A5101530396- Manufacturing Process and Optimization
- Scheduling and Optimization Algorithms
- Advanced Neural Network Applications
- Advanced Manufacturing and Logistics Optimization
- Metal Forming Simulation Techniques
- Industrial Technology and Control Systems
- Brain Tumor Detection and Classification
- Engineering Technology and Methodologies
- CCD and CMOS Imaging Sensors
- Advanced Computational Techniques and Applications
- Metallurgy and Material Forming
- Advanced Sensor and Control Systems
- Microstructure and mechanical properties
- Advanced Memory and Neural Computing
- Collaboration in agile enterprises
- Advanced machining processes and optimization
- Advanced Image Processing Techniques
- Advanced MRI Techniques and Applications
- Advanced Measurement and Detection Methods
- Rough Sets and Fuzzy Logic
- Advancements in Battery Materials
- Adversarial Robustness in Machine Learning
- Domain Adaptation and Few-Shot Learning
- Resource-Constrained Project Scheduling
- Simulation and Modeling Applications
ShanghaiTech University
2024
Affiliated Hospital of Taishan Medical University
2023
Shandong First Medical University
2023
Nanjing University
2019-2022
Capital Medical University
2022
Central South University
2005-2021
Nanjing General Hospital of Nanjing Military Command
2021
Tongji University
2019
Harbin Institute of Technology
2003-2013
China Telecom (China)
2008
Abstract Self‐supporting Sn foil is a promising high‐volumetric‐capacity anode for lithium ion batteries (LIBs), but it suffers from low initial Coulombic efficiency (ICE). Here, mechanical prelithiation adopted to improve ICE, and found that foils with coarser grains are prone cause electrode damage. To mitigate damage prepare thinner lithiated electrodes, 3Ag0.5Cu96.5Sn used has more refined (5–10 µm) instead of (50–100 µm), where the abundant grain boundaries (GBs) offer sliding systems...
Epilepsy is a chronic brain disease that causes persistent and severe damage to the physical mental health of patients. Daily effective prediction epileptic seizures crucial for epilepsy patients especially those with refractory epilepsy. At present, large number deep learning algorithms such as Convolutional Neural Networks Recurrent have been used predict obtained better performance than traditional machine methods. However, these methods usually transform Electroencephalogram (EEG) signal...
Quantization is a promising technique to compress the size of Convolutional Neural Network (CNN) models. Recently, various precision-scalable designs have been presented reduce computational complexity in CNNs. However, most them adopt straightforward calculation scheme implement CNN, which causes high bandwidth requirement and low hardware utilization efficiency. This paper proposes new architecture can fully CNN inference meanwhile has finely simplified scheme. Based on proposed scheme,...
Clinical radiotherapy (RT) is severely limited by hypoxic tumor microenvironment and a lack of targeting precision. Therefore, it crucial to develop highly efficient radiosensitizers enhance RT efficacy. Herein, novel kind epidermal growth factor receptor (EGFR)-antagonistic affibody-functionalized Pt-based nanozyme for sensitization EGFR-positive tumors was developed. In this study, porous platinum nanoparticles (pPt NPs) featuring catalase (CAT)-like activity strong radiation energy...
Low-latency and low-power implementations of Convolutional Neural Network (CNN) are highly desired for budget-restricted scenarios. Pruning Winograd algorithm two representative approaches to reduce the computation complexity CNNs. Coupling them is very attractive, but transformation removes data sparsity brought by pruning. In this paper, we present a low-latency sparse-Winograd CNN accelerator (LSW-CNN) pruned Wino-grad models. The ReLU-modified employed solve zero refilling issue. Our...
Due to the intensive computational complexity and various types of convolution, it is a challange implement different CNN models on specific hardware. Many previous works focus data reuse sparsity exploration accelerate computation but fail support convolution efficiently. When dealing with variants conventional such as deconvolution or dilated accelerators waste time padding zeroes convolving padded feature maps. In this paper, we propose unified algorithm intelligently combine several...
Abstract The purpose of the current study was to explore feasibility training a deep neural network accelerate process generating T1, T2, and T1ρ maps for recently proposed free‐breathing cardiac multiparametric mapping technique, where recurrent (RNN) utilized exploit temporal correlation among multicontrast images. RNN‐based model developed rapid accurate estimation. Bloch simulation performed simulate dataset more than 10 million signals time correspondences with different noise levels...
Setup planning is a complex and intuitive process. An efficient setup system essential to cut costs reduce lead-time. In this paper, an integrated based on Internet for presented. The implemented simple three-tier thin client-fat server architecture. utilizing of (extensible modeling language) XML as file format provides means the transfer information between various manufacturing systems. A methodology also presented, in which machining precedence, tolerance requirement fixturing...
