Mai Miyabe

Use trend analysis of twitter after the great east japan earthquake

After the Great East Japan Earthquake in Japan 2011, numerous tweets were exchanged on Twitter. Several studies have already pointed out that micro-blogging systems have shown potential advantages in emergency situations, but it remains unclear how people use them. This paper presents a case study of how people used Twitter after the Great East Japan Earthquake. First, we gathered tweets immediately after the earthquake and analyzed various factors, including locations. The results revealed two findings: (1) people in the disaster area tend to directly communicate with each other (reply-based tweet). On the other hand,(2) people in the other area prefer spread the information from the disaster area by using Re-tweet.

Effects of undertaking translation repair using back translation

Translation repair plays an important role in intercultural communication that involves machine translation. It can be used to create messages that have very few translation mistakes. The accuracy of the results of translation repair when an original sentence is rewritten has not yet been evaluated. The improvement brought about by translation repair has to be demonstrated in order to apply translation repair to multilingual communication. Therefore, we have evaluated the translation repair of Japanese-English, Japanese-Chinese, and Japanese-Korean translations using back translation. We have used test sentences with a character count ranging from 15 to 32. On the basis of these evaluation experiments, we have estimated the accuracy and the cost of translation repair. (1) After nearly six rounds of translation repair work in three languages, the average translation accuracy of the sentences used in the experiment was improved (the meaning of the translated sentences was almost the same as that of the original sentences). In the experiment, 65% of the sentences were improved to the level of highly accurate. Moreover, 99% of the sentences were improved to the level of moderately accurate. 2) The cost of repairing a sentence depended on the number of translation-difficult words or phrases that were contained in the sentence. When the quality of a translation was low, finding the word or phrase to be modified was a difficult task. Thus, the cost corresponded to the quality of the translation.

Parallel-text based support system for intercultural communication at medical receptions

Differences in languages have to be bridged in order for intercultural communication to take place. Although the medical field requires highly accurate translations for promoting intercultural communication, the present level of support is insufficient. In this paper, we propose a system that uses parallel texts to support intercultural communication at hospital reception desks. Using parallel texts that have been translated accurately, the proposed system can provide this level of intercultural medical communication. We conducted an experiment to test the effectiveness of the system, which confirmed that it is possible to provide accurate intercultural communication using parallel texts. The experiment compared our system with the conventional system, which uses a printed list of parallel texts, and showed that the retrieval time of the proposed system is about the same as that of the conventional system.

Effects of Repair Support Agent for Accurate Multilingual Communication

Translation repair plays an important role in intercultural communication using machine translation. It can be used to create messages that have very few translation mistakes. However, translation repair is a laborious task. It is important to carry out translation repair efficiently. Therefore, we propose a repair support agent that provides the segments that have not been translated accurately. We perform experiments on the translation repair efficiency to evaluate the effectiveness of the repair support agent. The results of these experiments are as follows. (1) Providing inaccurately translated segments improves the ability to detect inaccurate segments. (2) The inaccurate-judgment rate can affect the improvement of the efficiency of translation repair.

Accuracy evaluation of sentences translated to intermediate language in back translation

The back-translation method is used to check the accuracy of a sentence translated to a native language. We believe that there exits a positive correlation between the accuracy of sentences translated to an intermediate language and that of back-translated sentences. However, this has not yet been verified. However, some back-translated sentences have high accuracy even if the translated sentence is inaccurate. Therefore, we have to verify the correlation between the accuracy of sentences translated to an intermediate language and that of back-translated sentences. We have evaluated the accuracy of back-translated sentences and that of sentences translated to an intermediate language to establish the correlation between the two accuracies. We have obtained the following results: (1) There exists a positive correlation between the accuracy of sentences translated to an intermediate language and that of back-translated sentences. (2) The occurrence rate of an accuracy mismatch case, wherein a back-translated sentence is accurate but the translated sentence is inaccurate, is less than or equal to 0.5%. (3) Back-translation can be used to check the accuracy of a translated sentence.

A web-based multilingual parallel corpus collection system for the medical field

We have developed a Web-based multilingual parallel corpus collection system named TackPad for medical workers and foreign patients. The number of foreign visitors and residents in Japan is increasing. However, it is not sufficient to provide linguistic support via medical translators to people who do not understand Japanese. In particular, poor communication may lead to medical errors. Moreover, there are limitations to a medical translators support. It is difficult to use machine translations in the medical field because of their inaccuracy. A system that uses parallel corpora with guaranteed accuracy appears to be useful. However, a huge amount of parallel corpora related to specialty fields and in several languages is required in the medical field. It would be difficult for a limited number of people to accumulate a sufficient amount of parallel corpora. Therefore, we have specially designed a support system named TackPad for the medical field; this system can accurately register parallel corpora and supply this data to other systems. TackPad adopts a Web 2.0 approach and can be used to tap peoples capabilities for knowledge accumulation. The purpose of TackPad is to collect parallel corpora; other systems can use the corpus via the Internet. We have obtained various responses from the result of the trial experiment. The experiment has demonstrated the practicality of the system, for example, people of various age groups and from various occupations can actually use the system.

