Roberto S. Legaspi

• Doctor of Philosophy in Engineering
• Core Research Scientist at KDDI Research Inc.

Multivariate long-term time series forecasting (MLTSF), applicable across various domains, has gained increasing research attention. Channel-independent (CI) models, including Linear and Transformer-based architectures, have recently achieved state-of-the-art (SOTA) performance for MLTSF. Notably, Linear models can deliver satisfactory forecasting...

The Sense of Agency (SoA) refers to the individual’s perception of control over actions and their subsequent impact on the external environment. SoA encompasses multiple dimensions, such as implicit/local and explicit/general, which can be quantitatively assessed through cognitive tasks and psychometric questionnaires, respectively. The explicit an...

Linear causal models are important tools for modeling causal dependencies and yet in practice, only a subset of the variables can be observed. In this paper, we examine the parameter identifiability of these models by investigating whether the edge coefficients can be recovered given the causal structure and partially observed data. Our setting is...

Persuasion plays a pivotal role in a wide range of applications from health intervention to the promotion of social good. Persuasive chatbots can accelerate the positive effects of persuasion in such applications. Existing methods rely on fine-tuning persuasive chatbots with task-specific training data which is costly, if not infeasible, to collect...

We propose a framework for predicting sensor event sequences (SES) in smart homes, which can proactively support residents’ activities and alert them if activities are not completed as intended. We leverage ongoing activity recognition to enhance the prediction performance, employing a GPT2-based model typically used for sentence generation. We hyp...

Customizing persuasive conversations related to the outcome of interest for specific users achieves better persuasion results. However, existing persuasive conversation systems rely on persuasive strategies and encounter challenges in dynamically adjusting dialogues to suit the evolving states of individual users during interactions. This limitatio...

To investigate how the interaction effects of dispositional and regional factors affect the susceptibility to persuasive strategies, we conducted a large-scale online survey (n = 3,116 in eight major cities in Japan). Our results indicate that (1) the association between the persuasiveness of persuasive strategies and personality types varies acros...

People perceive psychological characteristics (PCs), such as the personality and values of a marriage partner, as extremely important factors in partner selection. Due to its importance, considerable work has investigated the relationship between couples’ PCs and their marital satisfaction, and their findings have been adopted by matchmaking servic...

Spatiotemporal data aggregated over regions or time windows at various resolutions demonstrate heterogeneous patterns and dynamics in each resolution. Meanwhile, the multi-resolution characteristic provides rich contextual information, which is critical for effective long-sequence forecasting. The importance of such inter-resolution information is...

Long-term spatio-temporal prediction (LTSTP) over different resolutions plays a crucial role in planning and dispatching smart city applications, such as smart transportation and smart grid. The Transformer, which has demonstrated superiority in capturing long-term dependencies, was recently studied for spatio-temporal prediction. However, it is di...

Existing research shows that it is effective for the persuader to personalize persuasion based on the persuadee’s personality. However, prior studies overlooked the interactions of multiple personality traits, which undermines the effectiveness of personalization. In this paper, we investigated the utility of personality types for predicting behavi...

Event detection has been proved important in various applications, such as route selection to avoid the congestion an event causes or deciding whether to join an event that one is interested in. While geotagged tweets are popular sources of information for event detection, they are usually insufficient for accurate detection when scarce. On the oth...

Existing research shows that it is effective for the persuader to per-sonalize persuasion based on the persuadee's personality. However, prior studies overlooked the interactions of multiple personality traits, which undermines the effectiveness of personalization. In this paper, we investigated the utility of personality types for predicting behav...

Learning user attributes is essential for providing users with a service. In particular, for e-commerce portals which deal in variety of goods ranging from clothes to foods to home electronics, it is especially important to learn “domain-independent” attributes such as age, gender, and personality that affect people’s behavior across various domain...

Despite the significance of assortativity as a property of networks that paves for the emergence of new structural types, surprisingly, there has been little research done on assortativity. Assortative networks are perhaps among the most prominent examples of complex networks believed to be governed by common phenomena, thereby producing structures...

The notion of a persuasive technology (PT) that is autonomous and intelligent, and more importantly, cognizant of and sensitive to human sense of agency (SoA), i.e., the subjective feeling or judgement that oneself is in control of situations, remains to be theorized, conceptualized and elucidated. Three important questions have emerged from our in...

