Palladium‐Based Nanomaterials: A Platform to Produce Reactive Oxygen Species for Catalyzing Oxidation Reactions

Oxidation reactions by molecular oxygen (O₂) over palladium (Pd)‐based nanomaterials are a series of processes crucial to the synthesis of fine chemicals. In the past decades, investigations of related catalytic materials have mainly been focused on the synthesis of Pd‐based nanomaterials from the angle of tailoring their surface structures, compositions and supporting materials, in efforts to improve their activities in organic reactions. From the perspective of rational materials design, it is imperative to address the fundamental issues associated with catalyst performance, one of which should be oxygen activation by Pd‐based nanomaterials. Here, the fundamentals that account for the transformation from O₂ to reactive oxygen species over Pd, with a focus on singlet O₂ and its analogue, are introduced. Methods for detecting and differentiating species are also presented to facilitate future fundamental research. Key factors for tuning the oxygen activation efficiencies of catalytic materials are then outlined, and recent developments in Pd‐catalyzed oxygen‐related organic reactions are summarized in alignment with each key factor. To close, we discuss the challenges and opportunities for photocatalysis research at this unique intersection as well as the potential impact on other research fields.     

基于循环神经网络的汉语语言模型建模方法

语言模型是语音识别系统的重要组成部分,目前的主流是n-gram模型。然而n-gram模型存在一些不足,对语句中长距信息描述差、数据稀疏是影响模型性能的两个重要因素。针对不足,研究者提出循环神经网络(Recurrent Neural Network,RNN)建模技术,在英语语言模型建模上取得了较好的效果。根据汉语特点将RNN建模方法应用于汉语语言建模,并结合两种模型的优点,提出了模型融合构建方法。实验结果表明:相比传统的n-gram语言模型,采用RNN训练的汉语语言模型困惑度(Per PLexity,PPL)有了下降,在对汉语电话信道的语音识别上,系统错误率也有下降,将两种语言模型融合后,系统识别错误率更低。     

100μL微量移液器国际关键比对的数学模型研究

为验证各国对微量移液器的校准能力,组织了国际关键比对。通过比对发现大气压力对移液器的校准结果有着较大影响,为此对数学模型进行了修正,经过修正后计算所得各参比实验室的校准结果有了较好的一致性。同时通过大量实验得出,微量移液器由于其自身结构存在不稳定性所带入的不确定度分量是不确定度分析中的一个主要来源。比对中研究所得的数学模型其算法准确、结果可靠,很好地解决了移液器在日常检定及校准中的数据处理问题。