Michimasa Inaba

Neural Utterance Ranking Model for Conversational Dialogue Systems

In this study, we present our neural utterance ranking (NUR) model, an utterance selection model for conversational dialogue agents. The NUR model ranks candidate utterances with respect to their suitability in relation to a given context using neural networks; in addition, a dialogue system based on the model converses with humans using highly ranked utterances. Specifically, the model processes word sequences in utterances and utterance sequences in context via recurrent neural networks. Experimental results show that the proposed model ranks utterances with higher precision relative to deep learning and other existing methods. Furthermore, we construct a conversational dialogue system based on the proposed method and conduct experiments on human subjects to evaluate performance. The experimental result indicates that our system can offer a response that does not provoke a critical dialogue breakdown with a probability of 92% and a very natural response with a probability of 58%.

Investigation of the effects of nonverbal information on werewolf

Werewolf is one of the popular communication games all over the world. It treats ambiguity of human discussion including the utterances, gestures and facial expressions in a broad sense. In this research, we pay attention to this imperfect information game werewolf. The purpose of the research is to develop an intelligent agent “AI werewolf” which is enabled to naturally play werewolf with human. This paper aims to investigate how behavior contribute to victory of own-side players by using machine learning as a first step. As the results of investigation and analysis of the playing movie, we found that nonverbal information in the game of werewolf has importance to winning or losing the game.

Movement design of a life-like agent for the werewolf game

In this research, we target at the interactive communication game “werewolf” with a subject of research. Werewolf is a popular party game all over the world, and the relevance studies have been advanced in recent years. However, the life-like agent who does werewolf has not been developed. Therefore the purpose of this research is to analyze non-verbal information from movies which play the werewolf with face-to-face communication and to make clear the impression for others by the movements of players in the game. Moreover, we verify whether the life-like agent gives an impression like human in werewolf game by mounting the movements on a life-like agent.

Strategy Selection by Reinforcement Learning for Multi-car Elevator Systems

This paper discusses the group control of elevators for improving efficiency, an efficient control method for multi-car elevator using reinforcement learning is proposed. In the method, the control agent selects the best strategy among four strategies, namely Transportation strategy, Passenger strategy, Zone strategy, and Difference strategy according to traffic flow. The control agent takes the number of total passengers and the distance from the departure floor to the destination floor of a call into account. Through experiments, the performance of the proposed method is shown, the average service time of the proposed method is compared with the average service time obtained for the cases where the car assignment is made by each of the three or four strategies.

Experiments of displaying images to keep the motivation in e-learning

E-learning systems that use computers and the Internet have become popular. E-learning systems have many advantages. However, the users often lose their motivation for learning in the process of studying, and the frequency that they use e-learning systems sometimes decreases. In order to improve or keep their motivation, we add a function that the users are praised or scolded by displaying images of a teacher, actors, and friends. We check to see if the function is effective or not through experiments.

AI Wolf Contest — Development of Game AI Using Collective Intelligence —

In this study, we specify the design of an artificial intelligence (AI) player for a communication game called “Are You a Werewolf?” (AI Wolf). We present the Werewolf game as a standard game problem in the AI field. It is similar to game problems such as Chess, Shogi, Go, and Poker. The Werewolf game is a communication game that requires several AI technologies such as multi-agent coordination, intentional reading, and understanding of the theory of mind. Analyzing and solving the Werewolf game as a standard problem will provide useful results for our research field and its applications. Similar to the RoboCup project, the goal of this project is to determine new themes while creating a communicative AI player that can play the Werewolf game with humans. As an initial step, we designed a platform to develop a game-playing AI for a competition. First, we discuss the essential factors in Werewolf with reference to other studies. We then develop a platform for an AI game competition that uses simplified rules to support the development of AIs that can play Werewolf. The paper reports the process and analysis of the results of the competition.

Constructing a Non-task-oriented Dialogue Agent using Statistical Response Method and Gamification

This paper provides a novel method for building non-task-oriented dialogue agents such as chatbots. The dialogue agent constructed using our method automatically selects a suitable utterance depending on a context from a set of candidate utterances prepared in advance. To realize automatic utterance selection, we rank the candidate utterances in order of suitability by application of a machine learning algorithm. We employed both right and wrong dialogue data to learn relative suitability to rank the utterances. Additionally, we provide a low-cost and quality-assured learning data acquisition environment using crowdsourcing and gamification. The results of an experiment using learning data obtained via the environment demonstrate that the appropriate utterance is ranked on the top in 82.6% of cases and within the top 3 at 95.0% of cases. Results show that using context information that is not used in most existing agents is necessary for appropriate responses.

Constructing a Human-like agent for the Werewolf Game using a psychological model based multiple perspectives

In this paper, we focus on the Werewolf Game. The Werewolf Game is an advanced communication-game in which winning or losing is directly linked to ones success or failure in communication. Therefore, we expect exponential developments in artificial intelligence by studying the Werewolf Game. In this current study, we propose a psychological model that considers multiple perspectives to model the play of a human such as inferring the intention of the other side. As one of the psychological models, we constructed a “ones self model” that models the role of others as viewed from their own viewpoint. In addition, to determine whether ones opinion is reliable after inferring others intentions, we also constructed an “others model” that models the role of others as viewed from their viewpoints. Combining these models, we showed through experimentation that a combined approach achieved better results, i.e., higher win percentages.

Machine-Learned Ranking Based Non-Task-Oriented Dialogue Agent Using Twitter Data

This paper describes a method for developing a non-task-oriented dialogue agent (also called chat-oriented or conversational dialogue agents) that can cover broad range of topics. Our method extracts a topic from a users utterance and acquires candidate utterances that contain the topic from Twitter. Our agent selects a suitable utterance for dialogue context from candidates using machine-learned ranking method. Results of an experiment demonstrate that a dialogue agent based on the proposed method can conduct more natural and enjoyable conversation compared to other dialogue agents.

Experiments Assessing Learning of Agent Behavior using Genetic Programming with Multiple Trees

This paper proposes Genetic Programming(GP) with control nodes using the conditional probability and the island model for efficient learning of agent behavior. In the methods, each individual has a chromosome representing agent behavior as several trees. In GP using the conditional probability, individuals with high fitness values are used to produce conditional probability tables to generate individuals in the next generation. In GP using the island model, the population is divided into two islands of individuals: one island keeps diversity of individuals and the other puts emphasis on the accuracy of the solution. The methods are applied to a garbage collection problem and Santa Fe Trail problem. The proposed methods are compared with traditional GP, GP with control nodes, and Genetic Network Programming(GNP) with control nodes. Experimental results show that the proposed methods are efficient.