We propose a model that combines only simple techniques to meet the challenge of cooking activity recognition. The challenge dataset is basically small, consisting only of four subjects where three are used for training and one for validation. In order not to overfit the small training data, we employed two simple classifiers, LightGBM and Naive Ba...

Sense of agency (SoA) refers to the experience or belief that one's own actions caused an external event. Here we present a model of SoA in the framework of optimal Bayesian cue integration with mutually involved principles, namely reliability of action and outcome sensory signals, their consistency with the causation of the outcome by the action,...

The concept of sense of agency (SoA) has garnered considerable attention in human science at least in the past two decades. Coincidentally, about two decades ago, artificial intelligence (AI) research witnessed an explosion of proposed theories on agency mostly based on dynamical approaches. However, despite this early burst of enthusiasm, SoA mode...

Despite the increasing significance of sense of agency (SoA) research, the literature lacks a formal model: what computational principles underlie SoA, the registration that oneself initiated an action that caused something to happen? We theorize SoA in the framework of optimal Bayesian cue integration with mutually involved principles, namely, rel...

The activity of the user is one example of context information which can help computer applications respond better to the needs of the user in a seamless manner based on the situation without needing explicit instruction. With potential applications in many fields such as health-care, assisted living and sports, there has been considerable interest...

We have previously introduced the concept of perception-based resilience that accounts for the impact of human perception in resilience thinking. In this paper, we further this concept as we argue for our novel perception-based resilience model by elucidating its theoretic and model bases and how they fit coherently in the model. To provide traces...

We tackled in this study the difficult problem of topic extraction in Thai tweets on the country's historic flood in 2011. After using Latent Dirichlet Allocation (LDA) to extract the topics, the first difficulty that faced us was the inaccuracy the word segmentation task that affected our interpretation of the LDA result. To solve this, we refined...

With our world witnessing critical systemic changes, we argue for a deeper understanding of what fundamentally constitutes and leads to critical system changes, and how the system can be resilient, i.e., persist in, adapt to, or transform from dramatically changing circumstances. We position our argument with long-standing theories on complexity, s...

Resilience is the ability of a complex system to persist in, adapt to, or transform from dramatically changing circumstances. Our objective is to characterize the resilience of a complex system in depth by looking at what fundamentally constitutes and leads to system changes and how the system can be resilient to these changes. Our characterization...

Marketers have devoted their efforts to comprehend the essence of negative electronic word-of-mouth (eWOM) since it was first introduced to understand brand crisis and the fact that social media facilitates rapid consumer communications that exacerbate the situation. This motivates marketers and researchers to investigate the complexities behind ne...

Researches on inertial navigation systems (INS) have formulated complex step detection algorithms and stride length estimations. But for current systems to work, INSs have to correctly identify negative pedestrian locomotion. Negative pedestrian locomotion are movements that a user can naturally make without any real position displacement, but has...

The predictive technique proposed in this project was initially designed for an indoor smart environment wherein intrusive tracking techniques, such as cameras, mobile phones, and GPS tracking systems, could not be appropriately utilized. Instead, we installed simple motion detection sensors in various areas of the experimental space and observed m...

Researches on inertial navigation systems (INS) have formulated complex step detection algorithms and stride length estimations. But for current systems to work, INSs have to correctly identify negative pedestrian locomotion. Negative pedestrian locomotion are movements that a user can naturally make without any real position displacement, but has...

Students engage in many learning activities outside of class but, it is not easy for them to learn on their own because they also need to identify what activities to perform, decide how long to engage in them, evaluate their progress, shift to other activities if needed and avoid distractions aside from others. This research designed and implemente...

The trend of multimodal interaction in interactive gaming has grown significantly as demonstrated for example by the wide acceptance of the Wii Remote and the Kinect as tools not just for commercial games but for game research as well. Furthermore, using the player’s affective state as an additional input for game manipulation has opened the realm...

Resilience is the capacity of a system to withstand large disturbances and dramatically changing conditions without losing its core purpose and integrity and achieve generalized recovery once failure happens. Due to the many dimensions of this concept, we have previously argued for a taxonomy that can account for its various contexts. What is signi...

