New Frontiers in Artificial Intelligence

This year’s workshop was the seventh LENLS and the second held at the Campus Innovation Center in Mita, Tokyo in November of 2010 as part of the JSAI International Symposia on AI program of the Japanese Society for Artificial Intelligence. The workshop featured invited talks by Alexandru Baltag, on dynamic epistemic logic, and Shunsuke Yatabe, on paradoxes—in particular, paradoxes arising from theories of truth—and coinduction. A paper based on that second talk, “Yablo-like Paradoxes and Co-induction”, appears in the present volume. In addition, 18 papers were selected by the program committee (see Acknowledgements) from the submitted abstracts for presentation and as alternates.

As always with the LENLS workshops, the content of the presented papers was rich and varied. This year’s theme was salience: its realization in natural language (as well as human reasoning and knowledge representation) and how it should be given a formal representation. In addition to papers on this topic, a wide range of topics were represented, including, on the empirical side, focus phenomena, negative polarity items, reduplication, factivity and vagueness, and, on the theoretical side, variable-free semantics, game-theoretic pragmatics, categorial grammar, and various sorts of dynamic logics. In the remainder of this introduction, I will briefly indicate the content of the papers selected from this wide array of topics to appear in the present volume.

Segmented Discourse Representation Theory (SDRT) [2,7] provides a dynamic semantics for discourse that exploits a rich notion of discourse structure. According to SDRT, a text is segmented into constituents related to each other by means of rhetorical relations; the resulting structure, known as a

segmented discourse representation structure

or SDRS has various semantic effects. This theory has shown how discourse structure makes contributions to the interpretation of a variety of linguistic phenomena, including tense, modality, presupposition, the interpretation of anaphoric pronouns and ellipsis. SDRT exploits dynamic semantics [20,14] to interpret SDRSs. We investigate here the advantages of integrating SDRT within continuation style semantics of the sort developed in [17].

This paper aims to provide a logical background for Combinatory Categorial Grammar (CCG) and its typological discussions. Based on the Curry-Howard correspondence between Gentzen-style proof systems and Lambek Lamda Calculi, and those between Hilbert-style proof systems and substructural

BCWK

-logic, I define a new class of logic which provides subclasses for each of the substructural combinatory logics, called

Subdirectional Combinatory Logic

, and propose that CCG is a subdirectional logic of a relevance logic (

Combinatory Hypothesis

). This hypothesis allows us to discuss typological parameters in universal grammar in terms of the presence/absence of a certain directional combinators.

This paper details a method of providing interpretations for kakari-uke dependency structures between bunsetsu (phrasal units) for utterances of Japanese. The technique employs a definition of evaluation against a structured assignment to generate meaning representations. An implementation is described which has led to a robust and efficient system for processing unrestricted texts of Japanese. In addition the paper offers a theory for grammatical/valid bunsetsu dependencies in terms of the class of kakari-uke structures combined with case frame information that successfully support evaluation.

Vagueness is a pervasive feature of natural languages that is challenging semantic theories and theories of language evolution alike. We focus here on the latter, addressing the challenge of how to account for the emergence of vague meanings in signaling game models of language evolution. We suggest that vagueness is a natural property of meaning that evolves when

boundedly rational

agents repeatedly engage in cooperative signaling.

This paper examines how the binding relation can be established between discontinuous possessor and possessee NPs in Japanese. Japanese has various peculiar constructions involving relational nouns, in which their possessors seem to appear in positions quite far from their original positions. Without positing empty pronouns as necessary in the standard theory of binding, we adopt the variable-free semantics proposed by Jacobson [3], [4] and [5] to give a unified account to binding phenomena in these constructions. In the binding-without-pronoun approach, we can associate discontinuous possessor and possessee NPs using the type-shift rule for binding, assuming that binding should not be dealt with as relations between linguistic expressions, but as relations between argument slots of functional expressions.

