Automatic portfolio optimization (APO) is the process of automatically optimizing the allocation of assets in an investment portfolio through algorithms and models. Empirical evidence suggests that stock prices are likely to follow the mean reversion theory. Although existing mean reversion strategies for APO have been shown to achieve good empirical performance across many real-world datasets, they tend to construct single-stock portfolios, which results in high risk. In this paper, we propose a risk diversification version of Online Moving Average Reversion (OLMAR) [16], one of the major mean reversion strategies. The parameter k of our proposed method OLMAR-k controls the search space for optimization to guarantee the existence of solutions and prevent from solutions close to uniform. According to experimental results in resent datasets, OLMAR-k outperforms other state-of-the-art APO methods in terms of popular metrics for return and risk.

This study presents the development and evaluation of a critical paper reading support system named "Forest/CR." The system aims to enhance learners' critical reading skills by facilitating the formation of reading strategies. It achieves this by recognizing questions that learners were unable to generate independently and by allowing them to compare their interpretations with those of others. The system features functions such as paper content organization, comparison of reading content, and comprehensive evaluation support. An experimental study involving third-year undergraduate students demonstrated that the system effectively broadened students' reading activities, promoted the generation of critical questions, and supported the learning of new reading strategies.

In this study, we adopt the Thinking Like a Historian Approach (TLH), which systematizes the historical inquiry process of historians as a learning model. By expanding Linked Open Data (LOD) using a large language model (LLM) based on an ontology that defines TLH concepts, we construct a white-box-oriented learning resource that ensures comprehensiveness, reproducibility, learning controllability, and maintainability. The historical learning support system developed based on this learning resource has adaptive question generation function and historical information visualization function in an open-ended historical inquiry space.

Research meetings are critical forums for sharing and evolving ideas, where the depth of the discussion can significantly enhance the research content. Effective meetings require not just a recapitulation of the discussions but a thorough examination of the thoughts and decisions preceding the meeting. In this paper, we propose a support system that promotes metacognitive reflection by aligning pre-meeting preparations with post-discussion feedback, thus enhancing the overall research process and outcomes.

Language learning resources, such as dialogue scenarios, are important for supporting second language (L2) learning by engaging learners in simulated conversations or role-plays. However, there are often barriers to the wider dissemination of these resources, including the lack of a standard format for distributing them, the difficulties of maintaining and updating them over time, and the need for effective means of tracking their use and impact. This paper presents an integrated framework for fostering the wider dissemination of L2 learning resources, with a focus on dialogue scenarios. The framework is designed to address the above challenges by providing a novel approach for creating and distributing dialogue scenarios, as well as a set of guidelines to ensure that they are pedagogically sound and compatible with learning management systems (LMS). The paper also discusses the potential benefits of the framework for L2 learners and educators and outlines future directions for the project.

To form and maintain smooth interpersonal relationships in various situations in social life, the ability to appropriately understand the mental states of others, such as their intentions and feelings, is required. However, it is difficult to infer the unobservable mental states of others, and it is also difficult to learn from others how to perform inferring activities that are not externalized. In this study, we focus on “dramatic scripts” as a learning material that contributes to the training of the ability to understand others and propose a learning support system that aims to improve the ability to understand others and to cultivate an attitude of understanding others through inference activities about the emotions and conflicts of characters.

Hand-eye coordination (HEC) is an important ability in a variety of sports, as it involves the ability to quickly and accurately control motion using visual information. This ability is extensively involved in badminton smash reception. In existing studies on HEC evaluation, this ability was evaluated according to information collected after the motion had been performed. Accordingly, it is impossible to determine the specific phase of perception or motion that causes problems in HEC. The purpose of this study was to develop a system for identifying the abilities of perception involved in HEC as it relates to badminton smash it is considered as badminton's most effective shot. Two systems were developed to identify HEC perceptual abilities, one focused on perception in a static state (i) and the other focused on perception-to-motion comprehensiveness (ii). Next, two types of tests were conducted, one involving (i) and the other using (ii). The tests were randomly administered for a total of 18 participants in three groups: six people who had previously played badminton, six people who had played ball sports other than badminton, and six people who were inexperienced at ball sports. Participants were healthy adults. The tests aimed to identify factors of perceptual ability by comparing them across systems. The results showed no significant difference between experienced badminton players and others according to the conducted tests, where test (i) was used for measuring the distance between the shuttlecock and gaze. A significant difference between the group of experienced badminton players and the other two groups was, however, observed in the test using (ii). The results suggested the ability to follow the shuttlecock with the eyes to be a perceptual ability related to HEC in badminton smash receptions. The results of this study can help evaluate perceptual ability during smash receptions.

