Citations using URL (URL citations) that appear in scholarly papers can be used as an information source for the research resource search engines. In particular, the information about the types of cited resources and reasons for their citation is crucial to describe the resources and their relations in the search services. To obtain this information, previous studies proposed some methods for classifying URL citations. However, their methods trained the model using a simple fine-tuning strategy and exhibited insufficient performance. We propose a classification method using a novel intermediate task. Our method trains the model on our intermediate task of identifying whether sample pairs belong to the same class before being fine-tuned on the target task. In the experiment, our method outperformed previous methods using the simple fine-tuning strategy with higher macro F-scores for different model sizes and architectures. Our analysis results indicate that the model learns the class boundaries of the target task by training our intermediate task. Our intermediate task also demonstrated higher performance and computational efficiency than an alternative intermediate task using triplet loss. Finally, we applied our method to other text classification tasks and confirmed the effectiveness when a simple fine-tuning strategy does not stably work.

Negation scope resolution is the task that identifies the part of a sentence affected by the negation cue. The three major corpora used for this task, the BioScope corpus, the SFU review corpus and the Sherlock dataset, have different annotation schemes for negation scope. Due to the different annotations, the negation scope resolution models based on pre-trained language models (PLMs) perform worse when fine-tuned on the simply combined dataset consisting of the three corpora. To address this issue, we propose a method for automatically converting the scopes of BioScope and SFU to those of Sherlock and merge them into a unified dataset. To verify the effectiveness of the proposed method, we conducted experiments using the unified dataset for fine-tuning PLM-based models. The experimental results demonstrate that the performances of the models increase when fine-tuned on the unified dataset unlike the simply combined one. In the token-level metric, the model fine-tuned on the unified dataset archived the state-of-the-art performance on the Sherlock dataset.

To accelerate open science, it is encouraged to publish and share research artifacts, e.g., scholarly papers, datasets, and code, with their metadata. Although the metadata facilitates the discovery of research artifacts, the metadata is not always complete. In this study, we verify the feasibility of estimating metadata that has not yet been filled (unfilled metadata) using metadata that has been filled (filled-in metadata) for realizing automatic completion of metadata. We implement an estimation method and evaluate its performance through experiments. The method takes filled-in metadata expressed in text as input and estimates unfilled metadata using a BERT-based model. In our experiments, the filled-in metadata was set as the name of the research artifacts, and the unfilled metadata was set as their type. Experimental results showed a high estimation performance.

In Japan, the number of single-person households, particularly among the elderly, is increasing. Consequently, opportunities for people to narrate are being reduced. To address this issue, conversational agents, e.g., communication robots and smart speakers, are expected to play the role of the listener. To realize these agents, this paper describes the collection of conversational responses by listeners that demonstrate attentive listening attitudes toward narrative speakers, and a method to annotate existing narrative speech with responsive utterances is proposed. To summarize, 148,962 responsive utterances by 11 listeners were collected in a narrative corpus comprising 13,234 utterance units. The collected responsive utterances were analyzed in terms of response frequency, diversity, coverage, and naturalness. These results demonstrated that diverse and natural responsive utterances were collected by the proposed method in an efficient and comprehensive manner. To demonstrate the practical use of the collected responsive utterances, an experiment was conducted, in which response generation timings were detected in narratives.

Narrating is a fundamental human desire. In order to satisfy the desire, it is more effective to have multiple listeners listen to narratives than to have one listener does. Conversational agents, e.g., communication robots and smart speakers, are expected to listen to narratives, instead of human listeners. For such agents to be recognized as narrative listeners, it is necessary to generate responsive utterances at appropriate timings. This paper proposes a method to estimate the generation timings of responsive utterances by active listeners of spoken narratives. The proposed method estimates them by encoding acoustic and linguistic information extracted from narratives with a transformer-based approach. To evaluate the performance of the proposed method, an experiment was conducted using 131,616 responsive utterances to narratives. The experimental results demonstrated that the proposed method was effective for the estimation and both the acoustic and linguistic information was useful.

