Advanced Intelligent Systems

Analog electronic and photonic crossbar arrays have been emerging as energy‐efficient hardware implementations to accelerate computationally intensive general matrix–vector and matrix–matrix multiplications in machine learning (ML) algorithms. However, the inevitable nonuniformity in large‐scale electronic and optoelectronic devices and systems prevents scalable deployment. Herein, a calibration approach is reported that enables accurate calculations in crossbar arrays despite hardware imperfections. This approach is experimentally validated in a small‐scale free‐space photonic crossbar array based on cascaded spatial light modulators and demonstrated the scalability and universality of this approach in various large‐scale electronic and photonic crossbar arrays. The improved performance of calibrated crossbar arrays in an ML model inference is further demonstrated to classify handwritten digital images.

Jichao Fan

Yingheng Tang

Weilu Gao

Universal Approach for Calibrating Large‐Scale Electronic and Photonic Crossbar Arrays

The emergence of memristors with potential applications in data storage and artificial intelligence has attracted wide attentions. Memristors are assembled in crossbar arrays with data bits encoded by the resistance of individual cells. Despite the proposed high density and excellent scalability, the sneak‐path current causing cross interference impedes their practical applications. Therefore, developing novel architectures to mitigate sneak‐path current and improve efficiency, reliability, and stability may benefit next‐generation storage‐class memory (SCM). Moreover, conventional digital computers face the von‐Neumann bottleneck and the slowdown of transistors’ scaling, imposing a big challenge to hardware artificial intelligence. Memristive crossbar features colocation of memory and processing units, as well as superior scalability, making it a promising candidate for hardware accelerating machine learning and neuromorphic computing. Herein, first, crossbar architecture is introduced. Then, for storage, the origin of sneak‐path current is reviewed and techniques to mitigate this issue from the angle of materials and circuits are discussed. Computing wise, the applications of memristive crossbars in both machine learning and neuromorphic computing are surveyed, focusing on the structure of unit cells, the network topology, and the learning types. Finally, a perspective on future engineering and applications of memristive crossbars is discussed.

Huihan Li

Shaocong Wang

Xumeng Zhang

Wei Wang

Rui Yang

Zhong Sun

Wanxiang Feng

Peng Lin

Zhongrui Wang

Linfeng Sun

Yugui Yao

Memristive Crossbar Arrays for Storage and Computing Applications

忆阻交叉阵列用于存储和计算应用

Due to the rapid development of artificial intelligence (AI) and internet of things (IoTs), neuromorphic computing and hardware security are becoming more and more important. The volatile memristors, which feature spontaneous decay of device conductance, own the distinct combination of high similarity to the biological neurons and synapses and unique physical mechanisms. They are excellent candidates for mimicking the synaptic functions and ideal randomness source of the entropy for hardware‐based security. Herein, the recent advances of volatile memristors in devices, mechanisms, and application aspects are summarized. First, a brief introduction is presented to describe the switching type, materials, and temporal response of volatile memristors. Second, the volatile switching mechanisms are discussed and grouped into ion effects, thermal effects, and electrical effects. Third, attention is focused on the applications of volatile memristors for access devices, neuromorphic computing (artificial neurons and synapses), and hardware security (true random number generators and physical unclonable functions). Finally, major challenges and future outlook of volatile memristors for neuromorphic computing and hardware security are discussed.

Ruopeng Wang

Jia-Qin Yang

Jing-Yu Mao

Zhan-Peng Wang

Shuang Wu

Maojie Zhou

Tianyi Chen

Ye Zhou

Su-Ting Han

Recent Advances of Volatile Memristors: Devices, Mechanisms, and Applications

挥发性忆阻器的最新进展：器件、机理及应用

Neuromorphic electronics draw attention as innovative approaches that facilitate hardware implementation of next‐generation artificial intelligent system including neuromorphic in‐memory computing, artificial sensory perception, and humanoid robotics. Among the various neuromorphic devices, optoelectronic synapses are promising neuromorphic devices that use optical means to mimic synaptic plasticity and related functions. Compared with classical neuromorphic chip based on electronic synapses using electrical means, photonic neuromorphic chip using light as input spike signal can be attractive alternative approach for next‐generation artificial intelligent system capable of high density, low power consumption, and low crosstalk. Thus, various optoelectronic synaptic electronics have been developed to overcome the drawback of conventional artificial intelligent system based on electrical synapses. Herein, the recent progresses in transistor‐based optoelectronic synapses for artificial intelligent system and review their device architecture, neuromorphic operational mechanisms, manufacturing methodologies, and advanced applications for artificial intelligent computing and visual perception systems are focused. Finally, the future challenges and research direction in the optoelectronic synaptic research are discussed.

