NSF East Asia and Pacific Summer Institute (EAPSI) for FY 2013 in Japan

这项行动资助哥伦比亚大学的Yves Petinot于2013年夏天在日本东京的统计数学研究所开展计算机与信息科学与工程领域的研究项目。这个项目的题目是“用于自然语言生成和Web摘要的分层语言模型”。主持会议的科学家是Mochihashi教授。这个研究项目旨在开发和评估自然语言的层次模型，这些模型可以支持统计生成人类可读的句子，特别是网络专家。Web定位器是静态的，通常是Web统一资源定位器（url）的单句长摘要。该项目探索了统计和机器学习方法来自动预测语言元素如何出现在Web列表中。有机可用的Web资源，包括基于层次结构的Web目录，被用作学习和实验验证的基础。EAPSI奖学金的广泛影响包括为研究员提供美国以外的第一手研究经验；介绍该地区的科学、科学政策和科学基础设施；以及对社会、文化和语言的了解。这些活动符合美国国家科学基金会的目标，即在其科学家、工程师和教育工作者的职业生涯早期培养国际合作，从而确保美国科学劳动力具有全球意识。此外，该研究项目旨在促进大规模、高效和无偏见地发现信息和服务，这些信息和服务可以通过人类和机器代理通过Web访问。

This action funds Yves Petinot of Columbia University to conduct a research project in the Computer and Information Science and Engineering area during the summer of 2013 at the Institute of Statistical Mathematics in Tokyo, Japan. The project title is "Hierarchical Language Models for Natural Language Generation and Web Summarization." The host scientist is Professor Daichi Mochihashi.This research project aims at developing and evaluating hierarchical models of natural language that can support the statistical generation of human-readable sentences, and in particular, of Web gists. Web gists are static, typically single-sentence long summaries of Web Uniform Resource Locators (URLs). The project explores statistical and machine learning approaches to automatically predict how elements of language appear in Web gists. Organically available Web resources, including hierarchy-based Web directories, are being used as a basis for learning and experimental validation.Broader impacts of an EAPSI fellowship include providing the Fellow a first-hand research experience outside the U.S.; an introduction to the science, science policy, and scientific infrastructure of the respective location; and an orientation to the society, culture and language. These activities meet the NSF goal to educate for international collaborations early in the career of its scientists, engineers, and educators, thus ensuring a globally aware U.S. scientific workforce. Furthermore, this research project aims at facilitating the large-scale, efficient and non-biased discovery of information and services accessible through the Web both by human and machine agents.