This paper proposes a new representation for CCG derivations.CCG derivations are represented as trees whose nodes are labeled with categories strictly restricted by CCG rule schemata.This characteristic is not suitable for span-based parsing models because they predict node labels independently.In other words, span-based models may generate invalid CCG derivations that violate the rule schemata.Our proposed representation decomposes CCG derivations into several independent pieces and prevents the span-based parsing models from violating the schemata.Our experimental result shows that an off-theshelf span-based parser with our representation is comparable with previous CCG parsers.

A gapping construction consists of a coordinated structure where redundant elements are elided from all but one conjuncts.This paper proposes a method of parsing sentences with gapping to recover elided elements.The proposed method is based on constituent trees annotated with grammatical and semantic roles that are useful for identifying elided elements.Our method outperforms the previous method in terms of F-measure and recall.

Nowadays, spoken dialogue agents such as communication robots and smart speakers listen to narratives of humans. In order for such an agent to be recognized as a listener of narratives and convey the attitude of attentive listening, it is necessary to generate responsive utterances. Moreover, responsive utterances can express empathy to narratives and showing an appropriate degree of empathy to narratives is significant for enhancing speaker’s motivation. The degree of empathy shown by responsive utterances is thought to depend on their type. However, the relation between responsive utterances and degrees of the empathy has not been explored yet. This paper describes the classification of responsive utterances based on the degree of empathy in order to explain that relation. In this research, responsive utterances are classified into five levels based on the effect of utterances and literature on attentive listening. Quantitative evaluations using 37,995 responsive utterances showed the appropriateness of the proposed classification.

The Penn Treebank (PTB) represents syntactic structures as graphs due to nonlocal dependencies.This paper proposes a method that approximates PTB graph-structured representations by trees.By our approximation method, we can reduce nonlocal dependency identification and constituency parsing into single treebased parsing.An experimental result demonstrates that our approximation method with an off-the-shelf tree-based constituency parser significantly outperforms the previous methods in nonlocal dependency identification.

This paper proposes a left-corner parser which can identify nonlocal dependencies.Our parser integrates nonlocal dependency identification into a transition-based system.We use a structured perceptron which enables our parser to utilize global features captured by nonlocal dependencies.An experimental result demonstrates that our parser achieves a good balance between constituent parsing and nonlocal dependency identification.

We introduce a subway information system which utilize WiFi location technology for supporting a person in the underground. The system is composed of a mobile terminal with a WiFi device and a communication server. We have developed seven location aware applications for the mobile terminal. Each of the application helps the user with current location information. We have performed a demonstration experiment in the subway of Nagoya City with 35 subjects and got a positive acceptance of the system.

This paper proposes a new neuron architecture for a network similar to the radial basis function (RBF) network. The network with the proposed neuron, which we call a pseudo RBF network, is aimed for pattern classifications. Same as the conventional RBF network, each neuron in the hidden layer of the network is associated with a single cluster that represents a subclass. The proposed neuron effectively evaluates the possibility of the input vector belonging to its cluster. The pseudo RBF network with the proposed neuron is applied to a hand posture recognition system. Input image is preprocessed through horizontal/vertical projection followed by discrete Fourier transforms (DFTs) that calculate the magnitude spectrum. The magnitude spectrum is used as the feature vector to be fed to the network. Use of the magnitude spectrum makes the system very robust against the position changes of the hand image. The simulation results show that the average recognition rate of the system is 98% even though the hand positions are changed randomly.

Spoken monologues feature greater sentence length and structural complexity than do spoken dialogues. To achieve high parsing performance for spoken monologues, it could prove effective to simplify the structure by dividing a sentence into suitable language units. This paper proposes a method for dependency parsing of Japanese monologues based on sentence segmentation. In this method, the dependency parsing is executed in two stages: at the clause level and the sentence level. First, the dependencies within a clause are identified by dividing a sentence into clauses and executing stochastic dependency parsing for each clause. Next, the dependencies over clause boundaries are identified stochastically, and the dependency structure of the entire sentence is thus completed. An experiment using a spoken monologue corpus shows this method to be effective for efficient dependency parsing of Japanese monologue sentences.

