. However, sensing performance is quite poor due to the low sensing signal-to-noise ratio, especially in complex environments with severe channel fading. Fortunately, reconfigurable intelligent surface (RIS)-aided spectrum sensing can effectively tackle the above challenge due to its high array gain. Nevertheless, the traditional passive RIS may suffer from the ``double fading'' effect, which severely limits the performance of passive RIS-aided spectrum sensing. Thus, a crucial challenge is how to fully exploit the potential advantages of the RIS and further improve the sensing performance. To this end, we introduce the active RIS into spectrum sensing and respectively formulate two optimization problems for the passive RIS and the active RIS to maximize the detection probability. In light of the intractability of the formulated problems, we develop a one-stage optimization algorithm with inner approximation and a two-stage optimization algorithm with a bisection method to obtain sub-optimal solutions, and apply the Rayleigh quotient to obtain the upper and lower bounds of the detection probability. Furthermore, in order to gain more insight into the impact of the RIS on spectrum sensing, we respectively investigate the number configuration for passive RIS and active RIS and analyze how many reflecting elements are needed to achieve the detection probability close to 1. Simulation results verify that the proposed algorithms outperform existing algorithms under the same parameter configuration, and achieve a detection probability close to 1 with even fewer reflecting elements or antennas than existing schemes.

cs.IT↗