Chung-Hee Lee

Fine-Grained named entity recognition using conditional random fields for question answering

In many QA systems, fine-grained named entities are extracted by coarse-grained named entity recognizer and fine-grained named entity dictionary. In this paper, we describe a fine-grained Named Entity Recognition using Conditional Random Fields (CRFs) for question answering. We used CRFs to detect boundary of named entities and Maximum Entropy (ME) to classify named entity classes. Using the proposed approach, we could achieve an 83.2% precision, a 74.5% recall, and a 78.6% Fl for 147 fined-grained named entity types. Moreover, we reduced the training time to 27% without loss of performance compared to a baseline model. In the question answering, The QA system with passage retrieval and AIU archived about 26% improvement over QA with passage retrieval. The result demonstrated that our approach is effective for QA.

A LF based answer indexing method for encyclopedia question-answering system

This paper proposes a fast and effective question-answer system for encyclopedia domain using a new answer indexing method. We define about 160 answer types. The indexer generates AIU(Answer Index Unit) structures between answer candidates and content words within LF(Logical Form) and sentence boundary. We select essential terms among question terms using syntactic information for ranking the answer candidates. Experiments show our new method is good for the encyclopedia question-answering system.

An Intelligent TV interface based on Statistical Dialogue Management

In this paper, we propose an intelligent TV interface using a voice-enable dialogue system. This paper rests on the both directions: a new type of dialogue management model and its use for practical systems to commercialize. We devise a practical dialogue management model based on statistical learning methods. To analyze discourse context, we utilize statistical learning techniques for anaphora resolution and discourse history management. Contrary to the rule-based system, we develop an incremental learning method to construct dialogue strategies from the training corpus. Several dialogue service models equipped with the proposed TV interface are explained. To evaluate our model and its impact on an application task, we apply the stand-alone model to an TV settop box called eDi-TV.

A Case study of Edutainment Robot: Applying Voice Question Answering to Intelligent Robot

This paper presents a case study of edutainment robot, which is an intelligent robot for educational use with a voice-QA model applied. The emphatic functions of our robot are: analyzing spoken question from a student, finding an appropriate answer in Korean encyclopedia, and then serving the answer with speech synthesis. We develop the ESTk, which is an Automatic Speech Recognition (ASR) system based on Finite State Network (FSN) for processing Korean spoken questions. For answer extraction, we utilize machine learning techniques and pattern extraction method. With our live-update interaction method, our robot can be extended with new knowledge in real-time. By conducting a quiz game, we show a possibility of our robot as an edutainment robot.

Enhanced question answering with combination of pre-acquired answers

Recently there is a need for QA system to answer various types of user questions. Among these questions, we focus on record questions and descriptive questions. For these questions, pre-acquired answers should be prepared, while traditional QA finds appropriate answers in real-time. In this paper, we propose enhanced QA model by combining various pre-acquired answers in encyclopedia. We defined pre-acquired answer types, 55 Record Type(RT)s and 10 Descriptive Answer Type(DAT)s, in advance. To construct answer units, we built 183 Record Answer Indexing Templates and 3,254 descriptive patterns. We discussed how our proposed model was applied to the record and descriptive questions with some experiments.

Syllable-based Korean POS Tagging Based on Combining a Pre-analyzed Dictionary with Machine Learning

This study is directed toward the design of a hybrid algorithm for syllable-based Korean POS tagging. Previous syllable-based works on Korean POS tagging have relied on a sequence labeling method and mostly used only a machine learning method. We present a new algorithm integrating a machine learning method and a pre-analyzed dictionary. We used a Sejong tagged corpus for training and evaluation. While the machine learning engine achieved eojeol precision of 0.964, the proposed hybrid engine achieved eojeol precision of 0.990. In a Quiz domain test, the machine learning engine and the proposed hybrid engine obtained 0.961 and 0.972, respectively. This result indicates our method to be effective for Korean POS tagging.

Relevant Keyword Collection using Click-log

The aim of this paper is to collect relevant keywords from clicklog data including user`s keywords and URLs accessed using them. Our main hyphothesis is that two or more different keywords may be relevant if users access same URLs using them. Also, they should have higher relationship when the more same URLs are accessed using them. To validate our idea, we collect relevant keywords from clicklog data which is offered by a portal site. As a result, our experiment shows 89.32% precision when we define answer set to only semantically same words, and 99.03% when we define answer set to broader sense. Our approach has merits that it is independent on language and collects relevant words from real world data.

Korean Semantic Role Labeling Using Semantic Frames and Synonym Clusters

Semantic information and features are very important for Semantic Role Labeling(SRL) though many SRL systems based on machine learning mainly adopt lexical and syntactic features. Previous SRL research based on semantic information is very few because using semantic information is very restricted. We proposed the SRL system which adopts semantic information, such as named entity, word sense disambiguation, filtering adjunct role based on sense, synonym cluster, frame extension based on synonym dictionary and joint rule of syntactic-semantic information, and modified verb-specific numbered roles, etc. According to our experimentations, the proposed present method outperforms those of lexical-syntactic based research works by about 3.77 (Korean Propbank) to 8.05 (Exobrain Corpus) F1-scores.

Competition Relation Extraction based on Combining Machine Learning and Filtering

This study was directed at the design of a hybrid algorithm for competition relation extraction. Previous works on relation extraction have relied on various lexical and deep parsing indicators and mostly utilize only the machine learning method. We present a new algorithm integrating machine learning with various filtering methods. Some simple but useful features for competition relation extraction are also introduced, and an optimum feature set is proposed. The goal of this paper was to increase the precision of competition relation extraction by combining supervised learning with various filtering methods. Filtering methods were employed for classifying compete relation occurrence, using distance restriction for the filtering of feature pairs, and classifying whether or not the candidate entity pair is spam. For evaluation, a test set consisting of 2,565 sentences was examined. The proposed method was compared with the rule-based method and general relation extraction method. As a result, the rule-based method achieved positive precision of 0.812 and accuracy of 0.568, while the general relation extraction method achieved 0.612 and 0.563, respectively. The proposed system obtained positive precision of 0.922 and accuracy of 0.713. These results demonstrate that the developed method is effective for competition relation extraction.

Merging and Re-ranking Answers from Distributed Multiple Web Sources

Depending on questions, various answering methods and answer sources can be used. In this paper, we build a distributed QA system to handle different types of questions and web sources. When a user question is entered, the broker distributes the question over multiple sub-QAs according to question types. The selected sub-QAs find local optimal candidate answers, and then they are collected in to the answer manager. The merged candidates are re-ranked by adjusting confidence weights based on the question analysis result. The re-ranking algorithm aims to find global optimal answers. We borrow the concept from the margin and slack variables in SVM, and modify to project confidence weights into the same boundary by training. Several experimental results prove reliability of our proposed QA model.