查看更多>>摘要：2024 OCT 03 (NewsRx)-By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News-Fresh data on Machine Learning are presented in a new report. According to news reporting originating in Dalian, People's Republi c of China, by NewsRx journalists, research stated, "Fundamental understandings of the relationship between ion-electrode interaction and structural feature in porous carbon electrodes at a molecular level provides guidelines for the design of high-performance electric double layer supercapacitors. It is certified by e xperiments that porous carbon structures doped with nitrogen show enhanced capac itive performance." Financial support for this research came from National Natural Science Foundatio n of China (NSFC). The news reporters obtained a quote from the research from the Dalian University of Technology, "However, in the theoretical simulations, the fundamental charge storage mechanism is still elusive. In particular, the recent experimental resu lt shows that the generally ignored nitrolic nitrogen (N5) in porous carbon exhi bits a positive effect on capacitance, while graphitic nitrogen (N3) does the op posite, which is against with the simulation results based the 2D-modeled porous graphene structure. Here, we perform ab initio molecular dynamics simulations o n the N3 and N5-doped carbon/electrolyte interfaces, including both 2D planar an d 3D microporous carbon electrodes. Our calculation indicates that N3 in the 3D pore hinders the electrolyte transport, while N5-doped micropore still serves as an electrolyte transport channel through the formation of H-bond. The charge st orage mechanism is further elucidated by the analysis of the well equilibrated i nterfaces obtained from the machine learning force field accelerated molecular d ynamics. Our work provides a new insight into the effect of nitrogen doping in 3 D porous carbon, which is exactly opposite to the 2D planar graphene."
查看更多>>摘要：2024 OCT 03 (NewsRx)-By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-Investigators publish new report on Ro botics - Robotics and Automation. According to news reporting from Berkeley, Cal ifornia, by NewsRx journalists, research stated, "Motion planners are essential for the safe operation of automated vehicles across various scenarios. However, no motion planning algorithm has achieved perfection in the literature, and impr oving its performance is often time-consuming and labor-intensive." Financial supporters for this research include German Federal Ministry for Digit al and Transport (BMDV) for the project KoSi, German Research Foundation (DFG), Berkeley Deep Drive. The news correspondents obtained a quote from the research from the University o f California Berkeley, "To tackle the aforementioned issues, we present $ {\mathtt {DrPlanner} }$, the first framework designed to automatically diag nose and repair motion planners using large language models. Initially, we gener ate a structured description of the planner and its planned trajectories from bo th natural and programming languages. Leveraging the profound capabilities of la rge language models, our framework returns repaired planners with detailed diagn ostic descriptions. Furthermore, our framework advances iteratively with continu ous feedback from the evaluation of the repaired outcomes."
查看更多>>摘要：2024 OCT 03 (NewsRx)-By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-New research on Machine Learning is th e subject of a report. According to news originating from Wuxi, People's Republi c of China, by NewsRx correspondents, research stated, "The 3D reconstruction is generally defined as the process of capturing the shape and appearance of real objects. By reconstructing 3D digital model from a series of 2D slices, it bring s considerable convenience to visualize internal structure and decipher structur e-property relation of a material." Our news journalists obtained a quote from the research from Jiangnan University , "Nowadays, the 3D reconstruction is becoming a cutting-edge technique in depic ting the internal structure and evaluating the physical performance of targeted materials. Recent years, generative machine learning methods, such as generative adversarial networks (GAN), have achieved tremendous success in AI-generated ph ysical content (AIGPC). However, lots of technical challenges remain, including oversimplified models, oversized dataset requirement and inefficient convergence . These difficulties are caused by the insufficient ability to capture detailed features and the inadequacy of the generated model quality. To this end, a novel generative model is developed, which combines the multiscale features of U-net and the synthesis ability of GANs. With the help of the multiscale channel aggre gation module, the hierarchical feature aggregation module and the convolutional block attention module, our model is developed to capture the features of the m aterial microstructure better. The loss function is refined by combining the ima ge regularization loss with the Wasserstein distance loss. In addition, the anis otropy index is adopted to measure anisotropic degree of selected samples quanti tively. The results demonstrate that the 3D structures generated by the proposed model retain high fidelity with ground truth samples."