Students often face difficulty in self-directed learning scenarios (e.g., studying, research) because they need to control many aspects of the learning session. They need to decide what to learn, how long to perform a learning task, when to shift to a different learning task and manage distractions apart from others. We observed from our previous r...

The integration of low cost microelectromagnetic (MEM) sensors into smart phones have made inertial navigation systems (INS) possible for ubiquitous use. Many research studies developed algorithms to detect a user's steps, and to calculate a user's stride to know the position displacement of the user. Subsequent research have already integrated the...

An upcoming trend of affective gaming is where a player's emotional state is used to manipulate game play. This is an interesting field to explore especially for the survival horror genre that is excellent at producing player's intense emotions. In this research, we analyzed different player affective states prior to (i.e., Neutral, Anxiety, Suspen...

Self-regulated learners have been shown to learn more effectively. However, it is not easy to become self-regulated because learners have to be capable of observing and evaluating their thoughts, actions and behaviors while learning. In this work, we used Q-learning to reveal the effectiveness or ineffectiveness of a learning behavior that carries...

Music induces different kinds of emotions in listeners. Previous research on music and emotions discovered that different music features can be used for classifying how certain music can induce emotions in an individual. We propose a method for collecting electroencephalograph (EEG) data from subjects listening to emotion-inducing music. The EEG da...

Self-reflection and self-evaluation are effective processes for identifying good learning behavior. These are essential in self-directed learning scenarios because students have to be responsible for their own learning. Although students benefit from doing fine-grained analysis of their own behavior, which we observed in our previous work, asking t...

Self-regulated learners are successful because of their ability to select learning strategies, monitor their learning outcomes and adapt them accordingly. However, it is not easy to measure the outcomes of a learning strategy especially while learning. We present an architecture that allows students to gauge the effectiveness of learning behavior a...

This study proposes a methodology in building a multimodal stress-level model using different non-invasive physiological signals: Galvanic Skin Response (GSR), Blood Volume Pulse (BVP), and Respiratory Variability (Resp). Paced Stroop, Mental Math, and Game were used to induce stress to 4 subjects. A fixed window size of 5 seconds and 1 second slid...

Knowledge about position and activities of the participants is commonly used in location-based services and applications in smart environment, which need to know an approximated location of the users to provide a proper service. Assistive services provided in the smart space can be divided into two categories: prompt services and delayed services....

Self-regulation is an important skill for students to possess. It allows them to learn more effectively and it has been shown to cause better learning gains. Self-regulation is not an easy task especially for poor learners. This is the motivation behind researches that use computer-based learning environments to promote self-regulation through embe...

With the preference toward more action games from audiences, survival horror games need more of the fearful quality that makes it distinct in order to once again become attractive to the market. In this paper, we investigate the impact of game events, which consist of several visual and audio elements, on player affect as opposed to looking at the...

Rhythm is one of the most essential elements of music that can easily capture the attention of the listener. In this study, we explored various rhythm features and used them to build emotion models. The emotion labels used are based on Thayers Model of Mood, which includes contentment, exuberance, anxiety, and depression. Empirical results identify...

This paper uses brainwaves to recognize the computer activity of the user and provides music recommendation. Twenty-three (23) hours of data collection was performed by asking the computer user to wear a device that collects electroencephalogram (EEG) signals from his brain as he performed whatever tasks he wanted to perform while listening to musi...

Commonly attributed to digital natives is the ability to quickly, yet effectively, shift from one task to another. However, several works have debunked this assumption by showing that multitasking even among digital natives led to poor learning performance and productivity. Our aim is to provide a tool to help digital natives be self-aware of desir...

Every person reacts differently to music. The task then is to identify a specific set of music features that have a significant effect on emotion for an individual. Previous research have used self-reported emotions or tags to annotate short segments of music using discrete labels. Our approach uses an electroencephalograph to record the subject's...

Learning is commonly associated with knowledge transfer involving guidance from a teacher. However, as people grow older they are expected to know how to learn by themselves. In this research, we analyzed student learning in an unsupervised learning environment, i.e., performing academic research, wherein students have complete control over their l...

The current generation is much accustomed to new technologies which enable them to perform many activities online. More importantly, these technologies have been used by students even for learning. In this research we focused on student initiated learning online. Because students have control over their own learning, they are not bounded by a sylla...

