检索结果-内蒙古大学图书馆

IEEE International Symposium on Quality Electronic design

作者： Tarek Mohamed Hussam Amrouch Semiconductor Test and Reliability University of Stuttgart Stuttgart Germany Chair of AI Processor Design Technical University of Munich TUM School of Computation Information and Technology Munich Institute of Robotics and Machine Intelligence

ISBN: (数字)9798331509422

ISBN: (纸本)9798331509439

The relentless drive for advanced technologies, fueled by the demands of AI and safety-critical applications, has intensified the focus on transistor aginga pivotal concern that undermines both transistor reliability and overall circuit performance. This issue is further exacerbated by advancements in packaging and 3D integration, where elevated operating temperatures accelerate aging mechanisms. As technology nodes scales below 3 nm, transistor self-heating emerges as a fundamental challenge, driven by the thermal constraints of 3D confined structures. Traditional physics-based simulation tools, such as Technology CAD (TCAD), struggle to meet the growing computational demands of these intricate designs, with escalating simulation times that impede comprehensive design exploration and optimization. Here, we present a novel framework leveraging machine learning (ML) to accelerate transistor and circuit reliability analysis. These ML-driven methodologies achieve accurate predictions of self-heating and aging effects, enabling rapid identification of aging-prone transistors while drastically reducing computational overhead. Furthermore, by obviating the need to share proprietary, physics-based models from semiconductor foundries, these techniques preserve data confidentiality, addressing critical industry concerns. Such approach not only enhances the scalability of reliability assessments but also offers a transformative pathway for tackling the multifaceted challenges of next-generation semiconductor technologies.

关键词： Solid modeling Accuracy Accelerated aging machine learning Reliability engineering Foundries Transistors Integrated circuit reliability Integrated circuit modeling Three-dimensional integrated circuits

来源：评论

学校读者我要写书评

暂无评论

Measuring mutual policy divergence for multi-agent sequential exploration 24

Measuring mutual policy divergence for multi-agent sequentia...

引用

Proceedings of the 38th International Conference on Neural Information Processing Systems

作者： Haowen Dou Lujuan Dang Zhirong Luan Badong Chen National Key Laboratory of Human-Machine Hybrid Augmented Intelligence and National Engineering Research Center for Visual Information and Applications and Institute of Artificial Intelligence and Robotics Xi'an Jiaotong University School of Electrical Engineering Xi'an University of Technology

ISBN: (纸本)9798331314385

Despite the success of Multi-Agent Reinforcement Learning (MARL) algorithms in cooperative tasks, previous works, unfortunately, face challenges in heterogeneous scenarios since they simply disable parameter sharing for agent specialization. Sequential updating scheme was thus proposed, naturally diversifying agents by encouraging agents to learn from preceding ones. However, the exploration strategy in sequential scheme has not been investigated. Benefiting from updating one-by-one, agents have the access to the information from preceding agents. Thus, in this work, we propose to exploit the preceding information to enhance exploration and heterogeneity sequentially. We present Multi-Agent Divergence Policy Optimization (MADPO), equipped with mutual policy divergence maximization framework. We quantify the discrepancies between episodes to enhance exploration and between agents to heterogenize agents, termed intra-agent divergence and inter-agent divergence. To address the issue that traditional divergence measurements lack stability and directionality, we propose to employ the conditional Cauchy-Schwarz divergence to provide entropy-guided exploration incentives. Extensive experiments show that the proposed method outperforms state-of-the-art sequential updating approaches in two challenging multi-agent tasks with various heterogeneous scenarios. Source code is available at https://***/hwdou6677/MADPO.

关键词：

来源：评论

学校读者我要写书评

暂无评论

DenseKD: Dense Knowledge Distillation by Exploiting Region and Sample Importance

引用

IEEE Transactions on Neural Networks and Learning Systems 2025年第6期36卷 11243-11257页

作者： Haonan Zhang Longjun Liu Yi Zhang Xinyu Lei Fei Hui Bihan Wen National Key Laboratory of Human-Machine Hybrid Augmented Intelligence the National Engineering Research Center for Visual Information and Applications Institute of Artificial Intelligence and Robotics Xi’an Jiaotong University Xi’an China School of Electronics and Control Engineering Chang’an University Xi’an China School of Electrical and Electronic Engineering Nanyang Technological University Jurong West Singapore