查看更多>>摘要：2024 OCT 03 (NewsRx)-By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-New study results on artificial intell igence have been published. According to news originating from Bharathidasan Uni versity by NewsRx correspondents, research stated, "The purpose of this study is to investigate the factors that influence users' behavioral intentions to adopt Artificial Intelligence for personal financial management using an Innovation D iffusion Theory (IDT) framework." Our news reporters obtained a quote from the research from Bharathidasan Univers ity: "An analysis of empirical data collected from 246 users is conducted using Partial Least Squares Structural Equation Modeling (PLS-SEM). Findings reveal th at users' intentions to embrace AI in financial management are influenced by a v ariety of factors, including relative advantage, compatibility, and observabilit y."
查看更多>>摘要：2024 OCT 03 (NewsRx)-By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-Investigators publish new report on Ro botics - Robotics and Automation. According to news reporting originating in Tia njin, People's Republic of China, by NewsRx journalists, research stated, "With proper assistance, knee exoskeletons can benefit humans with impaired leg functi on. Prior studies found that misalignment between the knee and the exoskeleton m ay cause harm and undermine assistance performance." Financial support for this research came from National Natural Science Foundatio n of China (NSFC). The news reporters obtained a quote from the research from Nankai University, "S elf-aligning mechanisms can reduce misalignment but implementing them with a sim ple and lightweight design remains a challenge. In this letter, we designed a li ghtweight (740 g) knee exoskeleton that can provide assistive torque for knee ex tension and flexion during walking. A compact elastic limiter was proposed to re duce misalignment using an integrated shock-absorbing slider. Drawing inspiratio n from the cruciate ligaments, springs were equipped to reduce the nonlinearitie s in the alignment process. With selected spring stiffness, preliminary pressure measurements between the frames and the human body showed the exoskeleton can r educe undesired interaction forces by 50.9% during walking assista nce. Preliminary experimental results demonstrated the exoskeleton can achieve h igh torque control performance and reduce quadriceps activities during multiple gaits. For example, the activities of the rectus femoris, vastus medialis, and v astus lateralis were reduced by an average of 23.8%, 44.0% , and 34.0%, respectively during 8(degrees) uphill walking."
查看更多>>摘要：2024 OCT 03 (NewsRx)-By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-Researchers detail new data in Robotic s. According to news reporting from Rome, Italy, by NewsRx journalists, research stated, "Vision-based perception systems are typically exposed to large orienta tion changes in different robot applications. In such conditions, their performa nce might be compromised due to the inherent complexity of processing data captu red under challenging motion." Funders for this research include European Research Council (ERC), German Resear ch Foundation (DFG). The news correspondents obtained a quote from the research from Leonardo S.p.A., "Integration of mechanical stabilizers to compensate for the camera rotation is not always possible due to the robot payload constraints. This letter presents a processing-based stabilization approach to compensate the camera's rotational motion both on events and on frames (i.e., images). Assuming that the camera's a ttitude is available, we evaluate the benefits of stabilization in two perceptio n applications: feature tracking and estimating the translation component of the camera's ego-motion. The validation is performed using synthetic data and seque nces from well-known event-based vision datasets. The experiments unveil that st abilization can improve feature tracking and camera ego-motion estimation accura cy in 27.37% and 34.82%, respectively."
查看更多>>摘要：2024 OCT 03 (NewsRx)-By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-Researchers detail new data in Machine Learning. According to news originating from Shaanxi, People's Republic of Chin a, by NewsRx correspondents, research stated, "Compared with the traditional ele ctrocatalyst screening of the nitrogen reduction reaction (NRR), machine learnin g (ML) has achieved high-throughput screening with less computational cost. In t his paper, 140 TM@g-CxNy single-atom catalysts (SACs) are constructed for the NR R."
