Turn-taking prediction models output probabilities of turn shifts, yet they are typically evaluated by thresholding these probabilities into binary decisions and comparing them against corpus-observed labels. This practice implicitly treats corpus-observed turn shifts as definitive ground truth, even though under inherent turn-taking ambiguity they reflect one realized interactional outcome among multiple plausible outcomes, rather than a uniquely correct binary label. We argue that binary evaluation is a practical simplification rather than a theoretical necessity. Instead, predicted probabilities should be evaluated at the distributional level without being reduced to binary decisions. To this end, we propose a distribution-based evaluation framework that compares model output distributions with reference distributions and measures their divergence using the Wasserstein distance. We further show how discrepancies between model predictions and corpus-observed turn shifts can be used as a basis for training-data refinement. Experiments on Japanese conversational data, using linguistic information alone, showed that the proposed refinement reduced distributional divergence, indicating better alignment between predicted probabilities and the reference distributions. The refinement also improved balanced accuracy in a supplementary binary evaluation.

Visually impaired individuals often face challenges related to physical inactivity, stemming from limited accessibility to sports activities. While assistive technologies and adaptive sports have made progress, high-intensity contact sports remain largely inaccessible. This paper presents an approach to improve the accessibility of kickboxing for visually impaired individuals. We prioritize simple, immediate solutions that maximize the use of residual vision, diverging from conventional assistive technologies that rely on auditory or tactile feedback. Employing a participatory design methodology, we involved visually impaired individuals throughout the development process of specialized kickboxing equipment. This approach enabled us to address user-specific needs and challenges directly. Our research contributes to expanding sports participation opportunities for the visually impaired and fostering a more inclusive sports environment. We detail the development process of our adaptive kickboxing equipment and discuss its effectiveness in enhancing the sports experience for visually impaired individuals.

As increasing volumes of data have become available globally, there are growing expectations for the creation of value through the exchange of data across diverse fields and the enhancement of the value of existing services. Consequently, there is an increasing demand for markets and platforms to facilitate data trading and exchanges. However, unlike other well-established business ecosystems such as existing financial markets and service ecosystems, the overall structure and characteristics of the data exchange ecosystem remain largely unexplored. This study focuses on users who handle data, and aims to elucidate their behaviors and identify influential users on the platforms, thereby deepening our understanding of the data ecosystem. We conducted network analysis, community extraction, and time-series analysis of the behavior of Kaggle users. Our results indicate that the network of Kaggle users exhibits a scale-free structure similar to typical social networks and web links. This suggests that information dissemination among Kaggle users is likely to occur through a small number of hubs. Furthermore, we performed community extraction, analyzed the behavior of users in each community, and discovered that each community possessed distinct user behavior characteristics. Subsequently, we analyzed the number of influential users with numerous followers or high degree centrality and examined the time-series changes in the number of actions taken by these users. This analysis enabled us to identify the behaviors and formation patterns of those who played a central role in the data platform. The approach and findings of our study provide important considerations for stakeholders and potential participants in all the markets and platforms that handle data.

Character of spoken dialogue systems is important not only for giving a positive impression of the system but also for gaining rapport from users. We have proposed a character expression model for spoken dialogue systems. The model expresses three character traits (extroversion, emotional instability, and politeness) of spoken dialogue systems by controlling spoken dialogue behaviors: utterance amount, backchannel, filler, and switching pause length. One major problem in training this model is that it is costly and time-consuming to collect many pair data of character traits and behaviors. To address this problem, semi-supervised learning is proposed based on a variational auto-encoder that exploits both the limited amount of labeled pair data and unlabeled corpus data. It was confirmed that the proposed model can express given characters more accurately than a baseline model with only supervised learning. We also implemented the character expression model in a spoken dialogue system for an autonomous android robot, and then conducted a subjective experiment with 75 university students to confirm the effectiveness of the character expression for specific dialogue scenarios. The results showed that expressing a character in accordance with the dialogue task by the proposed model improves the user’s impression of the appropriateness in formal dialogue such as job interview.

