将绩效管理等同于绩效考核,过于看重考核结果,缺乏计划性,缺乏有效的激励和反馈机制等都是当前高校图书馆绩效管理存在的问题.运用人力资源管理的方法,结合"双一流"建设的要求,构建"双一流"背景下高校图书馆绩效管理体系.

根据李诫《营造法式》各版本传世序跋、各藏书家书目记录和校勘成果,在前人研究基础上对1949年前《营造法式》的版本进行考订和系统梳理,并以图示的形式揭示了该书各版本之间的源流关系.

Sentiment classification is an application of sentiment analysis, which is a popular research field in NLP. It can classify documents into different categories according to their sentiments. For a sentiment classification task, the first step is to extract sentimental features from documents, and then classify them using some classifiers. In the first step, a traditional way to extract sentimental features is to apply sentiment dictionaries. However, sentiment words may have different sentiment tendencies in different contexts, and traditional sentiment dictionaries does not consider this situation where wrong sentiment tendencies may be selected for sentiment words. In our research, we find that sentiment words will not have diverse meanings when they associate with the nearby aspects and entities in documents. Then, we propose a three layers sentiment dictionary, which can associate sentiment words with the corresponding entities and aspects together to reduce their multiple meanings. In the second step of the sentiment classification task, many classification models, such as SVM, GBDT, can be used to classify documents according to the extracted sentiment words. However, different classifiers have different weaknesses. A Stacking-based hybrid model is applied to combine SVM and GBDT together to overcome their weaknesses and reach higher performance. This hybrid model contains two layers, and the output of the first layer will become the input of the second layer. The first layer will generate different classification results according to different classifiers, while the second layer will automatically learn how to select a probable one as the final result. The experimental results show that our hybrid model outperforms the baseline single models.

加强特藏资源转化与利用对于处于大数据时代的高校图书馆核心竞争力建设具有重要作用.文章通过分析特藏资源转化与利用的意义、影响因素,总结了特藏资源转化与利用的主要方式,探讨了华南理工大学图书馆特藏资源转化与利用的成功经验.

学术评估乏力导致学术不端事件频频发生,为此产生了各种各样的学术不端行为检测系统.虽然检测系统快速、智能、功能强大,客观上对"学术不端行为"起到了警示和震慑作用,但在实践中发现其仍然存在着一些问题,对此提出了一些解决方法,并建议应结合专家意见合理看待、理性使用检测结果,加强宣传教育,预防为主,建立健全相应的监督体制和诚信档案.

现行高校图书馆绩效考核存在:未与图书馆发展战略规划相结合;考核指标产生方法缺乏科学性;考核指标偏宏观抽象、缺乏针对性等问题.根据调研结果,结合"双一流"的建设要求,将现行岗位分类由管理岗、专业技术岗和工勤岗等三类划分为行政管理岗、专业技术岗、技术管理岗、工勤岗、学生岗等五类.运用德尔菲法,建立起了一套包含4个一级指标、42个二级指标、127个三级指标的高校图书馆人员绩效考核指标体系.

With the development of e-commerce, websites such as Amazon and eBay have become very popular. Users post reviews of products and rate the helpfulness of reviews on these websites. Reviews written by a user and reviews rated by a user reflect the user's interests and disinterest. Thus, they are very useful for user profiling. In this study, the authors explore users' reviews and ratings of reviews for personalized searching and propose a review-based user profiling method. To satisfy a user's basic information needs, expressed in the form of a query, they also propose a priority-based result ranking strategy. For evaluation, they conduct experiments on a real-life data set. The experimental results show that their method can significantly improve retrieval quality.

In computer vision many tasks have achieved state-of-the-art performance using convolutional neural networks (CNNs) [11], typically at the cost of massive computational complexity. A key problem of the training is the low speed of the progress. It may cost much time especially when computational resources are limited. The focus of this paper is speeding up the training progress based on fine-tuned backpropagation progress. More specifically, we train the CNNs with standard backpropagation firstly. When the feature extraction layers got better features, then we start to block the standard backpropagation in the whole layers, the loss function values only back propagates between fully connected layers. So it can not only save time but also pay more attention to train the classifier to get the same or better result compared with training with standard backpropagation all the time. Comprehensive experiments on JD ( https://www.jd.com/ ) datasets demonstrate significant reduction in computational time, at the cost of negligible loss in accuracy.

