Diffusion Policy (DP) has demonstrated strong potential in robotic visuomotor control, offering robust generalization and seamless integration of multi-modal data. However, its complex model structure and increasing multi-modal inputs have brought latency and power challenges for edge resource-constrained robotic platforms. To address the above challenges, we identify the potential intra-model and inter-model redundancies in DP. We observe that DP relies on frequent multi-modal inputs such as images and text during execution. However, the demands of fine-grained robotic manipulation result in substantial intra-modal similarity across consecutive image frames, which, combined with inter-modal semantic redundancy between images and language, indicates that much of the input information is repetitive and potentially compressible. Yet prior works have not exploited these characteristics for targeted optimization. We therefore propose a hardware–software co-design accelerator. On the algorithmic side, we introduce self- and cross-guided modal compression, leveraging intra- and inter-modality similarity to reduce redundant computation within the key DP modules. On the hardware side, we design a tailored architecture that supports multiple operators with optimized sparse memory access, lightweight computation engines, and reconfigurable on-chip dataflow, substantially reducing energy cost. Experimental results demonstrate a 26× speedup over a high-performance GPU while consuming only 1.5 W, enabling low-power and real-time robotic control on edge robotic devices.

Smart contracts, closely intertwined with cryptocurrency transactions, have sparked widespread concerns about considerable financial losses of security issues. To counteract this, a variety of tools have been developed to identify vulnerability in smart contract. However, they fail to overcome two challenges at the same time when faced with smart contract bytecode: (i) strong interference caused by enormous non-relevant instructions; (ii) missing semantics of bytecode due to incomplete data and control flow dependencies. In this paper, we propose a multi-teacher based bytecode vulnerability detection method, namely Multi-Teacher Vulnerability Hunter (MTVHunter), which delivers effective denoising and missing semantic to bytecode under multi-teacher guidance. Specifically, we first propose an instruction denoising teacher to eliminate noise interference by abstract vulnerability pattern and further reflect in contract embeddings. Secondly, we design a novel semantic complementary teacher with neuron distillation, which effectively extracts necessary semantic from source code to replenish the bytecode. Particularly, the proposed neuron distillation accelerate this semantic filling by turning the knowledge transition into a regression task. We conduct experiments on 229,178 real-world smart contracts that concerns four types of common vulnerabilities. Extensive experiments show MTVHunter achieves significantly performance gains over state-of-the-art approaches.

Although deep unfolding networks are widely studied for low-level computer vision tasks and the Robust Principal Component Analysis (RPCA) is unfolded for high-level task of anomaly detection, almost no work focus on interpretable multi-class semantic segmentation network. To resolve this issue, we propose an edge-aware RPCA optimization model for brain tissue segmentation by formulating each-class segmentation as non-edge estimation and edge prediction. Firstly, according to the optimization model, three parts of Multi-class Low-rank Non-edge-estimation (MLN), Multi-class Edge Prediction (MEP) and Multi-class Semantic Prediction (MSP) can be formed. Secondly, we unfold these three parts as MLN, MEP and MSP networks at each stage to build an Edge-aware RPCA Segmentation Network (ER-SegNet). Thirdly, multi-class features are jointly leveraged by each one among these three networks rather than independent multi-class prediction. Experimental results have shown that the proposed ER-SegNet can compete against many newest segmentation methods in terms of pixel accuracy and Dice score.

Microservice systems, characterized by their distributed nature and dynamic environments, pose significant challenges for failure diagnosis. Traditional failure diagnosis methods based on a single source of observability data, such as logs, metrics, or traces, fall short due to their inability to manage the complexity of inter-service communications. Consequently, in recent years, many methods based on multi-source observability data have been proposed, which promise a comprehensive analysis by integrating logs, metrics, traces. However, despite the promising potential of multi-source methods, we find that existing works often overlook a critical question: whether multi-source data is truly necessary for failure diagnosis. To address this, we conduct a systematic evaluation of eleven representative failure diagnosis methods based on multi-source observability data, using three public datasets and our own curated dataset. Our experiments focus on multiple aspects of multi-source datasets and methods. The results indicate that the quality of existing open-source datasets is inconsistent, and not all studied methods consistently perform well. Surprisingly, we find that adding more data sources does not necessarily improve the performance of microservice failure diagnosis in some cases.