查看更多>>摘要：2024 OCT 03 (NewsRx)-By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-Fresh data on Robotics are presented i n a new report. According to news originating from Arrasate Mondragon, Spain, by NewsRx correspondents, research stated, "Many daily tasks exhibit a periodic na ture, necessitating that robots possess the ability to execute them either alone or in collaboration with humans. A widely used approach to encode and learn suc h periodic patterns from human demonstrations is through periodic Dynamic Moveme nt Primitives (DMPs)." Funders for this research include Basque Government, European Union (EU), Vinnov a. Our news journalists obtained a quote from the research from Mondragon Unibertsi tatea, "Periodic DMPs encode cyclic data independently across multiple dimension s of multi-degree of freedom systems. This method is effective for simple data, like Cartesian or joint position trajectories. However, it cannot account for va rious geometric constraints imposed by more complex data, such as orientation an d stiffness. To bridge this gap, we propose a novel periodic DMP formulation tha t enables the encoding of periodic orientation trajectories and varying stiffnes s matrices while considering their geometric constraints. Our geometry-aware app roach exploits the properties of the Riemannian manifold and Lie group to direct ly encode such periodic data while respecting its inherent geometric constraints . We initially employed simulation to validate the technical aspects of the prop osed method thoroughly."
查看更多>>摘要：2024 OCT 03 (NewsRx)-By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-Current study results on Machine Learn ing have been published. According to news reporting originating from Ottawa, Ca nada, by NewsRx correspondents, research stated, "Intelligent dataplanes (IDPs) can enhance network service performance and adaptation speed by executing one o r more machine learning (ML) models directly on the served flows. The real-time ML inference enables line-speed decision-making for some traffic management func tionalities." Financial support for this research came from Natural Sciences and Engineering R esearch Council of Canada (NSERC). Our news editors obtained a quote from the research from the University of Ottaw a, "Due to the inherent scarcity of both the computational and memory resources and the strict high-speed per-packet processing demands, existing IDP deployment s either realize only a limited set of ML models such as decision trees, or requ ire substantial modifications in the switch hardware. In this paper, we propose INQ-MLT, a novel ML-based management toolbox to address the aforementioned limit ations. INQ-MLT delegates the task of training various ML models to the control- plane. The latter adopts a tailored quantization-aware training process to compe nsate for the effect of precision loss resulting from quantization. The toolbox then employs a quantization mechanism to transform the trained ML model paramete rs (e.g., weights and activations) from floating-point representations to compac t low-precision fixed integer values that can be easily processed and stored in the data-plane. Finally, the trained model is deployed into the IDP pipeline by restricting all its inference operations to basic arithmetic operations. To anal yze the performance of INQ-MLT, we quantify the accuracy loss resulting from the quantization step through rigorous theoretical analysis. A proof-of-concept imp lementation of the proposed toolbox is developed using P4-based software switche s."
查看更多>>摘要：2024 OCT 03 (NewsRx)-By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-Current study results on Robotics - Ro botics and Automation have been published. According to news reporting from Harb in, People's Republic of China, by NewsRx journalists, research stated, "Samplin g-based motion planning (SBMP) algorithms are renowned for their robust global s earch capabilities. However, the inherent randomness in their sampling mechanism s often results in inconsistent path quality and limited search efficiency." Financial support for this research came from National Natural Science Foundatio n of China (NSFC). The news correspondents obtained a quote from the research from the Harbin Insti tute of Technology, "In response to these challenges, this work proposes a novel deep learning-based motion planning framework, named Transformer-Enhanced Motio n Planner (TEMP), which synergizes a Co-Regulation Environmental Information Enc oder (CEIE) with a Motion Planning Transformer (MPT). CEIE converts scenario dat a into encoded environmental information (EEI), providing MPT with an insightful understanding of the environment. MPT leverages an attention mechanism to dynam ically recalibrate its focus on EEI, task objectives, and historical planning da ta, refining the sampling node generation. To demonstrate the capabilities of TE MP, we train our model using a dataset consisting of planning results produced b y RRT*. CEIE and MPT are collaboratively trained, enabling CEIE to autonomously learn and extract patterns from environmental data, thereby forming informative representations that MPT can more effectively interpret and utilize for motion p lanning."
NETL
NSTL