Multiagent coordination in cooperative multiagent systems, as one of the fundamental problems in multiagent systems, and has been widely studied in the literature. In real environments, the interactions among agents are usually sparse and regulated by their underlying network structure, which, however, has received relatively few attentions in previous work. To this end, we firstly systematically investigate the multiagent coordination problems in cooperative environments under the networked social learning framework under four representative topologies. A networked social learning framework consists of a population of agents where each agent interacts with another agent randomly selected from its neighborhood in each round. Each agent updates its learning policy through repeated interactions with its neighbors via both individual learning and social learning. It is not clear a priori whether all agents are able to learn towards a consistent optimal coordination policy. Two types of learners are proposed: individual action learner and joint action learner. We evaluate the learning performances of both learners extensively in different cooperative (both single-stage and Markov) games. Besides, the influence of different factors (network topologies, different types of games, different topology parameters) is investigated and analyzed and new insights are obtained.

Sentiment analysis and opinion mining is a task to analyze people's opinions or sentiments from textual data, which is very useful for the analysis of many NLP applications. The difficulty of this task is that there are a variety of sentiments inside documents, and these sentiments have variety expressions. Hence, it is hard to extract all sentiments using a dictionary that is commonly used. In this paper, we construct the domain sentiment dictionary using external textual data. Besides, many classification models can be used to classify documents according to their opinion. However, these single models have strengths and weaknesses. We propose a highly effective hybrid model combining different single models to overcome their weaknesses. The experimental results show that our hybrid model outperforms baseline single models.

Graph-based ranking algorithm such as TextRank shows a remarkable effect on keyword extraction. However, these algorithms build graphs only considering the lexical sequence of the documents. Hence, graphs generated by these algorithm can not reflect the semantic relationships between documents. In this paper, we demonstrate that there exists an information loss in the graph-building process from textual documents to graphs. These loss will lead to the misjudgment of the algorithm. In order to solve this problem, we propose a new approach called Topic-based TextRank. Different from the traditional algorithm, our approach takes the lexical meaning of the text unit (i.e. words and phrase) into account. The result of our experiments shows that our proposed algorithm can outperform the state-of-the-art algorithms.

Sentence categorization is a task to classify sentences by their types, which is very useful for the analysis of many NLP applications. There exist grammar or syntactic rules to determine types of sentences. And keywords like negation word for negative sentences is an important feature. However, no all sentences have rules to classify. Besides, different types of sentences may contain the same keywords whose meaning may be changed by context. We address the first issue by proposing a hybrid model consisting of Decision Trees and Support Vector Machines. In addition, we design a new feature based on N-gram model. The results of the experiments conducted on the sentence categorization dataset in "Good Ideas of China" Competition 2015 show that (1) our model outperforms baseline methods and all online systems in this competition; (2) the effectiveness of our feature is higher than that of features frequently used in NLP.

In multiagent systems, social norms is a useful technique in regulating agents' behaviors to achieve coordination or cooperation among agents. One important research question is to investigate how a desirable social norm can be evolved in a bottom-up manner through local interactions. In this paper, we propose two novel learning strategies under the collective learning framework: collective learning EV-l and collective learning EV-g, to efficiently facilitate the emergence of social norms. Experimental results show that both learning strategies can support the emergence of desirable social norms more efficiently in a much broader range of multiagent interaction scenarios than previous work, and also are robust across different network topologies.

Many research works are concentrating in how to handle the opinion mining for English, Chinese and Japanese, etc. However, little work has been done on opinion mining for Cantonese which is a world-wide influential language with 70 million speakers. In this paper, we point out that instead of utilizing the Mandarin lexical database, it’s necessary to construct a particular lexical database for Cantonese. Besides, we explore some Cantonese special written-tradition rules and incorporate them into the feature-based opinion summarization system framework. The experimental results show that our framework significantly outperform the traditional Mandarin sentiment analysis method using ICTCLAS.

近年来,社区问答服务系统(CQA)越来越受到人们的欢迎,但随着提问规模的膨胀,获得回答的问题比重逐步降低,且答案质量无法得到保障.为了提高问答系统中问题被解答的概率,并提升答案可信度,文中提出了基于社交关系相似度的社交问答系统(SQA),主动寻找与提问者社交关系紧密且能够回答问题的用户,并提出了针对提问者与最佳回答者的推荐方法.实验结果表明,在主观性强或实时性强等问题集上,文中方法能更快地得到让提问者满意的答案.