Vagueness is a ubiquitous feature that we know from many expressions in natural languages. It can invite a serious problem: the Sorites Paradox. The aim of this paper is to proposed a new version of complete logic for vague predicates - salient-similarity-based vague predicate logic (

SVPL

) that can avoid the Sorites Paradox and give answers to all of the Semantic Question, the Epistemological Question and the Psychological Question given by Graff.

We review three pairwise similar paradoxes, the modest liar paradox, McGee’s paradox and Yablo’s paradox, which imply the

ω

-inconsistency. We show that is caused by the fact that co-inductive definitions of formulae are possible because of the existence of the truth predicate.

Notions of normal intensionality and intensionality preserving negation are introduced. Both these notions apply to intensional sentential operators. When applied to factive operators they allows us to understand the way these operators give rise to presuppositions and their duals, assertions. They allow us also to distinguish semantically emotive factives from non-emotive ones.

The Fourth InternationalWorkshop on Juris-Informatics (JURISIN 2010) was held on November 18–19, 2010 at Campus Innovation Center in Tokyo, Japan, as a part of the second JSAI International Symposia on AI (JSAI-isAI 2010), with supports of Japan Advanced Institute of Science and Technology (JAIST) and National Institute of Informatics (NII).

With the advent of the telecommunication technologies, a new form of disputes taking place in virtual environments started to emerge. In order to settle these disputes, Online Dispute Resolution tools appeared. In this paper we present such a tool, aimed at supporting mediation between two or more parties. Specifically, this tool looks at past known mediation processes and tries to guide the process into a successful outcome. Moreover, this tool places an emphasis on providing useful information for the parties before the resolution process actually starting, helping them take more rational decisions. Specifically, in this work we target scenarios in which one or more party exhibits avoiding or uncooperative conflict styles, i.e., the party cannot or is not willing to generate valid proposals for dispute resolution.

To safeguard fairness for all parties involved and proper procedure, actions within a legal context are heavily constrained. Detailed laws determine when actions are permissible and admissible. However, these restrictions do not prevent participants from acting. In this paper we present a methodology to support legal reasoning using institutions—systems that specify the normative behaviour of participants—and a corresponding computational model. We show how it provides a useful separation between the identification of real world actions, if and how they affect the legal model and how consequences

within

the legal model can be specified and verified. Thus, it is possible to define a context, introduce a real-world event and examine how this changes the state of the legal model: hence, the modeller can explore both model adequacy and that of the legal framework from which it is derived, as well as offering a machine-usable legal ‘oracle’ for software components. We illustrate the use of our framework by modelling contract cancellation under Japanese contract law.

This paper describes how to analyze Japanese statutory sentences. Although statutory sentences have many technical terms and characteristic structures, the design of the general Japanese corpus has no tags to handle such structures and usual Japanese parsers cannot analyze them correctly. Thus, we propose a new design of syntactic tags for statutory sentences and develop a support tool that corrects the output of a parser. In this paper, we focus on dependency structure of sentences and present an overview of the design and compilation of the corpus of Japanese statutory sentences.

In this paper, we propose a legal reasoning system called

PROLEG

(PROlog based LEGal reasoning support system) based on the Japanese “theory of presupposed ultimate facts” (called “Yoken-jijitsu-ron” in Japanese, the JUF theory, in short). The theory is used for decision making by judges under incomplete information. Previously, we proposed a translation of the theory into logic programming. However, it turns out that the knowledge representation in logic programming is difficult for lawyers to understand. So, in this paper, we change knowledge representation of rules in the JUF theory in PROLEG so that we reflect lawyers’ reasoning using the idea of “openness” proposed by a judge who is a main investigator of the JUF theory.