We have been attempting to realize simultaneous machine interpretation. However, determining the interpreting utterance timing is as difficult as determining translation units. This remains a major concern for the development of such a speech translation system. It is also crucial for the system’s users that the speech generated by the system is clear and easy to listen to. In this paper, we focus attention on the pauses that partly characterize simultaneous interpreters’ utterances. We attempt to analyze the results of an experiment conducted using 31 subjects on the relationship between listener-friendliness and the length of pauses in speech, using the CIAIR simultaneous interpretation database as the data source. The results generated some knowledge about listener impressions of simultaneous interpretation, which will be helpful for realizing simultaneous machine interpretation.

In applications of spoken monologue processing such as simultaneous machine interpretation and real-time captions generation, incremental language parsing is strongly required. This paper proposes a technique for incremental dependency parsing of Japanese spoken monologue on a clause-by-clause basis. The technique identifies the clauses based on clause boundaries analysis, analyzes the dependency structures of them, and tries to decide the dependency relations with another clauses, simultaneously with the monologue speech input. The dependency relations are generated at the stage before the input of the entire monologue, and therefore, our technique can be used for language parsing in simultaneous Japanese speech understanding. An experiment using Japanese monologues has shown that our technique had the same degree of the performance as the usual dependency parsing for monologue sentences.

This paper describes the design, analysis and utilization of a simultaneous interpretation corpus. The corpus has been constructed at the Center for Integrated Acoustic Information Research (CIAIR) of Nagoya University in order to promote the realization of the multi-lingual communication supporting environment. The size of transcribed data is about 1 million words, and the corpus would deserve to be called the simultaneous interpretation corpus of the largest-in-the-world class. The discourse tag and the utterance time tag were given to the corpus, and some software tools for corpus analysis in order to support the practical use of the corpus have been developed. Therefore, the corpus is expected to be useful not only for the development of simultaneous interpreting systems but also for the construction of an interpreting theory.

This paper presents construction of a spoken dialogue system using a large-scale spoken dialogue corpus with intention tags. In this system, all of main components, such as speech understanding, dialogue management, and response generation, are constructed with corpus-based methods. An evaluation experiment using a test set has shown that the performance of the corpus-based dialogue system is improved by adding examples.

Abstract This paper proposes a method of speech intention understand-ing based on a spoken dialogue corpus to which the intentiontags are given. The intention tag expresses the task-dependentintention of the speaker, and therefore, the proper understand-ing enables a spoken dialogue system to take appropriate ac-tions. We have tagged about 35000 utterances in the CIAIR in-car speech database. In our method, several decision trees forintention understanding are constructed. By constructing deci-sion trees and using them at the same time, the strong amount ofcharacteristic features related to intentions can be retrieved, andit can also be robustly coped with the diversity of the utterances.An experiment on inference of utterance intentions has shown73.1% accuracy. 1. Introduction In order to interact with a user naturally and smoothly, it is nec-essary for a spoken dialogue system to understand the intentionof the user exactly. As a method of speech intention under-standing, example-based approaches have been considered sofar [1, 3, 6]!%In general, these approaches involve comparinga spoken utterance with examples in a correctly-tagged corpus.The intention of the utterance is regarded as the intention tag ofthe most similar example in the corpus. However, it is difficultto infer the intention of the utterance to which any example inthe corpus is not similar.This paper proposes a method of speech intention under-standing based on a spoken dialogue corpus. The method con-structs several decision trees for intention understanding. Byconstructing several decision trees, the strong amount of char-acteristic features related to intentions can be retrieved, and itcan also be robustly coped with the diversity of the utterances.So far, we have designed an organization of the tags which iscalled Layered Intention Tag(LIT). These tags show more de-tailed utterance intention rather than the illocutionary act level,and have built the corpus[2, 3]. LIT is divided into several lay-ers considering the relevance between an intention and variousphenomena relevant to an utterance, such as a style, a keyword,a sentence structure. This method constructs several decisiontrees by using this corpus and infers the intention by combiningthem.In order to evaluate the effectiveness of our method, an ex-periment on inference of the utterance intentions was conductedusing the driver utterances about restaurant search recorded ona large-scale in-car spoken dialogue corpus of CIAIR[4, 5]. Asa result, the effectiveness of the method was confirmed.