Sung Woon Cho

Sung Min Kwon

Yong-Hoon Kim

Sung Kyu Park

Recent Progress in Transistor‐Based Optoelectronic Synapses: From Neuromorphic Computing to Artificial Sensory System

晶体管基光电突触的最新进展：从神经形态计算到人工感觉系统

Machine learning, particularly in the form of deep learning (DL), has driven most of the recent fundamental developments in artificial intelligence (AI). DL is based on computational models that are, to a certain extent, bio‐inspired, as they rely on networks of connected simple computing units operating in parallel. The success of DL is supported by three factors: availability of vast amounts of data, continuous growth in computing power, and algorithmic innovations. The approaching demise of Moore's law, and the consequent expected modest improvements in computing power that can be achieved by scaling, raises the question of whether the progress will be slowed or halted due to hardware limitations. This article reviews the case for a novel beyond‐complementary metal–oxide–semiconductor (CMOS) technology—memristors—as a potential solution for the implementation of power‐efficient in‐memory computing, DL accelerators, and spiking neural networks. Central themes are the reliance on non‐von‐Neumann computing architectures and the need for developing tailored learning and inference algorithms. To argue that lessons from biology can be useful in providing directions for further progress in AI, an example‐based reservoir computing is briefly discussed. At the end, speculation is given on the “big picture” view of future neuromorphic and brain‐inspired computing systems.

Adnan Mehonic

Abu Sebastian

Bipin Rajendran

Osvaldo Simeone

Eleni Vasilaki

Anthony J. Kenyon

Memristors—From In‐Memory Computing, Deep Learning Acceleration, and Spiking Neural Networks to the Future of Neuromorphic and Bio‐Inspired Computing

忆阻器——从内存计算、深度学习加速和脉冲神经网络到神经形态与仿生计算的未来

Ornithopters are bird‐like flapping‐wing robots. Only small ornithopters can hover, with long endurance at hummingbird size. Could larger ornithopters be improved further to hover longer? This paper reviews and examines the drive and power of hovering ornithopters, and elastic means of energy or thrust boosters. While the rotation of flexible wings enhance the thrust generation, two‐winged ornithopters did not scale up well because of higher disk loading. In comparison, the X‐winged or multiple‐V‐winged ornithopters enjoy a lower disk loading by beating multiple wings slower, at a smaller stroke angle or a longer span. Further, the clap‐and‐fling interaction of V and X‐wings boosts the thrust generation. Future works can explore the wing flexibility and morphology change to improve the hoverability and flight agility of ornithopters.

Gih‐Keong Lau

Yao‐Wei Chin

Shih‐Chun Lin

Yu‐Hsiang Lai

Boo Cheong Khoo

Spring and Power in Hovering Ornithopters

悬停扑翼飞行器中的弹簧与动力

Online learning aims to solve a sequence of consecutive prediction tasks by leveraging the knowledge gained from previous tasks. Linearized confidence‐weighted (LCW) learning is the first online learning algorithm introducing the concept of weight confidence into the prediction model through distributions over weights. It provides the flexibility for weights to update their values at different scales. The kernel trick in machine learning can be applied to LCW for a better prediction performance. However, the kernel‐based LCW algorithm is subject to the curse of kernelization which makes it vulnerable to the unlimited growth of the prediction model in runtime and memory consumption. In this study, we present the budgeted LCW (BLCW) algorithm which puts a limit on the growth by a predefined budget with optimization. Consequently, BLCW performs the LCW update and then reduces the information loss by projection. Based on the resource perspective that reinterprets LCW in terms of resources and utilization degrees, we demonstrated that BLCW approximates the kernel‐based LCW algorithm. We evaluate four budget maintenance strategies and suggest that the mean removal is the most stable. By various numerical experiments on real datasets, we demonstrate that BLCW performs competitively and effectively when compared to leading budgeted online algorithms.

Jacky Chung‐Hao Wu

Yu‐Shiou Lin

Henry Horng‐Shing Lu

Hsueh‐Ming Hang

Online Linearized Confidence‐Weighted Learning on a Budget

在线线性化置信加权学习预算

The study of cause and effect is of the utmost importance in many branches of science, but also for many practical applications of intelligent systems. In particular, identifying causal relationships in situations that include hidden factors is a major challenge for methods that rely solely on observational data for building causal models. This article proposes CAnDOIT, a causal discovery method to reconstruct causal models using both observational and interventional time‐series data. The use of interventional data in the causal analysis is crucial for real‐world applications, such as robotics, where the scenario is highly complex and observational data alone are often insufficient to uncover the correct causal structure. Validation of the method is performed initially on randomly generated synthetic models and subsequently on a well‐known benchmark for causal structure learning in a robotic manipulation environment. The experiments demonstrate that the approach can effectively handle data from interventions and exploit them to enhance the accuracy of the causal analysis. A Python implementation of CAnDOIT is developed and is publicly available on GitHub: https://github.com/lcastri/causalflow .