Modular fixtures are one of the important aspects manufacturing. This paper presents a desktop VR system for modular fixture design. The virtual environment is designed and design procedure proposed. It assists designer to make feasible decisions effectively efficiently. A hierarchical data model proposed represent assembly. Based on this structure, user can manipulate models precisely in VE during assembly processes. Moreover, machining simulation manufacturing interaction checking...
This work investigated the tensile characteristics of plain C-Mn steel with an ultrafine grained ferrite/cementite (UGF/C) microstructure and coarse-grained ferrite/pearlite (CGF/P) microstructure. The tests were performed at temperatures between 77 K 323 K. lower yield ultimate strengths significantly increased when was changed from CGF/P to UGF/C microstructures, but total elongation uniform decreased. A microstructural change had influence on athermal component not thermal component. a...
In multi-project environment, multiple project groups share and compete for the limited resources to achieve their own goals. order address dynamic requirements of production, this paper develops a planning scheduling system using distributed multi-agent approach. The hybrid framework is presented discussed with detail. This consists six types agents facilitated different functionalities. These generic are dynamically deployed customized at each location where executed. Further more,...
In this paper, a fuzzy set theory based intelligent approach for setup planning in manufacturing is introduced. The problem decomposed into three sub tasks the proposed approach: generation, operation sequence and sequence. setups are generated according to optimal machining direction of each feature, which determined by comprehensive judgment method. Using production rules theory, feature precedence relationships matrix (FPR) formed considering main influence factors such as geometry, datum...
In order to complete maintainability design and prediction in stage, case based reasoning (CBR), as an artificial intelligence (AI) method, is used this paper. The of structure considered stage. So the cases including not only function factors but also are presented by frame. retrieval algorithm about similar put forward through two layers: factors. With CBR, method could help designers, referencing previous cases, make a conceptual quickly, finally finish detailed prediction. paper new predict
As an approximation to SDSCI [static-dynamic-static (SDS) configuration interaction (CI), a minimal MRCI; Theor. Chem. Acc. 133, 1481 (2014)], SDSPT2 [Mol. Phys. 115, 2696 (2017)] is CI-like multireference (MR) second-order perturbation theory (PT2) that treats single and multiple roots on equal footing. This feature permits the use of selection over large complete active space (CAS) $P$ end up with much reduced reference $\tilde{P}$, which connected only portion ($\tilde{Q}_1$) full...
In order to plan, manage, coordinate, and control the concurrent design process of CAD/CAPP effectively, this paper proposes a Petri net based model as well its abstract model. The proposed has been validated using simulation tool - HPSim. Its characteristics are then discussed four modes described
With the popularity of deep learning, hardware implementation platform learning has received increasing interest. Unlike general purpose devices, e.g., CPU or GPU, where algorithms are executed at software level, neural network accelerators directly execute to achieve higher energy efficiency and performance improvements. However, as evolve frequently, engineering effort cost designing greatly increased. To improve design quality while saving cost, automation for was proposed, space...
This study intends to propose an intelligent system with the integration of workflow modeling technology and max-min ant (MMAS) in order realize visualization intelligence computer aided process planning (CAPP). Workflow was applied represent multi-process routes decision-making problem (MRDP) a modified colony algorithm proposed exploit best solution by improved pheromone strategy way updating pheromone. Finally, simulation results were reported indeed had admirable performance for MRDP.
Parts locating design involves a lot of information, using the rule based reasoning (RBR) alone to determine part is not reliable when facing with large amount fuzzy information. For solving problem, this paper presents new hybrid approach comprehensive judgment method and select modes references automatically for 3D parts. The uses RBR find all potential candidates modes, location each mode. provided optimally surfaces. realization described in detail. An example presented illustrate issues...
It is a crucial issue that constructing successful knowledge base to satisfy an efficient adaptive scheduling for the complex manufacturing system. Therefore, hybrid inductive learning-based acquisition algorithm presented in this paper. We combined genetic (GA) with simulated annealing (SA) develop optimization method, which GA was introduced present parallel search architecture and SA increase escaping probability from local optima ability neighbor search. The method utilized resolve...
This paper presents an intelligent system to automate the task of setup planning in Internet environment. The requirements a are discussed and how proposed meets is also described. In developing organization, WWW technology, including XML markup language Java programming utilized. By that, seamless integration with other manufacturing systems realized. Furthermore, multi-objective optimization approach based on fuzzy set theory for introduced. this approach, problem decomposed into three...