This paper presents a monocular projector-camera (procam) system using modular architecture based on relay optics. Conventional coaxial procam systems cannot support (1) online changes to lens settings (zoom and focus) and (2) wide-angle projection mapping. We develop design guidelines for a proposed procam system that would solve these restrictions and address the proposed system's unique technical issue of crosstalk between the camera and projector pixels. We conducted experiments using prototypes to validate the feasibility of the proposed framework. First, we confirmed that the proposed crosstalk reduction technique worked well. Second, we found our technique could achieve correct alignment of a projected image onto a moving surface while changing the zoom and focus of the objective lens. The monocular procam system also achieved radiometric compensation where a surface texture was visually concealed by pixel-wise control of a projection color based on the captured results of offline color pattern projections. Finally, we demonstrated the high expandability of our modular architecture, through the creation of a high dynamic range projection.

In recent years, spoken dialogue systems have been used in job interviews where an applicant talks to a system that asks pre-defined questions, called on-demand and self-paced job interviews.We propose a simultaneous job interview system, where one interviewer can conduct one-on-one interviews with multiple applicants simultaneously by cooperating with multiple autonomous interview dialogue systems.However, it is challenging for interviewers to monitor and understand all parallel interviews done by the autonomous system simultaneously.To address this issue, we implement two automatic dialogue understanding functions: (1) response evaluation of each applicant's responses and (2) keyword extraction for a summary of the responses.In this system, interviewers can intervene in a dialogue session when needed and smoothly ask a proper question that elaborates the interview.We have conducted a pilot experiment where an interviewer conducted simultaneous job interviews with three candidates.

It is important for photographers to have the best possible lighting configuration at the time of shooting; otherwise, they need post-processing on images, which may cause artifacts and deterioration. Thus, photographers often struggle to find the best possible lighting configuration by manipulating lighting devices, including light sources and modifiers, in a trial-and-error manner. In this paper, we propose a novel computational framework to support photographers. This framework assumes that every lighting device is programmable; that is, its adjustable parameters (e.g., orientation, intensity, and color temperature) can be set using a program. Using our framework, photographers do not need to learn how the parameter values affect the resulting lighting, and even do not need to determine the strategy of the trial-and-error process; instead, photographers need only concentrate on evaluating which lighting configuration is more desirable among options suggested by the system. The framework is enabled by our novel photographer-in-the-loop Bayesian optimization, which is sample-efficient (i.e., the number of required evaluation steps is small) and which can also be guided by providing a rough painting of the desired lighting configuration if any. We demonstrate how the framework works in both simulated virtual environments and a physical environment, suggesting that it could find pleasing lighting configurations quickly in around 10 iterations. Our user study suggests that the framework enables the photographer to concentrate on the look of captured images rather than the parameters, compared with the traditional manual lighting workflow.

Real-time captioning is a useful technique for deaf and hard-of-hearing (DHH) people to talk to hearing people. With the improvement in device performance and the accuracy of automatic speech recognition (ASR), real-time captioning is becoming an important tool for helping DHH people in their daily lives. To realize higher-quality communication and overcome the limitations of mobile and augmented-reality devices, real-time captioning that can be used comfortably while maintaining nonverbal communication and preventing incorrect recognition is required. Therefore, we propose a real-time captioning system that uses a transparent display. In this system, the captions are presented on both sides of the display to address the problem of incorrect ASR results, and the highly transparent display makes it possible to see both the body language and the captions.

We demonstrate the moderating abilities of a multi-party attentive listening robot system when multiple people are speaking in turns.Our conventional one-on-one attentive listening system generates listener responses such as backchannels, repeats, elaborating questions, and assessments.In this paper, additional robot responses that stimulate a listening user (side participant) to become more involved in the dialogue are proposed.The additional responses elicit assessments and questions from the side participant, making the dialogue more empathetic and lively.

A job interview is a domain that takes advantage of an android robot's human-like appearance and behaviors. In this work, our goal is to implement a system in which an android plays the role of an interviewer so that users may practice for a real job interview. Our proposed system generates elaborate follow-up questions based on responses from the interviewee. We conducted an interactive experiment to compare the proposed system against a baseline system that asked only fixed-form questions. We found that this system was significantly better than the baseline system with respect to the impression of the interview and the quality of the questions, and that the presence of the android interviewer was enhanced by the follow-up questions. We also found a similar result when using a virtual agent interviewer, except that presence was not enhanced.