Quantitatively evaluating the psychological and perceptual effects of objects is an important issue, but is difficult. In cognitive studies, the psychological potential field (PPF), which represents psychological intensities in vision and can be calculated by applying computational algorithms to digital images, may help with this issue. Although studies have reported using the PPF to evaluate psychological effects, such as impressions, detailed investigations on how the PPF represents psychological perception and its limitations have not yet been performed. Another relevant tool is the fixation map, which visualizes human eye fixations; this map is generated from actual measurements acquired by eye-tracking and does not represent psychological effects directly. Although the PPF and the fixation map are based on visual imaging, they have never been compared. In this paper, we do so for the first time, using psychological and perceptual properties of line-drawing images. The results demonstrate the difference between these methods, including their representation of different properties with respect to visual perception. Moreover, the similarity between the two methods highlights the possibility of assessing perceptual phenomena such as categorization and cognition of objects based on human vision.

In this research, we focused on research execution activities as a field of ill-defined problem solving and developed a mechanism to grasp learners’ thought processes based on interpretation rules and a generic ontology of research activities. The idea behind this mechanism is to prompt learners’ awareness of their own thinking processes so as to foster their metacognitive skills. In this paper, we discuss the features of a support system that we developed to achieve this goal.

Although writing activities present a good opportunity to develop metacognitive skills, it is difficult for novice learners to be aware of logical inconsistencies or a lack of perspective, even in their own documents. To promote such learners’ metacognitive activities, a novel eye-movement visualization system is proposed that activates learners’ metacognitive inference activities by expanding a previously proposed framework. The system intuitively visualizes all reading processes of the expert, who critically reads the learner’s document, as a timeline graph. In addition, it has several support functions that provide information on the expert’s characteristic reading behaviors and the logical structure of documents. The initial experiment showed that the proposed graph visualization system has the potential to activate learners’ metacognitive inference activities, even after verifying the previously proposed eye-movement visualization tools. It was also confirmed that learners can generate the correction points by themselves, even in documents that they had already checked critically before submission.

In this paper, we discussed a shopping activity from the viewpoint of preventing dementia. The activity is one of the IADL(Instrumental Activities of Daily Living) activities and it is more natural, essential and intellectual than other IADL activities. In the related works, the activity has been regarded as the activity which stimulates human cognitive functions and several training methods including therapy or rehabilitation have been proposed. We conducted the observation of shopping activity and analyzed the intellectual aspect of the activity. In particular, we focused on the words about “planning” and “logicality” because they often appeared in the protocol of participants who often go shopping by themselves and are familiar with shopping activities.

When holding creative discussions with other people, it is important not only that the speaker accurately understand what he/she wants to convey but also that the speaker express knowledge in a way that is understandable to the listener. “Reflective thinking” (knowledge restructuring) in which one considers what one wants to say and “verbalization of reasoning” driven by thinking that is conscious of other people are mutually interactive and deepen understanding of what one is trying to say. This study focuses on academic research meetings as a forum conducive to creative discussions and on the preparation of research materials (research documentation) as the product of the verbalization of reasoning that takes knowledge transmission into account. It proposes, in particular, a self-conversation support system that supports the iteration of these two activities.

In data-parallel computing that uses a graphic processing unit (GPU), processing of large data requires that multiple GPUs be used in the computer to increase its execution performance. Increasing processing performance by using multiple computing resources has been enabled by the development of computing frameworks based on a directed acyclic graph (DAG). However, their performance degrades in out-of-core processing, which often occurs in processing of large data on GPUs with limited memory capacity. The GPU data input/output (I/O) for data swapping between host memory and GPU memory during the execution of a user DAG is usually a performance bottleneck. A computing framework called "Victream" is proposed to overcome this drawback. It uses a novel scheduler that involves two methods to minimize the total amount of GPU data I/O of data swapping. First, it performs locality-aware scheduling. When it schedules a task, it selects one that requires the minimum amount of data swapping and reuses as much of the data residing in GPU memory as possible. Second, it extends the locality-aware scheduling so that GPUs can execute data prefetching. Prefetching data that are swapped out from a GPU enables efficient use of bottleneck GPU I/O resources. To prefetch the input data of future tasks, it is required to determine the schedule of future tasks. Victream's scheduler (hereafter, the Victream scheduler) extends the locality-aware scheduling so that it can schedule future tasks to enable data prefetching that is executed in the way that minimizes the amount of data I/O of data swapping. Evaluation of a Victream prototype showed that the performance of Victream is better than that of conventional frameworks by up to 117%.