Vocabulary size in speech May Be an early indicator of cognitive impairment

Little is known about the relationship between mild cognitive impairment (MCI) and changes to language abilities. Here, we used the revised Hasegawa Dementia Scale (HDS-R) to identify suspected MCI in elderly individuals. We then analyzed written and spoken narratives to compare the language abilities between study participants with and without MCI in order to explore the relationship between cognitive and language abilities, and to identify a possible indicator for the early detection of MCI and dementia. We recruited 22 people aged 74 to 86 years (mean: 78.32 years; standard deviation: 3.36). The participants were requested to write and talk about one of the happiest events in their lives. Based on HDS-R scores, we divided the participants into 2 groups: the MCI Group comprised 8 participants with a score of 26 or lower, while the Healthy Group comprised 14 participants with a score of 27 or higher. The transcriptions of both written and spoken samples for each participant were used in the measurement of NLP-based language ability scores. Our analysis showed no significant differences in writing abilities between the 2 groups in any of the language ability scores. However, analysis of the spoken narrative showed that the MCI Group had a significantly larger vocabulary size. In addition, analysis of a metric that signified the gap in content between the spoken and written narratives also revealed a larger vocabulary size in the MCI Group. Individuals with early-stage MCI may be engaging in behavior to conceal their deteriorating cognition, thereby leading to a temporary increase in their active spoken vocabulary. These results indicate the possible detection of early stages of reduced cognition before dementia onset through the analysis of spoken narratives.

Analysis of Microblog Rumors and Correction Texts for Disaster Situations

Microblogging systems such as Twitter have become popular. They are especially useful and helpful for users in disaster situations. Microblogs have facilitated the spread of information of all kinds, even rumors. Rumors block adequate information sharing and cause severe problems. Several studies have analyzed rumors, but it remains unclear how rumors are spread on microblogging systems. As described in this paper, we present a case study of how rumors are spread on Twitter in a recent disaster situation, that of the Great East Japan earthquake in March 11 2011, based on comparison to a normal situation. We also specifically examine the correction of rumors because automatic extraction of rumors is difficult, but extracting rumor-correction is easier than extracting the rumors themselves. We (1) classify tweets in disaster situations, (2) analyze tweets in disaster situations based on users impression, and (3) compare the spread of rumor tweets in a disaster situation to that in a normal situation. A category with only the three required fields

Influence of detecting inaccurate messages in real-time remote text-based communication via machine translation

In multilingual communication using machine translation, translation correction based on back translation plays an important role. If users are unable to identify inaccurate translations, they will not attempt to correct them. This is an important issue because a considerable proportion of inaccurate machine translations go undetected, which prevents smooth communication. Therefore, it is necessary to develop a method for preventing users from transmitting inaccurate messages. This method can ensure that only accurate messages are exchanged between users. However, some problems may occur with the use of this method, because it rejects the users judgment. It is important to verify the effectiveness of this method. We propose a method that reduces the transmission of inaccurate messages. If the method identifies a message as inaccurate, the chat server does not send the message to the receiver; it will instead encourage the sender to correct the message. We verified the effectiveness of the proposed method in chat communication using the Wizard of Oz method. The following conclusions were drawn from experimental results: (1) The chat communication progressed even when a user occasionally sent an inaccurate message. However, inaccurate messages did disrupt communication to some extent. When the proposed method was applied, users were able to communicate with each other more accurately. We observed that there were no significant differences in the effectiveness of communication with or without the proposed method. This suggests that our method is best applied to communication situations that require high degrees of accuracy. (2) The use of the proposed method caused users some discomfort because it rejected their judgment; thus, they did not know how to repair the messages. Based on this, it would seem that it is not easy to apply the proposed method to chat communication.

Design of Face-to-Face Multilingual Communication Environment for Illiterate People

In the medical field, a serious problem exists with regard to communication between hospital staffs and patients. Currently, although a medical translator accompanies a patient to medical care facilities, round-the-clock or emergency support is difficult to provide due to increasing requests. The medical field has high expectations from information technology. Therefore, we have developed a support system for multilingual medical reception termed M3. We have installed our system in the Kyoto City Hospital in Japan. However, we found that our system cannot provide support to illiterate people. If an illiterate person and another person speak different languages, it is difficult the other person to communicate face to face with the illiterate person while explaining the meaning of texts shown on the display of the support system. This is one of the problems specific to the multilingual communication. There is a need to solve this problem. Therefore, we have developed a method to provide support to illiterate people engaging in multilingual face-to-face communication. We use a text-to-speech function implemented using a selector switch to provide support to illiterate people in performing operations using a touch screen. We performed an experiment to examine the effect of the proposed method. The results of the experiment are as follows. (1) From the results of the questionnaire, we find that the subjects are able to operate the selector switch easily. Therefore, we conclude that the method using the selector switch has little effect on the operation of the system. (2) Retrieval time using the text-to-speech function is five times that using the normal operation. We need to consider a structure that can retrieve the required information easily if many readings of texts are required.