Effectively preserving and encoding structure features from objects in irregular and sparse LiDAR points is a crucial challenge to 3D object detection on the point cloud. Recently, Transformer has demonstrated promising performance on many 2D and even 3D vision tasks. Compared with the fixed and rigid convolution kernels, the self-attention mechanism in Transformer can adaptively exclude the unrelated or noisy points and is thus suitable for preserving the local spatial structure in the irregular LiDAR point cloud. However, Transformer only performs a simple sum on the point features, based on the self-attention mechanism, and all the points share the same transformation for value. A such isotropic operation cannot capture the direction-distance-oriented local structure, which is essential for 3D object detection. In this work, we propose a Structure-Embedding transFormer (SEFormer), which can not only preserve the local structure as a traditional Transformer but also have the ability to encode the local structure. Compared to the self-attention mechanism in traditional Transformer, SEFormer learns different feature transformations for value points based on the relative directions and distances to the query point. Then we propose a SEFormer-based network for high-performance 3D object detection. Extensive experiments show that the proposed architecture can achieve SOTA results on the Waymo Open Dataset, one of the most significant 3D detection benchmarks for autonomous driving. Specifically, SEFormer achieves 79.02% mAP, which is 1.2% higher than existing works. https://github.com/tdzdog/SEFormer.

Unlike the grid-paced RGB images, network compression, i.e.pruning and quantization, for the irregular and sparse 3D point cloud face more challenges. Traditional quantization ignores the unbalanced semantic distribution in 3D point cloud. In this work, we propose a semantic-guided adaptive quantization framework for 3D point cloud. Different from traditional quantization methods that adopt a static and uniform quantization scheme, our proposed framework can adaptively locate the semantic-rich foreground points in the feature maps to allocate a higher bitwidth for these "important" points. Since the foreground points are in a low proportion in the sparse 3D point cloud, such adaptive quantization can achieve higher accuracy than uniform compression under a similar compression rate. Furthermore, we adopt a block-wise fine-grained compression scheme in the proposed framework to fit the larger dynamic range in the point cloud. Moreover, a 3D point cloud based software and hardware co-evaluation process is proposed to evaluate the effectiveness of the proposed adaptive quantization in actual hardware devices. Based on the nuScenes dataset, we achieve 12.52% precision improvement under average 2-bit quantization. Compared with 8-bit quantization, we can achieve 3.11× energy efficiency based on co-evaluation results.

Computing-in-memory (CIM) architecture utilizing emerging non-volatile devices is promising for energy-efficient neural network (NN) applications. However, the non-ideal factors of non-volatile devices and analog circuits may incur severe accuracy loss, which cuts the algorithm and hardware design apart. The algorithm/hardware designers are skilled in either macro-scope NN models or detailed circuit/device errors, while sophisticated research of the joint effect on accuracy loss is urgently needed. In this paper, we propose a user-friendly, fast, and accurate simulation framework (CIMUFAS) to explore the impact of various non-ideal devices/circuits on algorithm accuracy. Based on multiple architecture-level CIM mapping/scheduling workflow, sophisticated non-ideal factors with different error models are established. The CIMUFAS also provides easy-to-use interfaces to flexibly support user-defined models/parameters for specified devices/circuits. Besides, the CIMUFAS framework achieves reasonable simulation time. Compared with MNSIM 2.0, the simulation time is reduced by 45% even after adding a more realistic hardware configuration. This CIMUFAS framework is verified with two fabricated CIM chips with <0.04% accuracy mismatch. The source code of CIMUFAS is publicly available at https://github.com/Hlal/CIMUFAS.