We have collected several size of data during experiments of the online shopping game. Where we could collect various information about the participants' behaviour. For instance, how they moved in the supermarket and which things (products) they checked, returned or bought. We have an intention to use those data to determine how human beings think during shopping. In addition, we think we can estimate the intelligence level (cognitive function) from the data. In order to do so, it will be necessary to deal with the collected data correctly. In this paper, from the viewpoint of granular computing, we will discuss how to collect data, what data should be collected, and how to store data in the data base. In addition, we will discuss how to estimate the cognitive function.

In Japan, as the super-aged society, it has recently been important to take care of housebound and elderly persons. It is important to encourage elderly persons to visit ordinary places in their daily life. There are several researches to focus on shopping activities as a chance of going outdoors. In this paper, we will survey several issues of housebound and several approaches against them. Especially we will focus on shopping activities in order to encourage elderly persons to visit ordinary places with less physical and psychological stress. In addition, we will consider an intelligent aspect of shopping activities from the viewpoint of preventing housebound and dementia.

Networked devices in an Internet-of-Things (IoT) environment generally need to be controlled or organized by a central controlling host. These IoT devices are implemented together with a small computer that consists of a central processing unit (CPU), a memory, and input/output (I/O) devices to communicate with this host. The small computer uses its own software network stack for communication. This paper reports our study of another finer-grained IoT in which an IoT device is not implemented with a computer but just an I/O device. The I/O device works as the local device of a remote controlling host and the IoT device is controlled by data communication with the controlling host through the I/O device. Such an IoT system simplifies the internal configuration of an IoT device node and can lower its implementation costs. However, previous methods have required a special hardware interface to their controlling host to connect to a remote I/O device that is implemented together with an IoT device. This paper proposes a novel software virtualization architecture for a controlling host to connect to I/O devices distributed in an Ethernet. It virtualizes a remote I/O device that communicates through the standard network interface card (NIC) of a host as an I/O device that is contained in the I/O device tree of the host. Various kinds of computers ranging from hand-held to rackmount ones are enabled to control remote IoT devices with the installation of the proposed software. In addition, because virtualization is performed in the PCI Express (PCIe) layer, any complying I/O device is remotely interconnected and is driven by its device driver without modifications. The feasibility of the method was studied by evaluating the implemented prototype using a NIC and a storage controller.

Extending input/output (I/O) bus networks over wireless Ethernet is expected to provide an effective way of achieving communication between multiple hosts and I/O devices. However, because connections on wireless Ethernet are lossy and unreliable, retransmission of I/O packets increases transmission latency that exceeds the limits of the I/O bus protocol. In this paper, we propose an Ethernet transport mechanism that adopts rateless codes to provide reliable transmission without having to retransmit packets. We further devised a queuing model to set the operating parameters of rateless codes to bound the probability of end-to-end latency that exceeds the defined limit value. Our results obtained through simulations and an FPGA implementation show that the proposed method decreased the transmission latency of I/O packets and enabled it to meet the requirements of the I/O bus protocol.

It is difficult to capture a person’s thinking processes because of the invisible, shapeless, complex structure of thought. In order to approach implicit thinking processes, we propose a framework for capturing a portion of meta-level thinking processes from gaze behaviors. Here, we focus on the meta-level thinking processes in the dissolution of belief conflicts, and introduce the concept of semi-metacognitive-level thinking, which allows us to make assumptions about the monitoring/control cycle based on gaze behaviors toward externalization objects of base-level thinking. In this paper, we discuss a framework for capturing metacognitive thinking processes from gaze behaviors and propose an ontology to embody the concept of the framework.

The purpose of this research is to support learners in self-directed learning on the Internet using automatically generated support using the current state of the semantic web. The main issue of creating meaningful content-dependent questions automatically is that it requires the machine to understand the concepts in the learning domain. The originality of this work is that it uses Linked Open Data (LOD) to enable meaningful content-dependent support in open learning space. Learners are supported by a learning environment, the Semantically Enhanced Open Learning Space (SOLS). Learners use the system to build a concept map representing their knowledge. SOLS support learners following the principle of inquiry-based learning. Learners that request help are provided with automatically generated questions that give them learning objectives. To verify whether the current system can support learners with fully automatically generated support, we evaluated the system. The results showed that LOD based support was feasible. Learners felt that the support provided was useful and helped them learn. The question support succeeded in improving the development of learners’ historical considerations and deep historical thinking skills. In addition, the engagement and interest in history of learners was improved by the questions. The results are meaningful because they show that LOD based support can be a viable tool to support learners in open learning space and that the question support has potential to support learners during a long time study.

