Month reports

It was time for me to write my month reports. Although there were two days left for us to submit, I'd like to write it today. As this was the end day of this year. Why not finish this year's work in this year? If I suspend to tomorrow or the day after tomorrow, I would keep bad habit.

I had done the month reports easy. I made it based on my blog and my regular working documents. This time I introduced the papers management experience first. And then wrote the papers reading outline of seven papers. I made a detail plan of next month later.

Right now, I knew the advantage of writing blog. Why not keep on?

Dicsourse-new detectors for definite description resolution: a survey and a preliminary proposal

Title：Dicsourse-new detectors for definite description resolution: a survey and a preliminary proposal针对确定性描述的新描述识别：综述和初步建议
Author：Massimo Poesio; Olga Uryupina; Renata Vieira
Author organization:
Massimo Poesio, University of Essex,Computer Science and Cognitive Science(UK)
Olga Uryupina, Universitat des Saarlandes, Computerlinguistik (Germany)
Renata Vieira, Unisinos, Computacao Aplicada (Brazil)

Conference: Proceedings of the Workshop on Reference Resolution and its Applications. ACL2004
Summary:
English:
Vieira and Poesio (2000) proposed an algorithm for definite description (DD) resolution that incorporates a number of heuristics for detecting discoursenew descriptions. The inclusion of such detectors was motivated by the observation that more than 50% of definite descriptions (DDs) in an average corpus are discourse new (Poesio and Vieira, 1998), but whereas the inclusion of detectors for non-anaphoric pronouns in algorithms such as Lappin and Leass’ (1994) leads to clear improvements in precision, the improvements in anaphoric DD resolution (as opposed to classification) brought about by the detectors were rather small. In fact, Ng and Cardie (2002a) challenged the motivation for the inclusion of such detectors, reporting no improvements, or even worse performance. We re-examine the literature on the topic in detail, and propose a revised algorithm, taking advantage of the improved discourse-new detection techniques developed by Uryupina (2003).

中文：
Vieira and Poesio (2000)提出了一种确定描述（definite description, DD）消解算法。算法中采用了一系列的启发式规则来检测上文中未出现过的话语描述。这种研究的动机在于观察一些平衡语料发现超过50％的话语描述都是当前上文中未出现过的话语描述。但是非指代性代词的消解算法如Lappin and Leass(1994)在精确率上有所提高，而在指代性代词的消解算法由于不是分类问题而导致提高非常少。事实上，Ng and Cardie (2002a)在这类消解问题上进行过开创性研究，结果是没有提高甚至性能有所下降。我们重新详细调研了这个研究点，并在Uryupina (2003)提高未登录描述消解技巧的基础上提出了一种修改性的算法。

为什么要做这个题目:

Poesio and Vieira(1998)在语料上进行研究发现众多语料，诸如Penn Treebank等，有52％的确定性描述（definite description, DD）是当前上文中未出现过的。

别人怎么做的:

Vieira and Poesio (2000)提出了一种确定描述（definite description, DD）消解算法。算法中采用了一系列的启发式规则来检测当前上文中未出现过的话语描述。但是包含非指代性代词识别的消解算法如Lappin and Leass(1994)在精确率上有所提高，而在包含指代性代词识别的消解算法由于不是分类问题而导致提高非常少。事实上，Ng and Cardie (2002a)在这类消解问题上进行过开创性研究，结果是没有提高甚至性能有所下降。
众多相关研究（Poesio and Vieira,1998;Bean and Riloff,1999; Ng and Cardie, 2002a; Uryupina,2003）都一致认为在DDs的DN识别中许多因素都起着重要作用。绝大多数算法采用混合算法来识别肯定性的DDs,识别DN的专有名称，识别功能性的DDs,识别被修饰的DDs来确定关系。
这些研究一致认为DN识别不能和指代消解分开实施。

问题在哪里：

这些机器学习方法的一个问题是这些系统在菜地和DD消解上都没有达到很好的效果，对比一些特殊的算法如下：Ng and Cardie的最好的一个程序版本在各种指代表达式上的F＝65.8，但是在DD上的F=29.6(Vieira and Poesio的最好的结果为F=77)，代词上的F=28.2(Tetreault, 2001的代词消解算法评测达到F=80)。很明显，这些算法的对比只能在同一个数据集上。正如Mitkov2000讨论的那样，在指代消解的评测中预处理和后处理对指代消解算法的效果具有很大的影响。但是我们认为在效果很好的系统上进行DN识别的评测可以更好的达到我们预期的结果。

作者提出了怎样的新方法：

本文的工作首先是对比了DN识别前后各种算法的代消解效果。采用Vieira and Poesio的算法来检测不包含DN识别的指代消解算法效果，在采用Uryupina的特征集的基础上加上一个简单的统计消解模型来检测加上DN识别的消解效果，数据集都采用Cardie and Ng提到的MUC-7中的数据。
实验结果如下：
----------------------------------
--------| R | P | F |
----------------------------------
Pronouns| 65.5 | 63.0 | 64.2 |
----------------------------------
DDs | 56.7 | 56.1 | 56.4 |
----------------------------------
Table 7.Evaluation of the GUITAR system without DN detector off raw text