The First International Workshop on Advanced Methodologies for Bayesian Networks (AMBN 2010) was held on November 18–19, 2010, at the Campus Innovation Center in Tokyo, Japan. It was part of the second JSAI International Symposium on AI (JSAI–isAI 2010), and was co-hosted by the Minato Discrete Structure Manipulation System Project on Exploratory Research for Advanced Technology (ERATO) of the Japan Science and Technology Agency (JST). It was supported by Mathematical Systems Inc.

We present in this paper a framework of approximate probabilistic inference which is based on three simple concepts. First, our notion of an approximation is based on “relaxing” equality constraints, for the purposes of simplifying a problem so that it can be solved more readily. Second, is the concept of “compensation,” which calls for imposing weaker notions of equality to compensate for the relaxed equality constraints. Third, is the notion of “recovery,” where some of the relaxed equality constraints are incrementally recovered, based on an assessment of their impact on improving the quality of an approximation. We discuss how this framework subsumes one of the most influential algorithms in probabilistic inference: loopy belief propagation and some of its generalizations. We also introduce a new heuristic recovery method that was key to a system that successfully participated in a recent evaluation of approximate inference systems, held in UAI’10. We further discuss the relationship between this framework for approximate inference and an approach to exact inference based on symbolic reasoning.

Causal relationships among a set of observed variables are often modeled using directed acyclic graph (DAG) structures, and learning such structures from data is known as the causal discovery problem. We here consider the learning of linear non-Gaussian acyclic models [9] with hidden variables [5]. Estimation of such models is computationally challenging and hence only possible when the number of variables is small. We present an algorithm for obtaining partial but in the large sample limit correct information about pairwise total causal effects in such a model. In particular, we obtain consistent estimates of the total effects for all variable pairs for which there exist an unconfounded superset of observed variables. Simulations show that the estimated pairwise total effects are good approximations of the true total effects.

Item response theory (IRT) is widely used for test analyses. Most models of IRT assume local independence, meaning that when the ability variables influencing the test performance are held constant, an examinee’s responses to any pair of items are statistically independent. However, many factors might cause local dependence among items. Consequently, conditional independence (CI) tests are needed among items given a latent ability variable. Hashimoto and Ueno (2011) proposed the latent conditional independence (LCI) test. While other CI tests are sensitive to dependencies of items aside from the targets, the LCI test is robust to such dependencies. However, when the two target items affect the same items, the LCI test might fail to detect local independency between the targets. The previous work of Hashimoto and Ueno (2011) is improved on to obtain a more accurate detection method.

Optimal search on Bayesian network structure is known as an NP-hard problem and the applicability of existing optimal algorithms is limited in small Bayesian networks with 30 nodes or so. To learn larger Bayesian networks from observational data, some heuristic algorithms were used, but only a local optimal structure is found and its accuracy is not high in many cases. In this paper, we review optimal search algorithms in a constraint search space; The skeleton of the learned Bayesian network is a sub-graph of the given undirected graph called super-structure. The introduced optimal search algorithm can learn Bayesian networks with several hundreds of nodes when the degree of super-structure is around four. Numerical experiments indicate that constraint optimal search outperforms state-of-the-art heuristic algorithms in terms of accuracy, even if the super-structure is also learned by data.

Latent tree (LT) models are a special class of Bayesian networks that can be used for cluster analysis, latent structure discovery and density estimation. A number of search-based algorithms for learning LT models have been developed. In particular, the HSHC algorithm by [1] and the EAST algorithm by [2] are able to deal with data sets with dozens to around 100 variables. Both HSHC and EAST aim at finding the LT model with the highest BIC score. However, they use another criterion called the cost-effectiveness principle when selecting among some of the candidate models during search. In this paper, we investigate whether and why this is necessary.