With the development of the Internet, there are more and more shared resources on the Web. Personalized search becomes increasingly important as users demand higher retrieval quality. Personalized search needs to take users' personalized profiles and information needs into consideration. Collaborative tagging (also known as folksonomy) systems allow users to annotate resources with their own tags (features) and thus provide a powerful way for organizing, retrieving and sharing different types of social resources. To capture and understand user preferences, a user is typically modeled as a vector of tag: value pairs (i.e., a tag-based user profile) in collaborative tagging systems. In such a tag-based user profile, a user's preference degree on a group of tags (i.e., a combination of several tags) mainly depends on the preference degree on every individual tag in the group. However, the preference degree on a combination of tags (a tag-group) cannot simply be obtained from linearly combining the preference on each tag. The combination of a user's two favorite tags may not be favorite for the user. In this article, we examine the limitations of previous tag-based personalized search. To overcome their problems, we model a user profile based on combinations of tags (tag-groups) and then apply it to the personalized search. By comparing it with the state-of-the-art methods, experimental results on a real data set shows the effectiveness of our proposed user profile method.

In order to improve the operation efficiency of cache sensitive B +-tree (CSB +-tree)indexing,this pa-per deals with the parallel construction and query performance of CSB +-tree on graphic processing unit (GPU).In the investigation,first,the mapping relationship between each key in internal nodes and the corresponding leaf node of the index tree is analyzed,a lock-free parallel algorithm that once for all builds the CSB +-tree internal node keys is proposed,and the index tree is constructed at the maximum parallel speed.Moreover,dynamic arrays su-pporting the arbitrary expansion of CSB +-tree index data on GPU are designed to implement the dynamic allocation of memory space on GPU,and padding bits are added to the boundary of the internal nodes to reduce the number of branches,thus improving the query efficiency of CSB +-tree.Experimental results indicate that the proposed algo-rithm is 31.0 and 1.4 times faster respectively than the parallel algorithms based on single node and tree layer.

The problem of coordination in cooperative multiagent systems has been widely studied in the literature. In practical complex environments, the interactions among agents are usually regulated by their underlying network topology, which, however, has not been taken into consideration in previous work. To this end, we firstly investigate the multiagent coordination problems in cooperative environments under the networked social learning framework focusing on two representative topologies: the small-world and the scale-free network. We consider a population of agents where each agent interacts with another agent randomly chosen from its neighborhood in each round. Each agent learns its policy through repeated interactions with its neighbors via social learning. It is not clear a priori if all agents can learn a consistent optimal coordination policy and what kind of impact different topology parameters could have on the learning performance of agents. We distinguish two types of learners: individual action learner and joint action learner. The learning performances of both learners are evaluated extensively in different cooperative games, and the influence of different factors on the learning performance of agents is investigated and analyzed as well.

The problem of coordination in cooperative multiagent systems has been widely studied in the literature. We firstly investigate the multiagent coordination problems in cooperative environments under the networked social learning framework focusing on two representative topologies: the small-world and the scale-free network. We consider a population of agents where each agent interacts with another agent randomly chosen from its neighborhood in each round. Each agent learns its policy through repeated interactions with its neighbors via social learning. It is not clear a priori if all agents can learn a consistent optimal coordination policy and what kind of impact different topology parameters could have on the learning performance of agents. We distinguish two types of learners: individual action learner and joint action learner. The learning performances of both learners are evaluated extensively in different cooperative games.

There are many news articles reported online everyday. Within an ongoing topic, people can find a huge amount of news articles. A topic often consists of several events, and people are interested in the whole evolution of a topic along a timeline. This requests for finding and identifying the dependent relationships between events. In order to understand the whole evolution of a topic effectively, we propose a framework of event relationship analysis. We define three kinds of event relationships which are coccurrence dependence relationship, event reference relationship, and temporal proximity relationship for modeling how an event is dependent on another event within a topic. Through combining three kinds of relationships, we can discover an Event Evolution Graph (EEG) for users to view the evolution of a topic. Experiments conducted on a real data set show that our method outperforms baseline methods.