This paper focuses on the computational analysis of the individual communication skills of participants in a group. The computational analysis was conducted using three novel aspects to tackle the problem. First, we extracted features from dialogue (dialog) act labels capturing how each participant communicates with the others. Second, the communication skills of each participant were assessed by 21 external raters with experience in human resource management to obtain reliable skill scores for each of the participants. Third, we used the MATRICS corpus, which includes three types of group discussion datasets to analyze the influence of situational variability regarding to the discussion types. We developed a regression model to infer the score for communication skill using multimodal features including linguistic and nonverbal features: prosodic, speaking turn, and head activity. The experimental results show that the multimodal fusing model with feature selection achieved the best accuracy, 0.74 in R2 of the communication skill. A feature analysis of the models revealed the task-dependent and task-independent features to contribute to the prediction performance.

Input/output (I/O) devices such as a graphics processing unit and a solid-state drive are inserted into I/O slots of a host in data center platforms. With this sort of configuration the I/O devices are used exclusively by the host with resultant inefficient resource usage. In addition, the maximum number of I/O devices that can be assigned to each host is limited by the number of its I/O slots. This paper proposes a method to solve the resource usage inefficiency and assignment limitation by disaggregating I/O devices from hosts. An I/O device accommodated in a device pool is interconnected with multiple hosts by a standard Ethernet and can be flexibly assigned to one of them. Introducing assignment flexibility of I/O devices among hosts improves the resource usage efficiency and enables as many I/O devices as desired being assigned to each host. The authors further propose a method that enables an I/O device to be simultaneously shared among multiple hosts and multiple I/O packets to be aggregated into an Ethernet frame to suppress the performance overhead of the device disaggregation. Evaluation results showed that I/O devices that are compliant to PCI Express are shared with the overhead up to 27 percent.

With the goal of assessing participant attitudes and group activities in collaborative learning, this study presents models of participation styles based on co-occurrence patterns of nonverbal behaviors between conversational participants. First, we collected conversations among groups of three people in a collaborative learning situation, wherein each participant had a digital pen and wore a glasses-type eye tracker. We then divided the collected multimodal data into 0.1-second intervals. The discretized data were applied to an unsupervised method to find co-occurrence behavioral patterns. As a result, we discovered 122 multimodal behavioral motifs from more than 3,000 possible combinations of behaviors by three participants. Using the multimodal behavioral motifs as predictor variables, we created regression models for assessing participation styles. The multiple correlation coefficients ranged from 0.74 to 0.84, indicating a good fit between the models and the data. A correlation analysis also enabled us to identify a smaller set of behavioral motifs (fewer than 30) that are statistically significant as predictors of participation styles. These results show that automatically discovered combinations of multiple kinds of nonverbal information with high co-occurrence frequencies observed between multiple participants as well as for a single participant are useful in characterizing the participant's attitudes towards the conversation.

We propose an embodied conversational agent based on the willingness to communicate (WTC) model in L2 to help increase WTC in the context of English as a Foreign Language (EFL) by providing users with various daily conversation contexts. To simulate realistic and efficient conversations, we adopted a semantic approach in the response generation and created a system with flexible and adaptable domain knowledge, user's intent detection, and mixed-initiative conversation strategy. Our evaluation of the proposed system demonstrated its potential to increase WTC in the EFL context.

Many people would like to live long with healthy body and mind. If you help people live well, you should take psychological stress into account. In this paper, we will support people in the situation of shopping activities. We propose the Shopping Chat Method and let a participant perform shopping along this method. Finally, we consider how to extract “novel” information about shopping activities from participants.

Gaze is an important nonverbal feedback signal in multiparty face-to-face conversations. To build a robot that can convey the appropriate attentional behavior in human-robot multiparty conversations, this paper analyzes human attentional behaviors in multiparty conversations, and establishes gaze-transition models for speakers, addressees, and side participants. Further, the model was implemented in a humanoid robot that can control its gaze.