Service science has been raised, and coming to be established as a research domain all over the world. This workshop in Tokyo has been motivated by the systems-design dimension of the service science. We aim to share and discuss a progressive vision to develop methods for innovating systems of service resources where novel values are created and supplied sustainably. A "service system" here is an artificially organized or self-organized active integration of the following resources:

1

Participants, i.e., providers and consumers of services, where a provider of a service may turn into a consumer in different contexts

2

Money, or other entities representing value, and their active flows: An utterance of praise can be also this kind of entity

3

Supply chains i.e., the chain of interactions from creators of service resources (products, information, food, etc) to consumers of services, and

4

Tools (computers, robots, sensing devices, etc) aiding the activities of (1), (2), and (3)

In this workshop, we discussed methods for designing and realizing service systems and parts of a service system, with positioning resources of the four kinds above in the systems to be created. By this, we aim to respond to the social demand to design an environment for value-creative and dynamic interactions among participants via resources in the market, rather than merely passing existing products and services from providers to customers for predetermined prices.

The volumes and velocity of data are growing at unprecedented rates, often physically distributed, have access constraints, and requirements to leverage the diverse computational fabrics like clouds and grids. The Meandre data-intensive component-based application infrastructure can leverage diversity and enables extremely scalable server clusters and applications to address these challenges. Data-intensive flows can: be web-services and/or computational tasks; co-locate processing with data; orchestrate cloud computing resources; and leverage grid resources with distributed execution. Meandre from a laptop to a cloud, grid, or server as analytical computational tasks and/or web-services in data-intensive flows made up of components that provide deployment and execution strategies for extreme scalability.

The expanding tourism market, in particular of East Asia, has drawn great interests and has raised a series of significant issues for researchers and planners in sustainable development. Unsustainable tourism development caused problems such as loss of natural resources, conflicts between tourists and local residents, and so on. This research intends to understand the development process of Recreational Business Districts (RBDs) in tourism areas and to provide a framework for supporting sustainable tourism development by analyzing interactions between tourists and RBD. An Agent-Based Simulation (ABS) combined with Geographic Information System (GIS) provides planning supports to tourism bureaus and policy makers to help them assess possible future development plans in tourism under certain scenarios.

To recognize technical trends is essential for the interested parties to understand the development directions of a technology at the industry level. Therefore, a research design has been formed for conducting technology roadmapping where association analysis is employed to measure the co-occurrence in the patent data sets and data crystallization is adopted to build the relations between clusters. Consequently, a technology roadmap for thin-film solar cell was constructed and relations between topics and subtopics were identified. Finally, according to the relations, the technical trends of thin-film solar cell were recognized and depicted.

Reliable and efficient operations are essential for successful service business, and Performance Indicators (PIs) are useful tools for assessing appropriateness of service operations and providing cues to remedy flawed performance. Performance indicators should be based on objective data on operation performance, derivable by concrete and simple calculation rules, and exhaustively related to business goals. Development of such PIs is not an easy task, and this work tries to propose a framework for developing PIs using an application example of operations management in an airline. The proposed scheme of development is so general that it is applicable also to services other than airline business.

Technology is one of the powerful media in education. Especially, the remedial instruction by computer technology is widely used from K-12 to higher education. This study applied the grounded theory approach analyzing data with visual association map to find the impact factors in English remedial e-learning instruction. The significances of this paper are proposing the results with visual representation and with an event scenario to instructors or instructional designers. The results indicated that the engineer school, science school, management school and concrete-sequential learning style were the main factors in learner characteristic perspective. With these learner characteristic impact factors, the instructional factors appeared in remedial English e-learning course. Future studies are needed.

This research aims at investigating intermediate-scale workflows using the surgical staff’s movement pattern. In this study, we have introduced an ultrasonic location aware system to monitor intraoperative movement trajectories on surgical staffs for the workflow analysis. And we developed trajectory data mining for surgical workflow segmentation, and analyzed trajectory data with multiple cases. As a result, in 77.18% of total time, a kind of current operation stage could be correctly estimated. With high accuracy 85.96%, the estimation using trajectory data was able to distinguish whether a current 5 minutes was transition time from one stage to another stage or not.. Based on these results, we are implementing the surgery safe support system that promotes safe & efficient surgical operations.

