This paper presents a new way of formalizing the coalition structure generation problem (CSG) so that we can apply constraint optimization techniques to it. Forming effective coalitions is a major research challenge in AI and multi-agent systems. CSG involves partitioning a set of agents into coalitions to maximize social surplus. Traditionally, the input of the CSG problem is a black-box function called a characteristic function, which takes a coalition as input and returns the value of the coalition. As a result, applying constraint optimization techniques to this problem has been infeasible. However, characteristic functions that appear in practice often can be represented concisely by a set of rules, rather than treating the function as a black box. Then we can solve the CSG problem more efficiently by directly applying constraint optimization techniques to this compact representation. We present new formalizations of the CSG problem by utilizing recently developed compact representation schemes for characteristic functions. We first characterize the complexity of CSG under these representation schemes. In this context, the complexity is driven more by the number of rules than by the number of agents. As an initial step toward developing efficient constraint optimization algorithms for solving the CSG problem, we also develop mixed integer programming formulations and show that an off-the-shelf optimization package can perform reasonably well.

More and more companies are putting emphasis on communication skill in the recruitment of their employees and are adopting group discussion as part of recruitment interview. In our project, we aim to develop a system that can provide advices to its users in improving the impression of their communication skill during group discussion. In this paper, we focus on the functional roles of the participants in group discussion and report the results of the analysis of the relationship between communication skill impression and functional roles. This work is based on a group discussion corpus of 40 participants. The participants’ communication skill of the corpus was evaluated by 21 external experts who had experience of recruitment. In addition, seven functional roles: Follower, Gatekeeper, Information giver, Objector, Opinion provider, Passive participant, and Summarizer were defined and annotated. Furthermore, we analyzed the conversational situations of corpus and the difference of between participants with high-score and low-score communication skill in these situations.

In recent years, companies are seeking for communication skill from their employers. More and more companies adopt group discussions in employer recruitment to evaluate the applicants' communication skill. However, the opportunity to improve communication skill in group discussion is limited due to the lack of partners. In order to solve this issue, our ongoing project is aiming to build a virtual agent or a robot that can participate group discussion, so that its users can repeatedly practice group discussion with it. In this paper, we propose the models in directing the agent's attention toward the other participants in three situations: when the agent is speaking, when the agent is listening, and when no participant is speaking. First, we gathered a data corpus of the discussion of 10 four-people groups. We then use low-level non-verbal features including attention of other participant, voice prosody, head movements, and speech turn extracted in the 10-hour corpus to train support vector machine models to determine the agent's attention on the other participants, or the material. The performance of the detection models in F-measure range between 0.4 and 0.6.

The training program of high school teachers in Japan lacks the chance to practice teaching skills and the admission of classes. The result is, many young teachers left their jobs in the first year due to frustration and other mental issues. In order to relieve this issue, we are running a project to develop a platform of virtual school environment and virtual students which allows the teacher trainees to practice with. As part of this platform, in order to send feedbacks to the trainee to stimulate them to improve their performance, we propose the use of the whole group of virtual students to generate per- ceivable atmospheres. Three research issues then emerge:(1) Whether the atmospheres in a classroom can be expressed in a computational (parameterized) model? (2) If the answer to (1) is 'yes', what are the elements to construct atmospheres? (3) The perception of atmospheres can be considered as very subjective, how to make the model objective becomes the last research issue. This paper presents our investigation on these issues and propose a parameterized atmosphere generation model based on empirical results.

In this paper, we aim to make a real-time system which can provide advice that teach users what should to do in what kind of situation, and to lead the user get a good communication skill impression. Towards to this goal, we confirmed the connection between the communication impression evaluation of the participants and their roles which can cover multiple non-verbal information within the group discussion by using statistical analysis. For this purpose, we use a multi-modal group discussion corpus including audio signal data and head motion data for participants observed in 30 group discussion sessions. The corpus also includes participants’ evaluated values of communication skills by 21 external experts, and their roles’ annotation which are annotated by 3 coders. We extracted the duration ratios of roles to estimate the communication skills by creating the regression model and classification model. Experiment results show that the best performance of the proposed classification model of aggression of opinions obtained best accuracy (0.90), and the of the regression model of aggression of opinions was also highest (0.56). That is to say that we can confirm the difference of communication skill impression of participant from the duration ratios of roles in a high precision by using our model.