Weight pruning methods of deep neural networks (DNNs) have been demonstrated to achieve a good model pruning rate without loss of accuracy, thereby alleviating the significant computation/storage requirements of large-scale DNNs. Structured weight pruning methods have been proposed to overcome the limitation of irregular network structure and demonstrated actual GPU acceleration. However, in prior work, the pruning rate (degree of sparsity) and GPU acceleration are limited (to less than 50%) when accuracy needs to be maintained. In this work, we overcome these limitations by proposing a unified, systematic framework of structured weight pruning for DNNs. It is a framework that can be used to induce different types of structured sparsity, such as filterwise, channelwise, and shapewise sparsity, as well as nonstructured sparsity. The proposed framework incorporates stochastic gradient descent (SGD; or ADAM) with alternating direction method of multipliers (ADMM) and can be understood as a dynamic regularization method in which the regularization target is analytically updated in each iteration. Leveraging special characteristics of ADMM, we further propose a progressive, multistep weight pruning framework and a network purification and unused path removal procedure, in order to achieve higher pruning rate without accuracy loss. Without loss of accuracy on the AlexNet model, we achieve 2.58× and 3.65× average measured speedup on two GPUs, clearly outperforming the prior work. The average speedups reach 3.15× and 8.52× when allowing a moderate accuracy loss of 2%. In this case, the model compression for convolutional layers is 15.0× , corresponding to 11.93× measured CPU speedup. As another example, for the ResNet-18 model on the CIFAR-10 data set, we achieve an unprecedented 54.2× structured pruning rate on CONV layers. This is 32× higher pruning rate compared with recent work and can further translate into 7.6× inference time speedup on the Adreno 640 mobile GPU compared with the original, unpruned DNN model. We share our codes and models at the link http://bit.ly/2M0V7DO.

Although compression techniques like pruning or quantization are beneficial for accelerators' energy efficiency, the large search space makes finding the appropriate compression scheme difficult. Besides, most existing works ignore the combination of both pruning and quantization. In this paper, we propose a reinforcement learning (RL) based joint compression framework to find the appropriate pruning ratio and quantization bit-width for accelerators. By interacting with the energy model of the target accelerator, the RL agent can learn the effect of compression scheme on both accuracy and energy efficiency. Through a long trial-and-error process, the agent can finally reach an optimal trade-off between accuracy and energy efficiency. Compared with control groups whose compression hyper-parameters are not jointly optimized, the proposed framework can achieve at least 25% energy reduction with higher accuracy or much higher accuracy with small disadvantages on energy. Compared with 8-bit quantized baseline, the framework can achieve 90% and 85% energy reduction on Cifar10 and Cifar100 respectively.

Convolutional neural networks (CNNs) have become powerful algorithms in various tasks. Application-specific integrated circuit (ASIC) has been widely used to accelerate CNN on mobile platforms because of its tremendous energy efficiency and performance. Meanwhile, CNNs have become much sparser with the development of network pruning algorithms. Recent works have employed different methods to improve the energy efficiency and performance of ASIC accelerators by utilizing the sparsity character of CNN. However, some of these methods suffer from tremendous output memory overhead and performance degradation induced by hash collisions. To overcome the aforementioned problem, we propose GAAS: an efficient group associated architecture and scheduler module for sparse CNN accelerators. It achieves smaller output memory overhead and higher performance compared with the state-of-the-art accelerator. Our proposed method GAAS mainly consists of two parts: 1) an$n$-way group associated architecture to reduce the output memory overhead and 2) a scheduler module to improve the performance. Besides, a load-balancing algorithm is proposed and implemented in the scheduler module to improve the performance by reducing the hash collision rate. To demonstrate the efficiency of GAAS, we implement a 4-way image-principal associated architecture with a$16\times 16$PE array and the scheduler module on our proposed method. The experimental results on AlexNet, VGG16, ResNet18, and MobileNet show that GAAS can reduce the output memory overhead by 50%, and it can surely improve the performance of them by$1.53\times $,$1.62\times $,$1.46\times $, and$1.55\times $, respectively.