From the innovative diffusion theory, it is known that the early adopters have the characteristic of being liable to accept new products and generate new value from using new products. Through their social networks, they also transmit information about new products to the early majority. Moreover, from the use and gratification theory, if a new product provides users with usability of information, social value or entertainment, it is likely that this new product will be accepted because users’ physical/mental need is fulfilled. Based on the above two principles, this study thereby develop a Human-Computing Grounded System. Expert first divides data as technical specification or value induced by early adopters, grounded theory analysis is then performed interactively. Results in this study indicate the new value induced by early adopters using new products which supported by new technology embedded in the new product. Meanwhile, inferred from the use and gratification theory, it is easier for the majority to accept a new product if this product offers more new values. By integrating multi-type of values conceived by the early adopters, our research model shows that indeed products with more and new values attract more majority.

It is considered that many people are struggling with diseases that are difficult to cure. In general, it takes a long time to diagnose and to cure such diseases. In Japan, methods for curing 56 incurable diseases have been studied. However, methods for early diagnosis of incurable diseases have not been studied. To make early diagnoses for incurable diseases improves the quality of life of patients. This paper proposes a new service system that supports latent patients of incurable diseases by early diagnosis. For using this system, users have to prepare a text in which episodes about patients’ experiences that have been caused because of their diseases are written. The system takes such a text as input, and then the system extracts common factors among episodes, i.e., keywords appearing in several episodes. The system output the extracted keywords as keywords relating to symptoms of an incurable disease. We experimented the system and extracted keywords relating to the symptoms of retinitis pigmentosa which was one of the incurable eye diseases. Most of the symptoms relating to the extracted keywords were not known even by medical doctors. Some of them indicated symptoms in the early stage of the disease. The experiment brought us one step closer to the early diagnosis of incurable disease as one of the service systems.

The Innovators Market Game (IMG) has been proposed as a tool for aiding innovative thoughts and communication, coming after our 10-year experiences in chance discovery where tools of data visualization have been applied in cases of decision making of business teams. This game enabled to run and accelerate the process of innovation, as well as to train human’s talent of analogical and combinatorial thinking. In this paper, we compare the effects of IMG as table game and a Web-based environment of communication where the fundamental features of IMG are imported. The effects of these two agorae, one on the table and the other on the Web, are compared. It is shown that there are some aspects of innovative communication we can enhance, some on the table and some on the Web. This directs us to future work merging the two in one process as a generalization of IMG.

This paper proposes multiscale service design method through the development of support service for prevention and recovery from dementia. Proposed multiscale service model consists of tool, event, human, network, style and rule. Service elements at different scales are developed according to the model. Firstly, the author proposes and practices coimagination method, which is expected to prevent the progress of cognitive impairment. Coimagination support system and program were developed as “tool” and “event”. Then, Fonobono Research Institute was established as a “network” for “human” who studies coimagination, which is a multisector research organization including older adults living around university campus, companies providing welfare and medical services, local government, medical institution, researchers of the University of Tokyo and Keio University. The institute proposes and realizes lifelong research as a novel life “style” for older adults, and discusses second social system for older adults as an innovative “rule” for social system of aged society.

In our society, there are many opportunities to exchange opinions on decision. Problems may occur when we exchange opinions. It is desirable for solving problems to make an environment that can achieve a smooth opinion exchange. There are "divergence phase" and "convergence phase" in the opinion exchange for decision. The divergence phase is to enumerate various choices related to the subject, and to consider possibility. On the other hand, in the convergence phase, participants investigate alternatives arose in the divergence phase, and argue for making a single conclusion.

In this paper, we focus on the convergence phase, and propose a system that supports the process narrowing the listed alternatives by using RFID tags. That is, the system supports the smooth progress of opinion exchanges by controlling the time, the speech order, and the alternative evaluation for leading a single conclusion that many participants can accept.