水电资产管理面临复杂性、长周期等挑战,传统的资产管理模式难以满足大规模工程资产管理的要求。通过建立水电资产的建筑信息模型(building information modeling,BIM),形成工程数字资产,实现贯穿水电资产全生命周期的数字化管理。本文...水电资产管理面临复杂性、长周期等挑战,传统的资产管理模式难以满足大规模工程资产管理的要求。通过建立水电资产的建筑信息模型(building information modeling,BIM),形成工程数字资产,实现贯穿水电资产全生命周期的数字化管理。本文以此为基础,解读资产全生命周期管理(asset lifecycle management,ALM)和BIM的含义,剖析水电资产管理的现状与挑战,提出系统的解决方案,以Crossrail、NIEAP、中国长江三峡集团有限公司(简称三峡集团)为对象,重点研究基于BIM的水电资产全生命周期管理系统的应用实践。展开更多
The risk of developing neoplasia leading to colorectal cancer is significantly increased in ulcerative colitis (UC) and most likely in Crohn's disease. Several endoscopic surveillance strategies have been implemen...The risk of developing neoplasia leading to colorectal cancer is significantly increased in ulcerative colitis (UC) and most likely in Crohn's disease. Several endoscopic surveillance strategies have been implemented to identify these lesions. The main issue is that colitisassociated neoplasms often occurs in flat mucosa, often being detected on taking random biopsies rather than by identification of these lesions via endoscopic imaging. The standard diagnostic procedure in long lasting UC is to take four biopsies every 10 cm. Image enhancement methods, such as chromoendoscopy and virtual histology using endomicroscopy, have greatly im- proved neoplasia detection rates and may contribute toreduced random biopsies by taking targeted "smart" biopsies. Chromoendoscopy may effectively be performed by experienced endoscopists for routine screening of UC patients. By contrast, endomicroscopy is often only available in selected specialized endoscopic centers. Importantly, advanced endoscopic imaging has the poten- tial to increase the detection rate of neoplasia whereas the interplay between endoscopic experience and interpretation of histological biopsy evaluation allows the physician to make a proper diagnosis and to find the appropriate therapeutic approach. Colitis-associated intraepithelial neoplasms may occur in flat mucosa of endoscopically normal appearance or may arise as dysplasia-associated lesion or mass (DALM), which may be indistinguishable from sporadic adenomas in healthy or non-colitis mucosa [adenoma-like mass (ALM)]. The aim of this review was to summarize endoscopic and histological characteristics of DALM and ALM in the context of therapeutic procedures.展开更多
FPGA is an appealing platform to accelerate DNN.We survey a range of FPGA chip designs for AI.For DSP module,one type of design is to support low-precision operation,such as 9-bit or 4-bit multiplication.The other typ...FPGA is an appealing platform to accelerate DNN.We survey a range of FPGA chip designs for AI.For DSP module,one type of design is to support low-precision operation,such as 9-bit or 4-bit multiplication.The other type of design of DSP is to support floating point multiply-accumulates(MACs),which guarantee high-accuracy of DNN.For ALM(adaptive logic module)module,one type of design is to support low-precision MACs,three modifications of ALM includes extra carry chain,or 4-bit adder,or shadow multipliers which increase the density of on-chip MAC operation.The other enhancement of ALM or CLB(configurable logic block)is to support BNN(binarized neural network)which is ultra-reduced precision version of DNN.For memory modules which can store weights and activations of DNN,three types of memory are proposed which are embedded memory,in-package HBM(high bandwidth memory)and off-chip memory interfaces,such as DDR4/5.Other designs are new architecture and specialized AI engine.Xilinx ACAP in 7 nm is the first industry adaptive compute acceleration platform.Its AI engine can provide up to 8X silicon compute density.Intel AgileX in 10 nm works coherently with Intel own CPU,which increase computation performance,reduced overhead and latency.展开更多
文摘水电资产管理面临复杂性、长周期等挑战,传统的资产管理模式难以满足大规模工程资产管理的要求。通过建立水电资产的建筑信息模型(building information modeling,BIM),形成工程数字资产,实现贯穿水电资产全生命周期的数字化管理。本文以此为基础,解读资产全生命周期管理(asset lifecycle management,ALM)和BIM的含义,剖析水电资产管理的现状与挑战,提出系统的解决方案,以Crossrail、NIEAP、中国长江三峡集团有限公司(简称三峡集团)为对象,重点研究基于BIM的水电资产全生命周期管理系统的应用实践。
文摘The risk of developing neoplasia leading to colorectal cancer is significantly increased in ulcerative colitis (UC) and most likely in Crohn's disease. Several endoscopic surveillance strategies have been implemented to identify these lesions. The main issue is that colitisassociated neoplasms often occurs in flat mucosa, often being detected on taking random biopsies rather than by identification of these lesions via endoscopic imaging. The standard diagnostic procedure in long lasting UC is to take four biopsies every 10 cm. Image enhancement methods, such as chromoendoscopy and virtual histology using endomicroscopy, have greatly im- proved neoplasia detection rates and may contribute toreduced random biopsies by taking targeted "smart" biopsies. Chromoendoscopy may effectively be performed by experienced endoscopists for routine screening of UC patients. By contrast, endomicroscopy is often only available in selected specialized endoscopic centers. Importantly, advanced endoscopic imaging has the poten- tial to increase the detection rate of neoplasia whereas the interplay between endoscopic experience and interpretation of histological biopsy evaluation allows the physician to make a proper diagnosis and to find the appropriate therapeutic approach. Colitis-associated intraepithelial neoplasms may occur in flat mucosa of endoscopically normal appearance or may arise as dysplasia-associated lesion or mass (DALM), which may be indistinguishable from sporadic adenomas in healthy or non-colitis mucosa [adenoma-like mass (ALM)]. The aim of this review was to summarize endoscopic and histological characteristics of DALM and ALM in the context of therapeutic procedures.
文摘FPGA is an appealing platform to accelerate DNN.We survey a range of FPGA chip designs for AI.For DSP module,one type of design is to support low-precision operation,such as 9-bit or 4-bit multiplication.The other type of design of DSP is to support floating point multiply-accumulates(MACs),which guarantee high-accuracy of DNN.For ALM(adaptive logic module)module,one type of design is to support low-precision MACs,three modifications of ALM includes extra carry chain,or 4-bit adder,or shadow multipliers which increase the density of on-chip MAC operation.The other enhancement of ALM or CLB(configurable logic block)is to support BNN(binarized neural network)which is ultra-reduced precision version of DNN.For memory modules which can store weights and activations of DNN,three types of memory are proposed which are embedded memory,in-package HBM(high bandwidth memory)and off-chip memory interfaces,such as DDR4/5.Other designs are new architecture and specialized AI engine.Xilinx ACAP in 7 nm is the first industry adaptive compute acceleration platform.Its AI engine can provide up to 8X silicon compute density.Intel AgileX in 10 nm works coherently with Intel own CPU,which increase computation performance,reduced overhead and latency.