Knowledge distillation (KD) can compress deep neural networks (DNNs) by transferring the knowledge of the redundant teacher model to the resource-friendly student model, where cross-layer KD (CKD) conducts KD between each stage of students and the multiple stages of teachers. However, previous CKD schemes select the coarse-grained stagewise features of teachers to teach students, leading to improper channel alignment. Also, most of these methods conduct uniform distillation for all the knowledge, limiting students to focus more on important knowledge. To address these problems, we propose a dense KD (DenseKD) in this article, dubbed as DenseKD. First, to achieve more accurate feature alignment in CKD, we construct the learnable dense architecture to make each channel of student flexibly capture more diverse channelwise features from teacher. Moreover, we introduce region importance to investigate the region’s guiding potential, it distinguishes the influence of different regions by the variation of representations of teacher models. In addition, to make students pay more attention to useful samples in KD, we calculate sample importance by the loss of teacher models. Consistent improvements over state-of-the-art approaches are observed in experiments on multiple vision tasks. For example, in the classification task, DenseKD achieves 72.30% accuracy of ResNet-20 on CIFAR-100, which is higher than the results of previous CKD methods. In addition, in the object detection task, DenseKD gains 2.84% mean average precision (mAP) improvements of Faster R-CNN with ResNet-18 against vanilla KD.

关键词： Knowledge engineering Cross layer design Chronic kidney disease Knowledge transfer Feature extraction Accuracy Object detection Control engineering Visualization Training

来源：评论

学校读者我要写书评

暂无评论

Intention-Oriented Fast Joint Prediction and Planning for Safe Driving

Intention-Oriented Fast Joint Prediction and Planning for Sa...

引用

International Conference on Intelligent Transportation

作者： Mingyi Zhang Jianru Xue Chen Lv Jianwu Fang National Key Laboratory of Human-Machine Hybrid Augmented Intelligence Institute of Artificial Intelligence and Robotics Xi'an Jiaotong University China School of Mechanical and Aerospace Engineering Nanyang Technological University Singapore

ISBN: (数字)9798331505929

ISBN: (纸本)9798331505936

The primary challenges for Autonomous Vehicle (AV) navigation in complex scenarios are to effectively handle the interactions between the Ego-Vehicle (EV) and surrounding participants and generate safe yet non-conservative plans. Current approaches employ Planning-Informed Prediction (PIP) and Joint Prediction and Planning (JPP) to model these interactions. At the same time, information loss exists in PIP and JPP always ignores the intention interactions for a free space generation. To this end, this paper incorporates intention interactions into JPP, where a shared intention interaction module is adopted to both the planning and the prediction tasks via a cross-query mechanism, enabling the model to reduce the model size and accelerate inference speed. In contrast to previous methods that typically plan a single trajectory and result in conservative behaviors, we introduce a contingent-masked planner to balance the safety and efficiency, so as to fulfill the scene-level prediction and greatly improve computational efficiency. We evaluate our approach through closed-loop experiments on the CARLA longest6 and Town05 benchmarks, and faster inference speed and more robust performance against state-of-the-art methods.

关键词： Navigation Computational modeling Contingency management Predictive models Benchmark testing Planning Safety Computational efficiency Trajectory Noise measurement

来源：评论

学校读者我要写书评

暂无评论

LEMON: Localized Editing with Mesh Optimization and Neural Shaders

arXiv

引用

arXiv 2024年

作者： Algan, Furkan Mert Yazgan, Umut Salihu, Driton Eteke, Cem Steinbach, Eckehard Technical University of Munich Chair of Media Technology Munich Institute of Robotics and Machine Intelligence MIRMI School of Computation Information and Technology Department of Computer Engineering Germany

In practical use cases, polygonal mesh editing can be faster than generating new ones, but it can still be challenging and time-consuming for users. Existing solutions for this problem tend to focus on a single task, either geometry or novel view synthesis, which often leads to disjointed results between the mesh and view. In this work, we propose LEMON, a mesh editing pipeline that combines neural deferred shading with localized mesh optimization. Our approach begins by identifying the most important vertices in the mesh for editing, utilizing a segmentation model to focus on these key regions. Given multi-view images of an object, we optimize a neural shader and a polygonal mesh while extracting the normal map and the rendered image from each view. By using these outputs as conditioning data, we edit the input images with a text-to-image diffusion model and iteratively update our dataset while deforming the mesh. This process results in a polygonal mesh that is edited according to the given text instruction, preserving the geometric characteristics of the initial mesh while focusing on the most significant areas. We evaluate our pipeline using the DTU dataset, demonstrating that it generates finely-edited meshes more rapidly than the current state-of-the-art methods. We include our code and additional results in the supplementary material. © 2024, CC BY.

关键词： Mesh generation

来源：评论

学校读者我要写书评

暂无评论

A Bio-inspired Solution for Double Support Force Distribution in Humanoid Robot Locomotion

A Bio-inspired Solution for Double Support Force Distributio...