In this paper, we propose a new method to expand an examplebased spoken dialogue system to handle context dependent utterances. The dialogue system refers to the dialogue examples to find an example that is suitable to promote dialogue. Here, the dialogue contexts are expressed in the form of dialogue slots. By constructing dialogue examples with the text of utterances and the dialogue slots, the system handle context dependent dialogue. And we also propose a new framework of spoken dialogue, named “GROW architecture” that consists of the dialogue system and a Wizard-of-OZ (WOZ) system. By using the WOZ system to add dialogue examples via network, it becomes efficient to augment dialogue examples.

Grant-in-Aids for Young Scientists of the Ministry of Education, Science, Sports and Culture, Japan; The Tatematsu Foundation

This paper describes an advanced spoken language corpus which has been constructed by enhancing an in-car speech database. The corpus has the following characteristic features: (1) series Advanced tag: Not only linguistic phenomena tags but also advanced discourse tags such as sentential structures, and utterance intentions, have been provided for the transcribed texts. (2) series Large-scale: The sentential structures and the intentions are currently provided for 45,053 phrases and 35,421 utterance units, respectively. (3) series Multi-layer: The corpus consists of different levels of spoken language data such as speech signals, transcribed texts, sentential structures, intentional markers and dialogue structures, moreover, they are related with each other. It allows a very wide variety of analysis of spontaneous spoken dialogue to utilize the multi-layered corpus. This paper also reports the result of investigation of the corpus, especially, focusing on the relations between the syntactic style and the intentional style of spoken utterances.

This paper reports the details of the collection of the multimedia data such as audio, video and auxiliary information of the vehicle during a spoken dialogue in a moving car. The system specially built in a Data CollectionVehicle (DCV) supports synchronous recording of multi-channel audio data from 16 microphones, 3-channel video data and the vehicle related data. Multimedia data has been collected for three sessions of spoken dialogue in about a 60-minute drive by each of 200 subjects. Data has been collected for two dialogue modes:(1) prompted dialogue between the driver and an accompanying operator and (2) natural dialogue between the driver and a telephone operator for information access over a cellular phone while driving a car. The corpus can be used for analysis of multimedia data in a moving car environment and also for modeling spoken dialogue in scenarios such as information access while driving a car.

The paper proposes an effective incremental parsing method for such interactive natural language processing systems as real-time dialogue systems, simultaneous machine interpreting systems, etc. This method produces the analysis of the input while it is being received. It can efficiently deal with not only the normal input, which is piecemeal addition to the input from left to right, but also such changes of the input as insertion and deletion. For such changes of the input, this method exploits parts of the previous analyses. We implemented this method on a workstation and conducted an experiment. We confirm that the method can be expected to be useful for an online language processing system.

This paper describes a database consisting of speech and language, which we are currently constructing for the purpose of the research on machine interpretation. The database contains bilingual data of lectures and dialogues. We have collected the speech of about 72 hours in total and transcribed it into the text manually. We have investigated the database in order to acquire empirical knowledge of human interpreting. In this paper, we report the characteristic features of spoken language by Japanese-to-English interpreters.

Towards a real-time spoken dialogue system, several incremental parsing methods have been proposed so far. They constructs syntactic structures for an initial fragment of an input sentence. However, they have a problem that the structures do not necessarily represent the syntactic relation correctly. The problem is caused by the ambiguity of initial fragments. This paper proposes an incremental disambiguation method, which decides correct structures at an stage where the entire input is not completed. The method finds the structures which are correct independently of the remaining input. When correct structures cannot be decided, the method delays the decision. Since the disambiguation is executed for every word input, the method can find correct structures at an early stage.

The Center for Integrated Acoustic Information Research (CIAIR) at Nagoya University has been collecting speech corpora in moving cars which are made available as resources to advance the research and development of robust ASRs and spoken dialogue systems under high-noise conditions. The speech corpus consists of (1) phonetically balanced sentences, (2) digit strings, (3) discrete words and (4) transcribed spoken dialogues between drivers and information systems for navigation and information retrieval. These data are collected in vehicles under both idling and driving situations. The language of the corpus is currently Japanese. The number of subjects is currently about 300, total recording time is over 200 hours and total corpus size is about 160GByte. We have also been recording video images from three different angles, vehicle-control signals, and vehicle location, all synchronized with the speech recording. We report the objective of the speech corpus, the recording methods and the recording vehicle developed.