Cheol-Young Ock

Learning to classify short text from scientific documents using topic models with various types of knowledge

An efficient framework to classify short text from scientific documents is proposed.Topic models from various types of knowledge were used for enhancing features in documents.Two methods were presented to optimize external features that enhance relatedness in documents.The performances were evaluated by using real-world scientific documents from online publisher.Proposed methods are shown to outperform related work. Classification of short text is challenging due to data sparseness, which is a typical characteristic of short text. In this paper, we propose methods for enhancing features using topic models, which make short text seem less sparse and more topic-oriented for classification. We exploited topic model analysis based on Latent Dirichlet Allocation for enriched datasets, and then we presented new methods for enhancing features by combining external texts from topic models that make documents more effective for classification. In experiments, we utilized the title contents of scientific articles as short text documents, and then enriched these documents using topic models from various types of universal datasets for classification in order to show that our approach performs efficiently.

Word sense disambiguation as a traveling salesman problem

Word sense disambiguation (WSD) is a difficult problem in Computational Linguistics, mostly because of the use of a fixed sense inventory and the deep level of granularity. This paper formulates WSD as a variant of the traveling salesman problem (TSP) to maximize the overall semantic relatedness of the context to be disambiguated. Ant colony optimization, a robust nature-inspired algorithm, was used in a reinforcement learning manner to solve the formulated TSP. We propose a novel measure based on the Lesk algorithm and Vector Space Model to calculate semantic relatedness. Our approach to WSD is comparable to state-of-the-art knowledge-based and unsupervised methods for benchmark datasets. In addition, we show that the combination of knowledge-based methods is superior to the most frequent sense heuristic and significantly reduces the difference between knowledge-based and supervised methods. The proposed approach could be customized for other lexical disambiguation tasks, such as Lexical Substitution or Word Domain Disambiguation.

Using wiktionary to improve lexical disambiguation in multiple languages

This paper proposes using linguistic knowledge from Wiktionary to improve lexical disambiguation in multiple languages, focusing on part-of-speech tagging in selected languages with various characteristics including English, Vietnamese, and Korean. Dictionaries and subsumption networks are first automatically extracted from Wiktionary. These linguistic resources are then used to enrich the feature set of training examples. A first-order discriminative model is learned on training data using Hidden Markov-Support Vector Machines. The proposed method is competitive with related contemporary works in the three languages. In English, our tagger achieves 96.37% token accuracy on the Brown corpus, with an error reduction of 2.74% over the baseline.

Margin perceptron for word sense disambiguation

Word Sense Disambiguation (WSD) is an AI-complete problem where senses of words in the documents must be correctly selected from a senses inventory. Support Vector Machines (SVM) method has been successfully applied to supervised WSD. In contrast, perceptron has not been popular in supervised WSD. In this paper, a supervised method combining Margin Perceptron (MP) and Platts probabilistic output is proposed to solve the word sense ambiguity problem. Experiments were conducted on Senseval-3 English Lexical Sample Task data set. The performance is comparable with systems using SVMs. Our system is in line with the best system participating in Senseval-3, regarding that we only used given training data, and no classifiers combination technique was applied. The advantage of our method is mainly two-fold: Firstly, good achieved performance shows that MP can be applied to problem with limited training data, especially in natural language processing. Secondly, MP algorithm used in this work is easy to implement, which benefits the application and the extension of the algorithm.

RP-MAC: A cross-layer duty cycle MAC protocol with a Reduced Pipelined-forwarding feature for Wireless Sensor Networks

Recently, duty-cycle MAC protocols with pipelined-forwarding and routing-integrated features such as P-MAC and PRI-MAC were proposed for efficient communication in terms of power consumption and end-to-end delay in Wireless Sensor Networks (WSNs). However, they remain unnecessary idle listening time at each cycle and their handshake mechanism introduces unnecessary delay at each hop. In this paper, we propose a new cross-layer duty-cycle MAC protocol, called RP-MAC (Reduced Pipelined-forwarding) protocol, which improves performance by reducing idle listening time and control overhead. In RP-MAC, the RTS/CTS handshake procedure in the pipeline is shortened by taking advantage of ACK mechanism. A relaying node overhears the ACK message from its potential senders instead of listening to RTS. Right after receiving the data successfully, the relaying node just needs to listen to the CTS without sending the RTS. The simulation results using Qualnet show that RP-MAC achieves significant improvement in comparison with PRI-MAC in terms of energy consumption and end-to-end delivery latency.

Automatic Knowledge Extraction for Aspect-based Sentiment Analysis of Customer Reviews

It is challenging to figure out the most common appraisal of an online product since there are too many reviews about it uploaded on the internet. Several research methods using opinion mining in the context of the online reviews have been suggested to solve this issue. The existing research on opinion mining can be classified into three general levels: document-level, sentence-level, and aspect-level sentiment analysis. Aspect-based evaluation is the most meaningful application in opinion mining, and researchers are getting more interested in product aspect extraction; however, more complex algorithms are needed to address this issue precisely with larger corpora. This paper introduces a method to automatically gain a knowledge-based system, which then is used to capture product aspects and corresponding opinions from a large number of product reviews in a specific domain. Our efforts tend to improve accuracy and the usefulness of review summaries by leveraging knowledge of product aspect extraction and provide both appropriate level of detail and richer representation capabilities.

Neural Machine Translation Enhancements through Lexical Semantic Network

In most languages, many words have multiple senses, thus machine translation systems have to choose between several candidates representing different senses of an input word. Although neural machine translation has recently become a dominant paradigm and achieved great progress, it still has to confront with the challenge of word sense disambiguation. Neural machine translation models are trained to identify the correct sense of a word as part of an end-to-end translation task, and their performances on word sense disambiguation are not satisfactory. This paper presents a case study of machine translation for Korean language. We have manually built a Korean lexical semantic network - UWordMap - as a large-scale lexical semantic knowledge-based in which each sense of every polysemous word is associated with a sense-code constituting a network node. Then, based on UWordMap, we determine the correct sense and tag the appropriated sense-code for polysemous words of the training corpus before training neural machine translation models. Experiments on translation from Korean to English and Vietnamese show that UWordMap can significantly improve quality of Korean neural machine translation systems in terms of BLEU and TER cores.