中文网络聊天语言的奇异性与动态性研究

随着互联网走入社会生活,网络聊天逐渐成为一种新的沟通渠道,网络聊天语言便应运而生。这类语言的日益丰富,给语言信息处理带来了新的挑战。研究发现,困难主要来自网络聊天语言的奇异性和动态性。本文借助真实网络聊天语言文本,对网络聊天语言的奇异性和动态性进行详细分析和归纳,并设计了面向解决奇异性和动态性问题的网络聊天语言文本识别与转换方法。我们先以网络聊天语言语料库为基础建立网络聊天语言模型和语言转换模型,通过信源–信道模型实现网络聊天语言向标准语言的转换。但该方法过于依赖网络聊天语言语料库,虽然能较好解决奇异性问题,但不能处理动态性问题。因此,我们进而以标准汉语语料库为基础建立文字语音映射模型,对信源–信道模型进行改进,最终有效解决了网络聊天语言的动态性问题。     

基于小规模标注语料的机器学习方法研究

文中通过讨论机器学习和自然语言处理之间的关系，论述了语料库语言工程中机器学习的困境，概述分析了应用半监督学习的现状，研究有限样本下结合未标注样本的方法和统计学习理论框架的结合前景。     

基于语义知识的汉语句法结构排歧

汉语在词类这个语言层次上存在着许多歧义结构,这给汉语的自动句法分析带来了难以逾越的障碍。通过寻找汉语语义类之间可能存在的句法关系建立汉语语义关联网,这为用汉语语义知识来解决句法歧义开辟了道路。文章针对具体的汉语歧义结构研究具体的解决办法,从而减少了计算的复杂度。     

Answering XML Queries Using Path-Based Indexes: A Survey

The problem of answering XML queries using path-based indexes is to find efficient methods for accelerating the XML query with pre-designed index structures over the XML database. This problem received increasing interests and have been lucubrated in recent years. Regular path expression is the core of the XML query languages e.g., XPath and XQuery. Most of the state-of-the-art path-based XML indexes, therefore, hammer at how to efficiently answer the path-based XML queries. This paper surveys various approaches to indexing XML data proposed in the literature. We give a step by step analysis to show the evolution of index structures for XML path information, based on tree structures or more commonly, directed labeled graphs. For each approach, we first present the specific issue it aims to tackle, and then the proposed solution presented. Furthermore, construction, physical data storage and maintenance costs, are analyzed.     

Normalization of Chinese chat language

Real-time communication platforms such as ICQ, MSN and online chat rooms are getting more popular than ever on the Internet. There are, however, real risks where criminals and terrorists can perpetrate illegal and criminal abuses. This highlights the security significance of accurate detection and translation of the chat language to its stand language counterpart. The language used on these platforms differs significantly from the standard language. This language, referred to as chat language, is comparatively informal, anomalous and dynamic. Such features render conventional language resources such as dictionaries, and processing tools such as parsers ineffective. In this paper, we present the NIL corpus, a chat language text collection annotated to facilitate training and testing of chat language processing algorithms. We analyse the NIL corpus to study the linguistic characteristics and contextual behaviour of a chat language. First we observe that majority of the chat terms, i.e. informal words in a chat text, is formed by phonetic mapping. We then propose the eXtended Source Channel Model (XSCM) for the normalization of the chat language, which is a process to convert messages expressed in a chat language to its standard language counterpart. Experimental results indicate that the performance of XSCM in terms of chat term recognition and normalization accuracy is superior to its Source Channel Model (SCM) counterparts, and is also more consistent over time.

Glycyrrhizin Represses Total Parenteral Nutrition-Associated Acute Liver Injury in Rats by Suppressing Endoplasmic Reticulum Stress

Total parenteral nutrition (TPN) is an artificial way to support daily nutritional requirements by bypassing the digestive system, but long-term TPN administration may cause severe liver dysfunction. Glycyrrhizin is an active component of licorice root that has been widely used to treat chronic hepatitis. The aim of this study is to investigate the hepatoprotective effect of glycyrrhizin on TPN-associated acute liver injury in vivo. Liver dysfunction was induced by intravenous infusion of TPN at a flow rate of 20 mL/kg/h for three h in Sprague Dawley rats. The rats were pretreated with Glycyrrhizin (1, 3 and 10 mg/kg intravenously). After receiving TPN or saline (control group) for three h, the rats were sacrificed, blood samples were collected for biochemical analyses and liver tissue was removed for histopathological and immunohistochemical examination. We found that aspartate aminotransferase (AST), alanine aminotransferase (ALT), total bilirubin (TB) and triglyceride (TG) levels were significantly increased in the TPN group without glycyrrhizin pretreatment and decreased in the glycyrrhizin-pretreated TPN group in a dose-dependent manner. The stained liver sections showed that glycyrrhizin relieved acute liver injury. The upregulation of serum protein biomarkers of reactive nitrogen species, including nitrotyrosine and inducible NO synthase (iNOS), were attenuated by glycyrrhizin pretreatment. Levels of endoplasmic reticulum (ER) stress factors, such as phosphorylation of JNK1/2, p38 MAPK and CHOP, were decreased by glycyrrhizin pretreatment. In summary, our results suggest that glycyrrhizin decreases TPN-associated acute liver injury factors by suppressing endoplasmic reticulum stress and reactive nitrogen stress.     