Luca Castri

Sariah Mghames

Marc Hanheide

Nicola Bellotto

CAnDOIT: Causal Discovery with Observational and Interventional Data from Time Series

CAnDOIT：基于时间序列观测和干预数据的因果发现

Conventional techniques for extracting physics‐based model parameters are inherently slow processes and often yield less accurate model parameters because of the inflexibility of physical equations. This study presents a novel machine learning–based method to accelerate and enhance the accuracy of compact model generation for multiple devices simultaneously. By integrating a Berkeley short‐channel Insulated‐Gate Field‐Effect Transistor (IGFET) model‐common multigate (BSIM) model with an error‐correction neural network, the proposed approach refines predictions for critical electrical behaviors such as drain current and gate capacitance. Extraction networks dynamically optimize parameter sets for both models, eliminating manual tuning and reducing the need for separate training for each device. The method was validated using TCAD‐simulated 3 nm nanosheet field‐effect transistors devices, achieving a mean absolute percentage error of 1.8% for drain current, 2.8% for transconductance, 8.5% for output conductance, and 0.4% for capacitances. Compared with the BSIM model alone, error reductions of 75, 70, 39, 85, and 81%, respectively, were achieved. This approach showed significant error reductions compared to the BSIM model alone and demonstrated robust performance across devices with variations, proving its effectiveness for large‐scale applications.

Seungjoon Eom

Seunghwan Lee

Hyeok Yun

Kyeongrae Cho

Soomin Kim

Rockhyun Baek

Machine Learning‐Driven Extraction of Hybrid Compact Models Integrating Neural Networks and Berkeley Short‐Channel Insulated‐Gate Field‐Effect Transistor Model‐Common Multigate for Multidevice Applications

机器学习驱动的混合紧凑模型提取：集成神经网络与伯克利短沟道绝缘栅场效应晶体管模型-通用多栅极模型，适用于多器件应用

The ability of reconfigurable robots to adapt to varying tasks and environments furnishes versatility and efficiency in their operations. In this article, tiling robots are introduced as a novel paradigm of shape‐morphing reconfigurable robots, defining them as polyform‐inspired machines capable of transforming between at least two polymorphic shapes. In the study, the existing and future designs of tiling robots by varying base shapes and polygon selections are explored, identifying a significant gap for further exploration of polyforms in their design. The various reconfiguration‐enabling mechanisms and locomotion mechanisms of tiling robots are comparatively analyzed. Summarized electromechanical developments, along with a proposed generalized kinematic model and control scheme, contribute to a comprehensive understanding of tiling robots. A comparison of tiling robots with other established reconfigurable robots is conducted to position tiling robotics within the broader landscape of reconfigurable robotics. The introduction of a new naming convention addresses the absence of a standardized nomenclature for tiling robots. In this article, highlighting the current focus on area coverage in autonomy algorithms of tiling robots, future developments in diverse application domains like logistics, entertainment, and education are anticipated, emphasizing the adaptability of tiling robots as a critical feature for their proliferation across various domains.

短名	Adv. Intell. Syst.
Journal Impact	6.69
国际分区	ROBOTICS(Q1)
期刊索引	SCI Q1中科院 3 区

ISSN	2640-4567
h-index	8
国内分区	计算机科学(3区)计算机科学自动化与控制系统(3区)计算机科学计算机人工智能(3区)计算机科学机器人学(3区)

涉及主题	计算机科学工程类物理人工智能材料科学量子力学电气工程生物化学纳米技术数学操作系统复合材料医学机器人
出版信息	出版商: John Wiley and Sons Inc，出版周期: ，期刊类型: journal，开源期刊: 是
基本数据	创刊年份: 2023，原创研究文献占比: 92.27%，自引率:8.80%， Gold OA占比: 86.43%
期刊费用	$3510详情
平均审稿周期	网友分享经验：4 Weeks

Advanced Intelligent Systems

期刊引文格式

有 1 位以上作者的期刊

有 2 位作者的期刊

有 3 位作者的期刊

有 5 位以上作者的期刊

书籍引用格式

学位论文引用格式

网页引用格式

专利引用格式

手工熬夜修改参考文献？研飞自动匹配期刊，一键轻松成稿，支持 Word/WPS

投稿经验分享