In Japan, the training of teachers mainly relies on in-classroom lectures in universities. It is compensated with the practice for a relatively short period, say only two to three weeks in real schools. The teacher-training programs in Japan therefore lacks the practice of teaching skills and the admission of classes. The result is, many young teachers left their jobs in the first year due to frustration and other mental issues. In order to relieve this situation, we are developing a Wizard-of-OZ (WOZ) based simulation platform of a school environment with computer graphics (CG) animated virtual students. The trainees can interact with the virtual students in this immersive and realistic virtual classroom and practice their teaching and administration skills. The virtual students are operated by an operator (the wizard) from remote with a dedicated interface. In addition to the training purpose, the system is considered to be able to be used in the examination of teacher recruitment as well. In that case, the operator is supposed to be the examination investigator.

While the environment of schools become more and more complicated, the improvement of teachers' skills in teaching and management is required. In this study, we focus on the development of a Wizard-of-Oz (WOZ) platform of simulated school environment, which can be utilized for teacher training or the examination of teacher recruitment from remote. This system is comprised of two front ends, one is a simulated classroom for the trainee, the other one is the interface for the system operator / investigator. The virtual classroom contains a number of virtual students who are controlled by the operator from the remote. The operator can observe the trainee from a dedicated interface and control the behaviors of any individual student as well as the atmosphere of the whole class. The whole-class atmosphere created by relatively large number of students is modeled as a concentration-arousal two dimensional space. The prototype system is evaluated with subject experiment and the results are reported.

To make use of collective intelligence of many autonomous self-interested agents, it is important to form a team that all the agents agree. Two-sided matching is one of the basic approaches to form a team that consists of agents from two disjoint agent groups. Traditional two-sided matching assumes that an agent has totally ordered preference list of agents to be paired with. However, it is unrealistic to have a totally ordered list for a large-scale two-sided matching problem. Therefore, two-sided matching with indifferences is proposed. It allows indifferences in the preference list of agents. Two-sided matching with indifferences has two important characters weakly stable and Pareto efficiency. In this paper, we propose a new integer programming based algorithm “nucleolus” for two-sided matching with indifferences. This algorithm propose the matching which satisfies weakly stable and Pareto efficiency.

To make use of the collective intelligence of many autonomous self-interested agents, it is important to form a team on which all the agents agree. Two-sided matching is one of the basic approaches to form a team that consists of agents from two disjoint agent groups. Traditional two-sided matching assumes that an agent has a totally ordered preference list of the agents it is to be paired with, but it is unrealistic to have a totally ordered list for a large-scale two-sided matching problem. In this paper, we propose an integer programming based approach to solve a two-sided matching program that allows indifferences in agents’ preferences, and show how an objective function can be defined to find a matching that minimizes the maximum discontentedness of agents in one group.

As information and communication technologies advance, large amounts of data are created everyday. The demands for processing such big data are also increasing. To meet them, the MapReduce framework has been proposed and is now widely used. On the other hand, a rule-based system is used to implement such an intelligent system as an expert system. For applying a rule-based system to process large amounts of data, we propose a method that implements a rule system based on the MapReduce framework. We constructed a simple rule system using Hadoop, which is an open source implementation of the MapReduce framework, and compared several methods of executing a rule system. Our experimental results indicate the potential of a rule system implemented using the MapReduce framework.

Two-sided matching is a major matching problem for players in two groups. If a player joins or leaves a group, the matching must be calculated from scratch. In this paper, we propose a changeable two-sided matching problem for two-sided problems in changing environments where a player can join or leave groups. We also propose an algorithm, called the continuation algorithm, for this problem, which calculates matching using not only the preferences of the players of the two groups, but also the matching result of the original two-sided matching problem before changes occur. We demonstrate that, if the matching result of the original two-sided matching is stable, the result obtained by the continuation algorithm is also stable.

A Coalition Structure Generation (CSG) problem involves partitioning a set of agents into coalitions so that the social surplus is maximized. Recently, Ohta et al . developed an efficient algorithm for solving CSG assuming that a characteristic function is represented by a set of rules, such as marginal contribution networks (MC-nets). In this paper, we extend the formalization of CSG in Ohta et al . so that it can handle negative value rules. Here, we assume that a characteristic function is represented by either MC-nets (without externalities) or embedded MC-nets (with externalities). Allowing negative value rules is important since it can reduce the efforts for describing a characteristic function. In particular, in many realistic situations, it is natural to assume that a coalition has negative externalities to other coalitions. To handle negative value rules, we examine the following three algorithms: (i) a full transformation algorithm, (ii) a partial transformation algorithm, and (iii) a direct encoding algorithm. We show that the full transformation algorithm is not scalable in MC-nets (the worst-case representation size is Ω( n 2 ), where n is the number of agents), and does not seem to be tractable in embedded MC-nets (representation size would be Ω(2 n )). In contrast, by using the partial transformation or direct encoding algorithms, an exponential blow-up never occurs even for embedded MC-nets. For embedded MC-nets, the direct encoding algorithm creates less rules than the partial transformation algorithm. Experimental evaluations show that the direct encoding algorithm is scalable, i.e., an off-the-shelf optimization package (CPLEX) can solve problem instances with 100 agents and rules within 10 seconds.