引用

IEEE-RAS International Conference on Humanoid Robots

作者： Francisco Javier Andrade Chavez Vidyasagar Rajendran Katja Mombaur CERC Human-Centred Robotics and Machine Intelligence of Department of Systems Design Engineering University of Waterloo Canada Optimization and Biomechanics for Human-Centred Robotics Karlsruhe Institute of Technology (KIT) Institute of Anthropomatics and Robotics (IAR) Karlsruhe Germany

In locomotion, the likelihood of slipping or maintaining contact is determined by the forces applied in the environment. Therefore, it is crucial to find methods of keeping forces within friction constraints. In single support, the relationship between center of mass acceleration and forces is unique. However, in double support, it becomes a non-deterministic problem. It is often assumed that forces are distributed to minimize a certain effort criterion. An interesting alternative is to distribute the forces in a manner similar to how a human would, which could result in a more human-like gait for humanoid robots. With this in mind, we introduce the modified Twin Polynomial Method (mTPM) as a technique to accurately distribute double support forces in a human-like way. The relevant relationships used in this method are presented, and we compare its performance against the current state of the art in human gait analysis. The comparison is conducted across three different types of contacts with the environment: barefoot, socks, and shoes. We show how human data differs from a ‘typical’ robot walk and use the proposed method to generate a more human-like distribution for robots.

关键词：

来源：评论

学校读者我要写书评

暂无评论

TReCiM: Lower Power and Temperature-Resilient Multibit 2FeFET-1T Compute-in-Memory design 24

TReCiM: Lower Power and Temperature-Resilient Multibit 2FeFE...

引用

43rd International Conference on Computer-Aided design, ICCAD 2024

作者： Zhou, Yifei Kämpfe, Thomas Ni, Kai Amrouch, Hussam Zhuo, Cheng Yin, Xunzhao Zhejiang University Hangzhou China Fraunhofer IPMS Dresden Germany Department of Electrical Engineering University of Notre Dame Notre Dame United States Technical University of Munich TUM School of Computation Information and Technology Munich Institute of Robotics and Machine Intelligence Munich Germany Key Laboratory of Collaborative Sensing and Autonomous Unmanned Systems of Zhejiang Province China

ISBN: (纸本)9798400710773

Compute-in-memory (CiM) emerges as a promising solution to solve hardware challenges in artificial intelligence (AI) and the Internet of Things (IoT), particularly addressing the"memory wall" issue. By utilizing nonvolatile memory (NVM) devices in a crossbar structure, CiM efficiently accelerates multiply-accumulate (MAC) computations, the crucial operations in neural networks and other AI models. Among various NVM devices, Ferroelectric FET (FeFET) is particularly appealing for ultra-low-power CiM arrays due to its CMOS compatibility, voltage-driven write/read mechanisms and high ION/IOFF ratio. Moreover, subthreshold-operated FeFETs, which operate at scaling voltages in the subthreshold region, can further minimize the power consumption of CiM array. However, subthreshold-FeFETs are susceptible to temperature drift, resulting in computation accuracy degradation. Existing solutions exhibit weak temperature resilience at larger array size and only support 1-bit. In this paper, we propose TReCiM, an ultra-low-power temperature-resilient multibit 2FeFET-1T CiM design that reliably performs MAC operations in the subthreshold-FeFET region with temperature ranging from 0°C to 85°C at scale. We benchmark our design using NeuroSim framework in the context of VGG-8 neural network architecture running the CIFAR-10 dataset. Benchmarking results suggest that when considering temperature drift impact, our proposed TReCiM array achieves 91.31% accuracy, with 1.86% accuracy improvement compared to existing 1-bit 2T-1FeFET CiM array. Furthermore, our proposed design achieves 48.03 TOPS/W energy efficiency at system level, comparable to existing designs with smaller technology feature sizes. © 2024 Copyright is held by the owner/author(s).

关键词： MISFET devices

来源：评论

学校读者我要写书评

暂无评论

Whole-body Stabilization of the REEM-C Humanoid on Actuated Balance Boards in Sagittal and Frontal Directions

Whole-body Stabilization of the REEM-C Humanoid on Actuated ...

引用

IEEE-RAS International Conference on Humanoid Robots

作者： Vidyasagar Rajendran Francisco Javier Andrade Chavez Katja Mombaur Department of Systems Design Engineering CERC Human-Centred Robotics and Machine Intelligence University Waterloo Canada Optimization and Biomechanics for Human-Centred Robotics Institute of Anthropomatics and Robotics (IAR) Karlsruhe Institute of Technology (KIT) Karlsruhe Canada

This paper explores stabilizer design for humanoid balancing on actuated balance boards. We qualitatively study how humans balance on balance boards to motivate our humanoid controller design. The developed approach combines various strategies from the literature including ankle admittance, CoM admittance, base pitch admittance, foot force difference admittance, contact wrench distribution and angular momentum regulation in a new integration to achieve stabilization on an actuated balance board. We explore the frontal and sagittal plane balance board problems along with pushes at the upper body and show that our approach works in simulation and on real hardware. This brings us one step closer to balancing on unactuated balance boards, other dynamic environments such as moving platforms and transportation devices, skills that will make humanoids more versatile.