The transmitter identification of digital television (DTV) systems becomes crucial nowadays. Transmitter identification (TxID, or transmitter fingerprinting) technique is used to detect, diagnose and classify the operating status of any radio transmitter of interest. The TxID system is specified in Advanced Television System Committee (ATSC) A/110 standard where pseudo random sequences are proposed to be embedded into the DTV signals before transmission. Thus, the transmitter identification can be realized by invoking the cross-correlation functions between the received signal and the possible candidates of the pseudo random sequences. The buried ratio or injection level of injected Kasami sequences in DTV-TxID systems will both affect the identification correctness and the DTV reception quality. In this paper, we investigate the important unsolved optimization problem for injection level. We present the new analysis here for the realistic scenario consisting of multiple transmitters and receivers over the additive white Gaussian noise channel. The signal-to-interference-plus-noise ratios for the TxID signal detection and the subject TV signal reception are both considered as two essential measures for single-frequency networks. Besides, we design a novel efficient injection-level optimization scheme for TxID simply based on the given information including the signal-to-noise ratio at the receiver and the locations of the transmitters and the receiver(s). Finally, several examples are also demonstrated for the TxID injection-level optimization in this paper.

The transmitter identification (TxID) of digital television (DTV) systems becomes crucial nowadays. TxID (or transmitter fingerprinting) technique is used to detect, diagnose and classify the operating status of any radio transmitter of interest. The TxID system is specified in Advanced Television System Committee (ATSC) A/110 standard where pseudorandom sequences are proposed to be embedded into the DTV signals before transmission. The buried ratio or injection level of injected Kasami sequences in DTV-TxID systems will both affect the identification correctness and the DTV reception quality. In this study, the authors investigate the important unsolved optimisation problem for injection level. The authors present the new analysis here for the realistic scenario consisting of multiple transmitters and receivers over the additive white Gaussian noise channel. The signal-to-interference-plus-noise ratios for the TxID signal detection and the subject TV signal reception are both considered as two essential measures for single-frequency networks. Besides, the authors design a novel efficient injection-level optimisation scheme for TxID simply based on the given information including the signal-to-noise ratio at the receiver and the locations of the transmitters and the receiver(s).

The digital imaging technology has grown explosively for multimedia applications in recent years. The need for the copyrighted digitalized media becomes urgent nowadays. An approach for the digital copyright protection is to employ advanced watermarking techniques, where watermarks can reveal the ownership identities. Generally speaking, the watermarks are embedded into image or video signals. In this paper, we will investigate digital watermarking techniques and propose a new optimal watermarking scheme. When multiple embedded watermarks are considered, a new analysis for the signal-to-interference-plus-noise-ratios (SINRs) with respect to the subject signal and the watermark signals is carried out. The objective quality measure for the digital watermarking applications should essentially consist of both signal-to-interference-plus-noise-ratio for the subject signal and similarity coefficients for the watermarks. In order to optimize the aforementioned objective measure, we design a novel efficient scale-factor optimization scheme, which can lead to the maximum overall SINR for both subject signal and watermarks. Simulation results are also demonstrated to illustrate the effectiveness of our proposed method.

ENG is a broadcasting industry acronym which stands for electronic news gathering in digital terrestrial television systems. It usually means that a television crew take an ENG truck on location to do a live report for a newscast. The ENG transmitter identification becomes important in digital video broadcasting nowadays. However, the transmitter identification could be quite difficult in a distributed transmission network. A pseudo-random sequence was proposed to be embedded into the digital television (DTV) signal prior to transmission. Thus, the transmitter identification can be realized by invoking the cross-correlation functions between the received signal and the possible candidates of the pseudo random sequences. Kasami sequences are commonly used as the transmitter ID (Tx-ID) sequences as they provide a large family of nearly-orthogonal codes. In order to investigate the sensitivity of the transmitter identification performance to the arbitrary topologies and the Kasami sequence lengths, we present the new analysis here for Tx-ID over random geometric layouts. In our analysis, the lowest received co-channel signal-to-interference ratio is considered as the crucial factor for the multiple-transmitter identification. Moreover, the transmitter location uncertainties are also considered in this paper. The effect of such uncertainties on the minimum required Kasami sequence lengths is investigated. It turns out to be that the larger the Kasami sequence length, the larger the received signal-to-interference ratio. Our new analysis can be used to determine the required Kasami sequence length for an arbitrary ENG-transmitter topology.