Forming e ective coalitions is a major research challenge in AI and multi-agent systems. Coalition structure generation (CSG), which involves partitioning a set of agents into coalitions so that social surplus is maximized, is a central research topic due to its computational complexity. In this paper, we present new methods for CSG utilizing recently developed compact representation schemes for characteristic functions. We characterize the complexity of CSG under these representation schemes. In this context, the complexity is driven more by the number of synergy coalition groups than by the number of agents. Furthermore, we develop mixed integer programming formulations and show that an o -the-shelf optimization package can solve these problems quite e ciently.

This paper presents a new way of formalizing the Coalition Structure Generation problem (CSG), so that we can apply constraint optimization techniques to it. Forming effective coalitions is a major research challenge in AI and multi-agent systems. CSG involves partitioning a set of agents into coalitions so that social surplus is maximized. Traditionally, the input of the CSG problem is a black-box function called a characteristic function, which takes a coalition as an input and returns the value of the coalition. As a result, applying constraint optimization techniques to this problem has been infeasible. However, characteristic functions that appear in practice often can be represented concisely by a set of rules, rather than a single black-box function. Then, we can solve the CSG problem more efficiently by applying constraint optimization techniques to the compact representation directly. We present new formalizations of the CSG problem by utilizing recently developed compact representation schemes for characteristic functions.We first characterize the complexity of the CSG under these representation schemes. In this context, the complexity is driven more by the number of rules rather than by the number of agents. Furthermore, as an initial step towards developing efficient constraint optimization algorithms for solving the CSG problem, we develop mixed integer programming formulations and show that an off-the-shelf optimization package can perform reasonably well, i.e., it can solve instances with a few hundred agents, while the state-of-the-art algorithm (which does not make use of compact representations) can solve instances with up to 27 agents.

Coalition formation is an important capability for automated negotiation among self-interested agents. In order for coalitions to be stable, a key question that must be answered is how the gains from cooperation are to be distributed. Coalitional game theory provides a number of solution concepts for this. However, recent research has revealed that these traditional solution concepts are vulnerable to various manipulations in open anonymous environments such as the Internet. To address this, previous work has developed a solution concept called the anonymity-proof core, which is robust against such manipulations. That work also developed a method for compactly representing the anonymity-proof core. However, the required computational and representational costs are still huge. In this paper, we develop a new solution concept which we call the anonymity-proof Shapley value. We show that the anonymity-proof Shapley value is characterized by certain simple axiomatic conditions, always exists, and is uniquely determined. The computational and representational costs of the anonymity-proof Shapley value are drastically smaller than those of existing anonymity-proof solution concepts.

Coalition formation is a key aspect of automated negotiation among self-interested agents. In order for coalitions to be stable, a key question that must be answered is how the gains from cooperation are to be distributed. Various solution concepts (such as the Shapley value, core, least core, and nucleolus) have been proposed. In this paper, we demonstrate how these concepts are vulnerable to various kinds of manipulations in open anonymous environments such as the Internet. These manipulations include submitting false names (one acting as many), collusion (many acting as one), and the hiding of skills. To address these threats, we introduce a new solution concept called the anonymity-proof core, which is robust to these manipulations. We show that the anonymity-proof core is characterized by certain simple axiomatic conditions. Furthermore, we show that by relaxing these conditions, we obtain a concept called the least anonymity-proof core, which is guaranteed to be non-empty.

Coalition formation is an important capability of automated negotiation among self-interested agents. In order for coalitions to be stable, a key question that must be answered is how the gains from cooperation are to be distributed. Recent research has revealed that traditional solution concepts, such as the Shapley value, core, least core, and nucleolus, are vulnerable to various manipulations in open anonymous environments such as the Internet. These manipulations include submitting false names, collusion, and hiding some skills. To address this, a solution concept called the anonymity-proof core, which is robust against such manipulations, was developed. However, the representation size of the outcome function in the anonymity-proof core (and similar concepts) requires space exponential in the number of agents/skills. This paper proposes a compact representation of the outcome function, given that the characteristic function is represented using a recently introduced compact language that explicitly specifies only coalitions that introduce synergy. This compact representation scheme can successfully express the outcome function in the anonymity-proof core. Furthermore, this paper develops a new solution concept, the anonymity-proof nucleolus, that is also expressible in this compact representation. We show that the anonymity-proof nucleolus always exists, is unique, and is in the anonymity-proof core (if the latter is nonempty). and assigns the same value to symmetric skills.