时间:2013年12月23日上午9:30
地点:校本部理工楼504
报告一题目:基于移进-归约的成分句法分析研究
报告人:朱慕华研究员
报告摘要:句法分析作为自然语言处理的核心任务受到广泛的关注。报告将介绍移进-归约成分句法分析研究的最新进展。首先描述基于全局训练(global training)和集束搜索(beam-search)的基准系统。然后将详细介绍移进-归约句法分析模型的改进方法,以及如何从自动标注数据中自动抽取信息以改进移进-归约句法分析的准确性。最终构建了当前领域内最好的中文移进-归约句法分析器,达到了快速而且准确地进行成分句法分析的目标。
报告人简介:朱慕华2013年毕业于东北大学,并获得博士学位。现为Nuance(上海)公司研究员,从事语音合成方面的研究与开发。朱慕华的研究领域包括句法分析和文本处理等自然语言处理任务的研究,并取得系列研究成果,发表SCI源期刊论文和自然语言处理领域顶级会论文10余篇。目前主要关注如何利用句法信息改进韵律结构的识别。
报告二题目:个性化新闻推荐
报告人:刘水博士
报告摘要:个性化推荐是美国雅虎新闻博彩平台的主要特色,报告将简略介绍雅虎的个性化推荐模型及其系统实现。在个性化模型中,热门新闻和用户的兴趣是排序模型的关键:如何选择出用户普遍感兴趣的热门新闻,如何计算出用户的兴趣,并将二者结合起来,是我们要解决的一个关键问题。在用户的兴趣模型中,我们采用了基于正态先验假设的极化稀疏模型,在热门新闻中我们采用了基于衰减加权的CTR模型,在GBDT的框架下,计算二类特征对于浏览行为的权重,有效的将两类特征融合起来。
报告人简介:刘水,2011年毕业于哈尔滨工业大学,并获得博士学位。毕业后先后服务于雅虎搜索和雅虎个性化新闻推荐部门,现供职于百度的社区技术部,主要负责百度文库和百度阅读的各种推荐模块。主要研究领域,主要是机器翻译,完全句法分析,以及文本推荐算法。毕业至今先后参与开发雅虎搜索的语言学分析模块,雅虎博彩平台的个性化推荐,以及百度阅读产品和文库产品的推荐开发,对工业界的推荐算法和推荐框架有一定的理解和经验。
报告三题目:N-gram Modeling for Translation
报告人:Zhang Hui
报告摘要:Standard phrase-based translation models do not explicitly model context dependence between translation units. As a result, they rely on large phrase pairs and target language models to recover contextual effects in translation. We explore n-gram markov models over Minimal TranslationUnits (MTUs) to explicitly capture contextual dependencies across phrase boundaries in the channel model. As there is no single best direction in which contextual information should flow, we explore multiple decomposition structures as well as dynamic bidirectional decomposition. The resulting models are evaluated in an intrinsic task of lexical selection for MT as well as a full MT system, through n-best re-ranking. These experiments demonstrate that additional contextual modeling significantly benefit a phrase-based system(up to 2.8 BLEU score) and that the direction of conditioning is important.Integrating multiple conditioning orders provides consistent benefit, and the most important directions differ by language pair.
个人简介: Hui Zhang is a 4th year PhD student in Information Sciences Institute, University of Southern Californian. He received his Bachelor and Master degrees from Xiamen University in 2004 and 2007
respectively. Before joining USC/ISI, he was a research associate in Institute for Infocomm Research (A-STAR, Singapore). His research interests are non-isomorphic bilingual structure translation, syntax
parsing and language modeling.