----------------------------------------------
--------------------| R | P | F |
----------------------------------------------
Without DN detection| 44.7 | 54.9 | 49.3 |
----------------------------------------------
With DN detection | 41.7 | 80.0 | 54.6 |
----------------------------------------------
Table 8.Using an oracle

这个实验说明了DN识别可以提高精确率30％左右。但是还不能说明它在高性能的指代消解系统上的提高。

本文工作中又提出了一个新的DN识别的特征集，评测数据也是在MUC-7上进行。但是实验没有进行完，现在还没有实验数据和分析。

对个人的研究的指导意义

可以借助这种方法来更好的完成ACE的指代消解的算法和评测。

存在问题和个人想到的改进方案

暂时无

Happy New Year's Day!

There were only two days left to New Year's Day.

From 15:00, all the members of our laboratory celebrated in the Qianshoufo Hotel. There were so many entertainment items for us, such as swim, skate, billiards, table tennis, bowls, shuffleboard. We all enjoied ourself.

After the entertainment time, we all began to have the nice dinner. There were twenty-four persons in the dining room. We were divided into three tables. Our table members included Dr.Tliu, Hong Yu, Yu Haibin, Hu Xiaoguang, Zhao Yongzhen, me, Huang Yongguang, Gao Liqi, Chen Yihen. During the dinner time, each of our table members had sung a song at least. We all named ourself as All Star Table.

Yesterday, my neighbor Mr. Luo Junping, who went to harbin for some business, brought me some sausage. The sausage was prepared by my parents. That was my favourite. I ask the kitchen to cook it and then divided into three dishes. All the members of our table were like it. Me, too. I had a lot of it. So nice dish.

Abid was invited to sing a song. Although we all could not understand what he sang, we all thanked him.

After the dinner, we had a photo of the whole laboratory. We all happied today!

Thanks to Dr.Tliu, thanks to our laboratory.

Happy new year to every one!

Using word similarity lists for resolving indirect anaphora

论文题目:Using word similarity lists for resolving indirect anaphora采用词语相似度列表来消解间接指代

论文出处：ACL2004 workshop on coreference resolution

发表时间:2004, July 25-26

论文作者: Caroline Gasperin and Renata Viera

作者单位: PIPCA-Unisinos Sao Leopoldo, Brazil
Summary:
English:
In this work we test the use of word similarity lists for anaphora resolution in Portuguese corpora. We applied an automatic lexical acquisition technique over parsed texts to identify semantically similar words. After that, we made use of this lexical knowledge to resolve coreferent definite descriptions where the head-noun of the anaphora is different from the head-noun of its antecent, which we call indirect anaphora.

中文：
本文中我们采用词语相似度列表在葡萄牙语语料上进行指代消解。我们采用了一种自动获取词汇的技术在句法分析后的文本上识别语义相似的词语。然后利用词汇知识来消解确定修饰的间接指代。这里的间接指代是指先行词和指代词的核型名词不同的情况。


为什么要做这个题目

简介指代消解主要解决先行词和指代词的核心词不一样，但是语义比较相关的短语也需要进行消解。

别人怎么做的

以前在英文上有人用过词汇相似度列表来进行指代消解（Poesio etal., 2002; Schulte im Walde, 1997; Bunescu, 2003）。

本文主要是在葡萄牙语上的指代消解，而且论文中介绍的词语相似度列表的方法需要查阅许多其他的论文。不符合现在的调研方向。先阅读到此。以后有机会再阅读。

Learn Prolog, Now!

When I was finding some information on machine learning, I clicked into a web which introduced the natural language processing. Iin that page there were some links to the prolog syntax analysis programs.

I had seen many useful prolog programs that could solve the natural language processing problem. So, I'd like to know something about this programming language.

Prolog, shorted by Program in Logic, can induce lots of results from the knowledge base.

After finding some useful information, I finded a nice book named as Learn Prolog Now! The useful software of prolog programming is SWI-prolog.

创造性的研究

晚上实验室Reading group讨论会上闲暇时和师姐讨论中国博士和外国博士的创造性研究的问题。师姐认为现在咱们的创造性的研究太少了，美国那边的许多博士的论文本生就奠定了将来的成就。我补充道，据报道中国现在的博士数量已经超过了美国，但是科研实力上还是远远落后于美国。为什么呢？我想大概是因为要求不一样吧。

现在有人说博士非常难念，但是看看现在国内的博士们的研究内容，出了个别情况外，很多都是水平较为低下的。

为什么会出现这种情况呢？

我猜，呵呵，也只能是猜了，是因为科研的过程中存在一些问题：首先的调研工作进行的不是很彻底，以致于一些工作和别人的重复了也不知道；其次是吃苦精神不够，读博有人说是一种人生的历练，需要非常的吃苦才行；开放性不够，博士的研究应该是非常开放的，但是许多的研究，随着程度的深入慢慢的变得越来越窄，研究越多就越不愿意去吸纳其他学科的研究成果和思想。

