蜜桃君

天涯大盗

两件事

以前总有太多的目标。惭愧,我就是爱做白日梦的巨蟹座。昨天下午我想到此生要做两件事:一是用我半生的时间去写三五本英文小说。二是尽平生所学去研究人工智能。立字为据,始终不渝。我本来下决心在这两个都很庞大的事业中再筛选一个,可是决心半天这两个实在是无法再割舍了。上天啊,请原谅我的贪婪。我不是要娶两个老婆,我只是想拥有两个梦想。

最近我的成果第一是完成硕士论文,并通过答辩。导师看了我的初稿之后就给写了一封信

thesis.jpg

The thesis studies the figurative language processing by extracting Google Books Ngrams and building a structured knowledge base consisting of noun-adjective pairs, which are processed and listed according to their frequency. Based on the statistical analysis of the knowledgebase, this paper has identified the rules of the source-target conceptual mapping mechanism as represented in formulas, and further conducted the computation of the semantic similarity.

论文通过提取Google Books Ngrams,建立由名词-形容词对组成的结构化知识库,对其进行处理,并按频率列出,对形象化语言处理进行研究。在对知识库进行统计分析的基础上,确定源概念-目标概念映射机制的规则公示,并进行语义相似度的计算。

第二个成果是“中英文认知属性知识检索系统”上线。

cogbank1.png

例如在搜索框输入单词pig,可以看到pig的英文属性有greedy, fat, midget, gluttonous, lazy等,并有相应的使用频率,可以看出pig在英文文献中更多的被用来表现什么特征。

cogbank2.png

研究这个有何价值呢?

其实就想让计算机理解你说的话究竟是反讽,还是字面意思。例如你说一个人像猪一样笨,很容易理解。但说一个人聪明得像猪一样,这对计算机的理解和翻译来说是有难度的。那么就需要建构一个比喻知识库,让计算机了解人对猪的认知是什么。同时,我们也可以通过中英文词库的对比,看出中文里猪的显著特征与英文的差别,这对跨文化研究也是很有价值的。

针对另一个梦想,关于英文小说,我准备把我一段特殊的经历写下来,作为故事的其中一个背景。

In this story, I recalled my experience of working in a military institute, where I was reminded of C.P. Snow's concept of Two Cultures, the split between science and humanities. The engineers surrounding me had no time for philosophy, and my knowledge of their research was zero.

On June 3rd, 2006, after lunch, a group of 35 young engineers and 5 PLA crewmembers boarded a newly-converted early-warning aircraft, which originally had been a Y-8 military cargo plane. The plane took off from Wuxi Shuofang Base, and crashed into a hillside near the Yao village, Baidian Township, in Guangde County of the central Anhui province, killing all the crewmembers and engineers on board. I was involved in handling the aftermath of the worst air force accident in PRC history, and eventually found out how I was connected to the people on the plane, and how the Two Cultures can be bridged. This story, spanning the period of time from 2000 to 2006, is narrated in two parallel lines, with Chapters told in the second person (You) and Distracters in the first (I), which hints that I was only playing a supporting role in my story. It is delicate to find out the Chapters and Distracters eventually merge into one bigger story until you reach the end.

在这个故事中,我回忆了在军工研究所工作的经历,在那儿我想到C.P.斯诺的 "两种文化"(Two Cultures)概念,即科学与人文的分裂。我周围的工程师们无暇顾及哲学,而我对他们的研究知识为零。

2006年6月3日,午饭后,35名年轻的工程师和5名解放军机组人员登上了一架新改装的预警机,这架飞机原来是一架Y-8军用货机。飞机从无锡硕放基地起飞后,在安徽省中部广德县柏垫镇姚村附近的山坡上坠毁,机上机组人员和工程技术人员全部遇难。我参与处理了中国历史上最严重的空军事故的善后工作,最终发现我和飞机上的人是如何联系在一起的,"两种文化 "是如何衔接的。这个故事的时间跨度从2000年到2006年,以两条平行线叙述,Chapters以第二人称(你)讲述,Distracters以第一人称(我)讲述,这暗示了我在我的故事中只是一个配角。很微妙的是,当你读到结尾时,会发现Chapters和Distracters最终会融合成一个更大的故事。

Powered By mitaojun.com

版权所有,转载请注明来源及文章链接。苏ICP备2020049409号