We describe an attentive listening system for the autonomous android robot ERICA.The proposed system generates several types of listener responses: backchannels, repeats, elaborating questions, assessments, generic sentimental responses, and generic responses.In this paper, we report a subjective experiment with 20 elderly people.First, we evaluated each system utterance excluding backchannels and generic responses, in an offline manner.It was found that most of the system utterances were linguistically appropriate, and they elicited positive reactions from the subjects.Furthermore, 58.2% of the responses were acknowledged as being appropriate listener responses.We also compared the proposed system with a WOZ system where a human operator was operating the robot.From the subjective evaluation, the proposed system achieved comparable scores in basic skills of attentive listening such as encouragement to talk, focused on the talk, and actively listening.It was also found that there is still a gap between the system and the WOZ for more sophisticated skills such as dialogue understanding, showing interest, and empathy towards the user.

We investigate the influence of channel estimation error on the achievable system-level throughput performance of non-orthogonal access with successive interference cancellation (SIC) in the cellular downlink. The channel estimation error in non-orthogonal access causes residual interference in the SIC process, which decreases the achievable user throughput. Furthermore, the channel estimation error causes error in the transmission rate control for the respective users, which may result in decoding error at not only the destination user terminal but also other user terminals for SIC. However, we show that by using a simple transmission rate back-off algorithm, the impact of the channel estimation error is effectively alleviated and non-orthogonal access with SIC achieves clear average and cell-edge user throughput gains relative to orthogonal access similar to the case with perfect channel estimation.

Recently, a number of studies on walking support systems for users with visual impairment have been reported. Most proposals, however, would require significant infrastructure investment and the use of dedicated devices. Auditory route maps that guide users to their destination have been proposed as a cost-reducing solution, but existing maps cannot be adapted to new routes according to individual preferences. This paper therefore proposes a walking support system consisting of three components: 1) an optimal routing method that selects the best route for an individual; 2) camera-based landmark matching to prevent users from losing their way; and 3) a remote assistance system to guide users back to the route if necessary. This paper also proposes methods for evaluating these system components.

This paper describes a method for a heartbeat and respiratory rate monitoring system using air pressure sensors and ultrasonic oscillosensor. By using these sensors, we propose a detection method of the state of human and an extraction method of heartbeat and respiratory rate in bed by fuzzy logic. Our method was examined on four healthy volunteers. We successfully detected the state of human and extracted heartbeat and respiratory signals. In our method, fuzzy logic plays a primary role in the detection of the state and extraction of heartbeat and respiratory signals. An experiment on four healthy volunteers was done. Consequently, our proposed method noninvasively and successfully detects the state of human and extracted heartbeat and respiratory rate in the bed by using the unconstrained sensors.

This paper proposes functional assessment system of autonomic nervous system by the heart rate variability using an air cushion sensor. The air cushion sensor can unconstraintly detect vital information by sitting down on the sensor. We perform functional assessment of autonomic nervous system by heart rate variability obtained by the system. We built the real time display system for visualizing the autonomic nervous system functions. In this system, we employ fuzzy membership functions with dynamic parameter to detect RR intervals. The experimental results show that we detect RR intervals with the correlation coefficient of 0.846 with comparison to that of electrocardiograph. Then, the errors of the HF (index of parasympathetic system) and the LF/HF (index of sympathetic system) are 18.34% and 16.99%, respectively.

In order to realize heart rate monitoring on a bed, there are mainly two types of approaches similar to other signal processing applications: frequency domain analysis and time-series domain analysis. In frequency domain analysis, FFT is widely used to extract heart rate from obtained signals. Since FFT assumes constant frequency, it cannot be used for extracting microscopic variability of heart rate. In time-series domain analysis, pattern matching based on autocorrelation is commonly used. The method is not only advanced in sensitivity to heart rate variability, but it is also sensitive to unexpected noise. In response to these problems, heart rate monitoring technology is proposed by using air pressure sensor. In this paper, a heart rate estimation algorithm employing fuzzy logic is proposed and effectiveness of fuzzy logic applied to biomedical sensing is discussed. The experiments were conducted to validate the effectiveness of the proposed technology by comparing it with other methods such as pattern matching based on autocorrelation.

This paper proposes a functional assessment system of autonomic nervous system using an air pressure sensor. The air pressure sensor can unconstraintly detect vital information by placing it under the mattress in bed. We perform functional assessment of autonomic nervous system by heart rate variability obtained by the system. In this system, we employ fuzzy membership functions with dynamic parameter to detect RR intervals. The experimental results show that we detect RR intervals with the correlation coefficient of 0.851 with comparison to that of electrocardiograph. Then the errors of the HF (index of parasympathetic system) and the LF/HF (index of sympathetic system) are 11.98% and 22.18%, respectively.