Global Proteomic Analysis of Brain Tissues in Transient Ischemia Brain Damage in Rats

Ischemia-reperfusion injury resulting from arterial occlusion or hypotension in patients leads to tissue hypoxia with glucose deprivation, which causes endoplasmic reticulum (ER) stress and neuronal death. A proteomic approach was used to identify the differentially expressed proteins in the brain of rats following a global ischemic stroke. The mechanisms involved the action in apoptotic and ER stress pathways. Rats were treated with ischemia-reperfusion brain injuries by the bilateral occlusion of the common carotid artery. The cortical neuron proteins from the stroke animal model (SAM) and the control rats were separated using two-dimensional gel electrophoresis (2-DE) to purify and identify the protein profiles. Our results demonstrated that the SAM rats experienced brain cell death in the ischemic core. Fifteen proteins were expressed differentially between the SAM rats and control rats, which were assayed and validated in vivo and in vitro. Interestingly, the set of differentially expressed, down-regulated proteins included catechol O-methyltransferase (COMT) and cathepsin D (CATD), which are implicated in oxidative stress, inflammatory response and apoptosis. After an ischemic stroke, one protein spot, namely the calretinin (CALB2) protein, showed increased expression. It mediated the effects of SAM administration on the apoptotic and ER stress pathways. Our results demonstrate that the ischemic injury of neuronal cells increased cell cytoxicity and apoptosis, which were accompanied by sustained activation of the IRE1-alpha/TRAF2, JNK1/2, and p38 MAPK pathways. Proteomic analysis suggested that the differential expression of CALB2 during a global ischemic stroke could be involved in the mechanisms of ER stress-induced neuronal cell apoptosis, which occurred via IRE1-alpha/TRAF2 complex formation, with activation of JNK1/2 and p38 MAPK. Based on these results, we also provide the molecular evidence supporting the ischemia-reperfusion-related neuronal injury.     

A genetic algorithm-based clustering approach for database partitioning

In a typical distributed/parallel database system, a request mostly accesses a subset of the entire database. It is, therefore, natural to organize commonly accessed data together and to place them on nearby, preferably the same, machine(s)/site(s). For this reason, data partitioning and data allocation are performance critical issues in distributed database application design. We are dealing with data partitioning. Data partitioning requires the use of clustering. Although many clustering algorithms have been proposed, their performance has not been extensively studied. Moreover, the special problem structure in clustering is rarely exploited. We explore the use of a genetic search-based clustering algorithm for data partitioning to achieve high database retrieval performance. By formulating the underlying problem as a traveling salesman problem (TSP), we can take advantage of this particular structure. Three new operators for GAs are also proposed and experimental results indicate that they outperform other operators in solving the TSP. The proposed GA is applied to solve the data-partitioning problem. Our computational study shows that our GA performs well for this application.     

Mathematical expression recognition: a survey

Abstract. Automatic recognition of mathematical expressions is one of the key vehicles in the drive towards transcribing documents in scientific and engineering disciplines into electronic form. This problem typically consists of two major stages, namely, symbol recognition and structural analysis. In this survey paper, we will review most of the existing work with respect to each of the two major stages of the recognition process. In particular, we try to put emphasis on the similarities and differences between systems. Moreover, some important issues in mathematical expression recognition will be addressed in depth. All these together serve to provide a clear overall picture of how this research area has been developed to date. Received February 22, 2000 / Revised June 12, 2000     

The Processing and Improvement of Multi-Statement Queries in Chiql

Chiql is a novel Chinese relational database query language for Chinese users.It supports procedural query style in which users can specify a complex database request in multiple simple statements.This facility renders Chiql simple-to-use and easy-to-remember.However,direct execution of multi-statemen Chiql rueries(i.e.statement by statement)is often inefficient as potential index-based operations(e.g.join) are by-passed.Furthermore,it often incurs additional database operations,such as scan and projection.To improve this situation,the SMA(Statement Merging Algorithm)is proposed.The goal of SMA is to merge as many dependent statments within Chiql query as possible to form a more efficient Chiql query.The ability in achieving improved effciency without sacrificing the simplicity of the language is the major advantage of this algorithm.