Teruaki Aizawa

Automatic linguistic analysis for language teachers: the case of zeros

This paper presents the Natural Language Processing-based linguistic analysis tool that we have developed for Japanese as a Second Language teachers. This program, Zero Detector (ZD), aims to promote effective instruction of zero anaphora, on the basis of a hypothesis about ideal conditions for second language acquisition, by making invisible zeros visible. ZD takes Japanese written narrative discourse as input and provides the zero-specified texts and their underlying structures as output. We evaluated ZDs performance in terms of its zero detecting accuracy. We also present an experimental report of its validity for practical use. As a result, ZD has proven to be pedagogically feasible in terms of its accuracy and its impact on effective instruction.

Completion of Japanese sentences by inferring function words from content words

A method of generating a Japanese sentence by inferring function words from content words using valency patterns is presented.A procedure for selecting an appropriate function word, on the assumption that correct content words have been selected for a given phrase lattice, is described. A method of inferring a correct verb when verbs are recognized less accurately than nouns by the speech recognition system is described. Sentences are produced from content words as inputs by using the valency patterns obtained from collected dialogue sentences in a restricted task domain. Using the semantic features of preceding nouns and valency patterns allow a fairly restricted number of candidate verbs to be inferred.This method eliminates possible errors at the interface between speech recognition and machine translation (component technologies of an Automatic Telephone Interpretation system) and selects the most appropriate candidate from a lattice of typical phrases output by the speech recognition system.

A Description Method of Syntactic Rules on Japanese Filmscript

The purpose of this paper is to propose a description method of syntactic rules, particularly, to manage manually. Our aims are the decrease of the number of rules and the structuring of them. The former is achieved by describing the rules with regular expression, and the latter is constructed by step-by-step analysis. These descriptions make the management of the rules simple and easy. Considering such concerns, we report the effectiveness of our method with some experimental results.

An error measure for japanese morphological analysis using similarity measures

The aim of this paper is to propose a Japanese morphological error measure in order to automatically detect morphological errors when analyzing. In particular, we focused on three similarities as the measure and experimented using them. From our experimental results, it was found that the precision of one of measures was 74% and it functioned well.

Emotion-based planning evaluation method

It has been suggested by a number of researchers that emotional systems in agents might have great practical value. We propose a system that makes a decision about a plan, basing its judgment on emotion (i.e. pleasure/displeasure) extracted from knowledge, in an attempt to build a human-like computer agent, better able to communicate with users. We present an emotion-based plan evaluation method, and then provide the result simulated on naturally occurring short-story scenarios and show that plan evaluation based on artificial emotion generated by our method successfully matches the one based on human emotion.

Using Mitate-shi Related to the CONTAINER Schema for Detecting the Container-for-Contents Metonymy

This paper shows a new approach to detect metonymy by noting mitate-shi : kinds of Japanese expressions, related to how to recognize objects people see, such as “no naka.” This idea is based on the concept of CONTAINER schema which reflects on our cognitive capacity, that is, people recognize various objects as a kind of container. Using mitate-shi, we can automatically capture humanlike and detailed knowledge base from corpora. In order to detect the container for contents metonymy, we use such knowledge base and attempt to estimate our effectiveness of detecting the metonymy.

Automatic generation method for spoken sentences by genetic programming and its application to emotion oriented dialog systems

We have already developed an emotional interface system with facial expressions applying Emotion Generation Calculation. However, we found that users required an interface that supported natural spoken sentence generation in order to carry out a smooth dialog between human and computers. The authors propose a spoken sentence generation system using Genetic Programming (GP), which can generate various expressions while keeping the predetermined rules of the conversations. To verify the effectiveness of the proposed method, we applied the method to a WWW-based health care system.

Using a natural-artificial hybrid language for database access

In this paper we propose a natural-artificial hybrid language for database access. The global construction of a sentence in this language is highly schematic, but allows expressions in the chosen language such as Japanese or English. Its artificial language part, SML, is closely related to our newly introduced data model, called scaled lattice. Adopting Japanese as its natural language part, we implemented a Japanese-SML hybrid language processing system for our compact database system SCLAMS, whose database consists of scaled lattices. The main features of this implementation are (1) a small lexicon and limited grammar, and (2) an almost free form in writing Kana Japanese.