面对自己的研究，我能做些什么呢？现在我的状态是调研共指消解的研究现状，调研的工作还不太够。首先解决调研问题吧。写出一个像样的综述才能开始下一步的工作。

Applying Coreference to Improve Name Recognition

论文题目: Applying Coreference to Improve Name Recognition利用共指来提高名称识别

论文出处: ACL2004 workshop on Coreference resolution

发表时间:2004年7月25日

论文作者：Heng JI and Ralph GRISHMAN

作者单位:纽约大学计算机系

摘要：
English:
We present a novel method of applying the results of coreference resolution to improve Name Recognition for Chinese. We consider first some methods for gauging the confidence of individual tags assigned by a statistical name tagger. For names with low confidence, we show how these names can be filtered using coreference features to improve accuracy. In addition, we present rules which use coreference information to correct some name tagging errors. Finally, we show how these gains can be magnified by clustering documents and using cross-document coreference in these cluters. These combined methods yield an absolute improvement of about 3.1% in tagger F score.

中文：
我们提出了一种新颖的利用共指消解结果来提高中文名称识别的方法。首先，我们采用一个统计名称识别器来进行标注，然后采用一些方法来计算每个单独标记的可信度。对于可信度较低的名称，我们展示了如何利用共指特征来过滤并提高名称识别的准确率。然后，我们提出了一些规则来利用共指信息校正名称识别的错误。最后，我们利用文本聚类和跨文档共指消解来考察名称识别效果的提高。这种结合在标注F值上产生了3.1％的提高。




为什么要做这个题目：为了提高名称识别的准确率

别人怎么做的：

在做NE识别方面，有人利用HMM（Bikel et al, 1997），最大熵（Borthwick et al, 1998, Chieu and Ng 2002），决策树（Sekine at al, 1998）,条件随机域（McCallum and Li, 2003）,基于类别的语言模型（Sun et al. 2002）,基于Agent的方法（Ye et al, 2002）和支持向量机。

问题在哪里

这些机器学习方法应用在名称识别上时效果的好坏都会依赖于标注语料库的大小和使用特征的范围。更为特殊的是，多数方法都是采用规模很小的上下文信息，例如当前词的前一个或两个词语以及跟随的名称。如果测试时遇到一个没有见过的词语，并且出现在一个信息量很少的上下文，那么就会很难进行识别。


作者提出了怎样的新方法

作者利用共指消解的结果来挖掘全局信息从而增加识别的准确率。在一篇文档中进行共指消解可以很好的用于名称识别，在多篇文档中的共指消解可以更好的用于名称识别。本文中采用的共指消解算法是基于一些启发式规则的方法，规则如下
－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－
Rule Type Rule Description

Name & Name
----All
--------Ident(i, j) Mentioni and Mentionj are identical
--------Abbrev(i, j) Mentioni is an abbreviation of Mentionj
--------Modifier(i, j) Mentionj = Modifier + “de” + Mentioni
--------Formal(i, j) Formal and informal ways of referring to the same entity(Ex. “美国国防部 / American Defense Dept. & 五角大楼/ Pentagon”)
----PER
--------Substring(i, j) Mentioni is a substring of Mentionj
--------Title(i, j) Mentionj = Mentioni + title word; or Mentionj = LastName + title word
----ORG
--------Head(i, j) Mentioni and Mentionj have the same head
----GPE
--------Head(i, j) Mentioni and Mentionj have the same head
--------Capital(i, j) Mentioni: country name;Mentionj: name of the capital of this country Applied in restricted context.
--------Country(i, j) Mentioni and Mentionj are different names referring to the same country.(Ex. “中国 / China & 华夏 / Huaxia & 共和国 / Republic”)

Name & Nominal
----All
--------RSub(i, j) Namei is a right substring of Nominalj
--------Apposition(i, j) Nominalj is the apposite of Namei
--------Modifier2(i, j) Nominalj = Determiner/Modifier + Namei/ head
----GPE
--------Ref(i, j) Nominalj = Namei + GPE Ref Word (examples of GPE Ref Word: “方面 / Side”, “政府/Government”, “共和国 / Republic”, “自治政府/ Municipality”)

Nominal & Nominal
----All
--------IdentN(i, j) Nominali and Nominalj are identical
--------Modifier3(i, j) Nominalj = Determiner/Modifier + Nominali
－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－－

这些规则不包含代词的消解问题。
采用MUC的评测机制，在人工标注生成的mentions语料上评测的结果是R=82.7%,p=95.1%,F=88.47%
在机器自动生成mentions上进行的评测结果是R=74。3%,p=84.5%,F=79.07%


这种方法从理论上分析有何长处

对于名称识别：增加了特征的来源。并且很好的利用了全局信息。
对于共指消解：简单、快速。

还存在哪些问题

正如规则中的说明一样，很多规则是很难用机器自动进行的，比如Formal(i,j)。这些模块的实现还需要一些技术。

个人想到的改进方案或者个人的创新观点

规则加上一些统计的方法来完成任务。