Group discussions are used widely when generating new ideas and forming decisions as a group. Therefore, it is assumed that giving social influence to other members through facilitating the discussion is an important part of discussion skill. This study focuses on influential statements that affect discussion flow and highly related to facilitation, and aims to establish a model that predicts influential statements in group discussions. First, we collected a multimodal corpus using different group discussion tasks; in-basket and case-study. Based on schemes for analyzing arguments, each utterance was annotated as being influential or not. Then, we created classification models for predicting influential utterances using prosodic features as well as attention and head motion information from the speaker and other members of the group. In our model evaluation, we discovered that the assessment of each participant in terms of discussion facilitation skills by experienced observers correlated highly to the number of influential utterances by a given participant. This suggests that the proposed model can predict influential statements with considerable accuracy, and the prediction results can be a good predictor of facilitators in group discussions.

To understand collaborative learning interaction, it is important to analyze not only argument processes based on verbal information but also non-verbal interaction. In order to analyze learning situations in collaborative learning, our previous work proposed an estimation method for learning attitudes based on participants’ non-verbal features. Because the method used limited features, this research enhances the method of the participants’ collaborative attitudes by analyzing non-verbal features in detail. The model also considers participants’ knowledge of their learning subject in the analysis. The estimation model detects three levels of the participants’ collaborative attitudes based on multinomial logistic regression analysis. The results of the analysis show that the speech interval feature, in particular, affects the participants’ collaborative attitudes. In addition, the results indicate that speakers with knowledge of the learning subject receive more attention from participants with insufficient knowledge. The results of the model evaluation find that the f-measure for classifying the participants’ collaborative attitudes is 0.569; for participants with knowledge, the f-measure is 0.647.

In collaborative learning, a discussion sometimes comes to a standstill because no comments have been made or no conflicting opinions have been offered. Activation remarks, which might solve such situations using past discussions, may be effective for similar situations. Our research group proposed a framework that applies the activation remarks of past discussions that were stored in a database to ongoing discussions based on case-based reasoning (CBR). To utilize past activation remarks, they should be modified to fit the current situations. We introduce a method for modifying selected activation remarks that contain situation-dependent words about the current discussion especially on remarks that contain words about the remaining time or topics. Our method grasps the time and the topics of ongoing discussions and creates activation remarks by replacing situation dependent words to words of ongoing discussion. Our experimental result showed that our proposed method created valid and effective activation remarks when appropriate past remarks were selected from the only one case-base.

In conversational agents with multiparty communication functionality, a system needs to be able to identify the addressee for the current floor and respond to the user when the utterance is addressed to the agent. This study proposes some addressee identification models based on speech and gaze information, and tests whether the models can be applied to different proxemics. We build an addressee identification mechanism by implementing the models and incorporate it into a fully autonomous multiparty conversational agent. The system identifies the addressee from online multimodal data and uses this information in language understanding and dialogue management. Finally, an evaluation experiment shows that the proposed addressee identification mechanism works well in a real-time system, with an F-measure for addressee estimation of 0.8 for agent-addressed utterances. We also found that our system more successfully avoided disturbing the conversation by mistakenly taking a turn when the agent is not addressed.

In collaborative learning, participants interact with other members by exchanging not only verbal information but also nonverbal information, such as looking at other participants and note taking, which plays an important role in facilitating effective learning. In order to utilize such non-verbal information to analyze collaborative learning interaction, we conducted experiments to collect non-verbal information (gaze direction, speech intervals, and writing actions of participants) using multimodal measurement devices. In this paper, we propose a visualization method of collaborative learning based on a multimodal corpus of collaborative learning. The method intuitively visualizes the learning interaction with respect to collaborative and learning attitudes of each participant. We introduce the prototype system and show an example of the interface.

In collaborative learning in virtual learning environments, because of the restrictions of the communication means, grasping the learning situations of others is difficult. Learning activities taken in individual spaces, such as taking notes or investigating unknown topics, are especially important for grasping the understanding situations of others. In this paper, we propose an awareness support interface that provides actions in individual spaces as awareness information. By noticing other learners’ individual activities, smooth and effective communication is promoted. Learners do not need the awareness information of all other learners all the time. They may see the actions of others who may provide effective information for them. Effective learners can be detected by the roles in the discussion. Therefore, in this paper, we propose a method for estimating the roles of learners in discussions. Then based on the estimated roles, the learner’s focus for acquiring actions in individual spaces is detected.

In collaborative learning, learners discuss with other learners by exchanging their utterances. In order to represent effective utterances for knowledge acquisition timely, we propose a detection method of focusing utterances which are useful for understanding about their discussion. Focusing utterances are detected according to utterance types and understanding knowledge of learners. Focusing utterances are represented emphatically in the interface so as to distinguish them from normal utterances. Experimental result showed that our method could detect more than 60% of focusing utterances correctly.