关键词：

来源：评论

学校读者我要写书评

暂无评论

VLP-KAN: Low-Complexity and Interpretable RSS-Based VLP Using Kolmogorov-Arnold Networks

VLP-KAN: Low-Complexity and Interpretable RSS-Based VLP Usin...

引用

Annual Conference on Wireless On Demand Network Systems and Services (WONS)

作者： Fabian Seguel Driton Salihu Stefan Hägele Eckehard Steinbach Department of Computer Engineering Chair of Media Technology Munich Institute of Robotics and Machine Intelligence (MIRMI) Technical University of Munich School of Computation Information and Technology München Germany

ISBN: (数字)9783903176713

ISBN: (纸本)9798331522025

Received signal strength (RSS)-based Visible Light Positioning (VLP) has gained attention due to its relatively easy infrastructure deployment and high localization accuracy. Different low-complexity geometrical and machine learning (ML)-based models have been proposed for localization because of their robustness against the uncertainty produced by tilted devices, non-Lambertian sources, and the low dimensionality of RSS-vector. Inspired by the Kolmogorov-Arnold representation theorem, Kolmogorov-Arnold Networks (KANs) have been proposed as a promising alternative to Multi-Layer Perceptrons (MLPs). This paper evaluates the use and performance of KANs in VLP systems for the first time. Results show that in the proposed scenario, KANs achieve below cm-level accuracy. Moreover, symbolic regression (SR) can be implemented straightforwardly to find a function that relates RSS with distance. It is shown in this paper that both KANs and SR-KAN models outperform MLP and Weighted K-nearest neighbors based approaches.

关键词： Location awareness Wireless communication Performance evaluation Deep learning Accuracy Uncertainty Nearest neighbor methods Fingerprint recognition Robustness Network systems

来源：评论

学校读者我要写书评

暂无评论

Capacitor Minimization in Analog Synapses for Efficient and Compact Binarized SNNs

IEEE Transactions on Circuits and Systems for Artificial Int...

引用

IEEE Transactions on Circuits and Systems for Artificial intelligence 2025年第2期2卷 114-125页

作者： Mikail Yayla Ming-Liang Wei Simon Thomann Albi Mema Chia-Lin Yang Jian-Jia Chen Hussam Amrouch Design Automation for Embedded Systems Group TU Dortmund University Dortmund Germany Lamarr Institute for Machine Learning and Artificial Intelligence Dortmund Germany Department of the Graduate Institute of Electronics Engineering National Taiwan University Taipei Taiwan Technical University of Munich (TUM) Munich Germany TUM School of Computation Information and Technology Chair of AI Processor Design Munich Institute of Robotics and Machine Intelligence Munich Germany Department of Computer Science and Information Engineering Graduate Institute of Electronics Engineering National Taiwan University Taipei Taiwan Department of Computer Science TU Dortmund University Dortmund Germany Lamarr Institute Dortmund Germany

Accelerating the immense workloads of Neural Networks (NNs) is a critical challenge, and analog computing-based approaches present a promising solution. Among these, the Integrate-and-Fire (IF) Spiking Neural Network (SNN) stands out for its potential in efficient neural computation. However, achieving high inference accuracy through precise multiply-accumulate (MAC) operations necessitates a large membrane capacitor—224 to 547 times larger than the synapse array at the 14nm technology node—resulting in prohibitive area costs. Additionally, the large capacitor size introduces higher energy consumption and longer access times. Thus, a key research challenge is maintaining accuracy while mitigating the costs associated with the large membrane capacitor. In this work, we propose a HW/SW Codesign method, called CapMin, for capacitor size minimization in analog computing IF-SNNs. CapMin minimizes the capacitor size by reducing the number of spike times needed for accurate operation of the HW, based on the absolute frequency of MAC level occurrences in the SW. To increase the operation of IF-SNNs to current variation, we propose the method CapMin-V, which trades capacitor size for protection based on the reduced capacitor size found in CapMin. In our experiments, CapMin achieves more than a 14 $\times$ reduction in capacitor size along with a 34% reduction in energy consumption over the state of the art, while CapMin-V achieves increased variation tolerance in the IF-SNN operation, requiring only a small increase in capacitor size.

关键词： Capacitors Circuits Accuracy Capacitance Micromechanical devices Costs Neurons Integrated circuit modeling Spiking neural networks Hardware

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：