基于联合神经网络学习的中文电力计量命名实体识别

随着电力计量业务的不断扩展，迫切需要由业务信息、技术知识、行业标准及其内在联系所组成的电力计量知识图谱，为电网的决策和发展提供更为全面有效的支持。命名实体识别是构建知识图谱的基础。针对电力计量领域需要，结合中文分词技术特点，基于联合学习思想，提出了一种基于联合学习的中文电力计量命名实体识别技术。该技术联合CNN-BLSTM-CRF模型与整合词典知识的分词模型，使其共享实体类别和置信度；同时将2个模型的先后计算顺序改为并行计算，减少了识别误差累积。结果表明，在不需要人工构建特征的情况下，方法的正确率、召回率、F值等均显著优于以往方法。     

Visible-light-induced,autopromoted nickel-catalyzed three-component arylsulfonation of 1,3-enynes and mechanistic insights

Science China Chemistry - A light-induced, nickel-catalyzed three-component arylsulfonation of 1,3-enynes in the absence of photocatalyst is reported. This methodology exhibited mild conditions,...     

Stochastic relaxed inertial forward-backward-forward splitting for monotone inclusions in Hilbert spaces

Computational Optimization and Applications - We develop a globalized Proximal Newton method for composite and possibly non-convex minimization problems in Hilbert spaces. Additionally, we impose...     

A New Carotane Sesquiterpenoid Derivative with Antifungal Activity from Gymnopus sp. 0612-9

Chemistry of Natural Compounds - A new carotane sesquiterpenoid, methyl gymnabloate (1), was isolated from the fermentation broth of the wild macrofungus Gymnopus sp. 0612-9. Its structure was...     

Y掺杂WS2二维材料的第一性原理研究

WS₂由于其优异的物理和光电性质引起了广泛关注。本研究基于第一性原理计算方法,探索了本征单层WS₂及不同浓度W原子替位钇(Y)掺杂WS₂的电子结构和光学特性。结果表明本征单层WS₂为带隙1.814 eV的直接带隙半导体。进行4%浓度(原子数分数)的Y原子掺杂后,带隙减小为1.508 eV,依旧保持着直接带隙的特性,随着Y掺杂浓度的不断增大,掺杂WS₂带隙进一步减小,当浓度达到25%时,能带结构转变为0.658 eV的间接带隙,WS₂表现出磁性。适量浓度的掺杂可以提高材料的导电性能,且掺杂浓度增大时,体系依旧保持着透明性并且在红外光和可见光区对光子的吸收能力、材料的介电性能都有着显著提高。本文为WS₂二维材料相关光电器件的研究提供了理论依据。     

Detecting pore size distribution of activated carbon by low-field nuclear magnetic resonance

In this study, the transverse relaxation time (T₂) of activated carbon (AC) in different relative environment humidity was detected firstly by low-field nuclear magnetic resonance (LFNMR). The pore size (diameter) of AC distributions was calculated by the relationship between T₂ and surface relaxation rate (ρ), where ρ was obtained by the detection of nine porous materials with known pore size. The results showed that the pore size distributions of AC calculated by ρ < 0.19 nm/ms were in good agreement with that obtained by nitrogen adsorption method and proved that LFNMR as a new detection method was feasible for characterizing AC pore size distribution.     

Effect of temperature on the interactions between cellulose and lignin via molecular dynamics simulations

Cellulose - Heating is essential in various biomass pre-treatments and thermal conversion processes. It is of practical significance to study the characteristics of cellulose-lignin and...     

MWCNTs/hydroxypropyl cellulose/polyethylene glycol-based shape-stabilized phase change materials

Journal of Thermal Analysis and Calorimetry - Polyethylene glycol (PEG) is widely used as phase change materials (PCMs). However, the leakage and low thermal conductivity issues restrict its...