Time-interval sequential patterns provide information not only on frequently occurring items and the order in which they happen but also reveal the temporal dimension between successive items. Although time-interval data have been dealt with in the past - as single or multiple, regular or irregular, and/or with definite ranges, what we are proposin...

Annotation of emotion in music has traditionally used human tagging approaches. We propose a novel approach of identifying important musical features that can lead to automatic emotion annotation for music. Using psychophysiological responses of a subject listening to music and chord sequences of the songs, we identify music segments that can be us...

Research in psychology and SSP often describe posture as one of the most expressive nonverbal cues. Various studies in psychology particularly link posture mirroring behaviour to rapport. Currently, however, there are few studies which deal with the automatic analysis of postures and none at all particularly focus on its connection with rapport. Th...

Although the design of our constructive adaptive user interface (CAUI) for an affect-based music compositional artificial intelligence has been modified on several fronts since the time it was introduced, what has become a persisting limitation of our research is the extent by which it should efficiently cover music theory effectively. This paper r...

Emotion is an important field of study shared by many disciplines such as psychology, language and computer science. Most studies in emotion require the collection and annotation of data, but many issues arise from this process. This research focused on the following issues: collecting data from multiple sources, providing flexibility in emotion la...

Music can induce different emotions in people. We propose a system that can identifymusic segmentswhich induce specific emotions from the listener. The work involves building a knowledge base with mappings between affective states (happiness, sadness, etc.) and music features (rhythm, chord progression, etc.). Building this knowledge base requires...

In this paper we investigated student online learning and non-learning related activities. The data collected in the research showed that students felt certain affective states when performing particular activity types and performed particular activity types when they felt certain affective states. These transitions were further investigated by gen...

Advancement in ambient intelligence is driving the trend towards innovative interaction with computing systems. In this paper, we present our efforts towards the development of the ambient intelligent space TALA, which has the concept of empathy in cognitive science as its architecture's backbone to guide its human-system interactions. We envision...

Data-centric affect modeling may render itself restrictive in practical applications for three reasons, namely, it falls short of feature optimization, infers discrete affect classes, and deals with relatively small to average sized datasets. Though it seems practical to use the feature combinations already associated to commonly investigated senso...

Students have different ways of learning and have varied reactions to feedback. Thus, allowing a system to predict how students would appraise certain feedback gives it the capability to adapt to what would help a student learn better. This research focuses on the prediction of a student's appraisal of feedback provided in an intelligent tutoring s...

We demonstrate a method to locate relations and constraints between a music score and its impressions, by which we show that machine learning techniques may provide a powerful tool for composing music and analyzing human feelings. We examine its generality by modifying some arrangements to provide the subjects with a specified impression. This demo...

How do students learn? It is important to ask this question to know how to help them. Because of the power of the internet and the boom of social networking sites, students of the digital native generation have been able to use these tools to improve and expand their learning. However, not much support is given to students as they learn in these so...

Empathy is a learnable skill that requires experiential learning and practice of empathic ability for it to improve and mature. In the context of human-system interaction (HSI) this can mean that a system should be permitted to have an initial knowledge of empathy provision that is inaccurate or incomplete, but with this knowledge evolving and prog...

This research investigates the use of emotion data derived from analyzing change in activity in the autonomic nervous system (ANS) as revealed by brainwave production to support the creative music compositional intelligence of an adaptive interface. A relational model of the influence of musical events on the listener’s affect is first induced usin...

Studies have emphasized that empathy is a learnable skill that can be developed through learning from experience. Applying this in the context of human-system interaction, an empathic system is one that automatically acquires an initial knowledge of empathy that is permitted to be incomplete, hence inaccurate and imperfect, but improves this knowle...

The consideration of human feelings in automated music generation by intelligent music systems, albeit a compelling theme, has received very little attention. This work aims to computationally specify a system's music compositional intelligence that tightly couples with the listener's affective perceptions. First, the system induces a model that de...

This research investigates the use of emotion data derived from analyzing change in activity in the autonomic nervous system (ANS) as revealed by brainwave production to support the creative music compositional intelligence of an adaptive interface. A relational model of the influence of musical events on the listener's affect is first induced usin...

This paper presents the results of recent modification in the Constructive Adaptive User Interface (CAUI) that induces a model of emotional impressions towards certain musical piece structures and improvises a piece based on the model. The CAUI previously employed a ready-made melody generating module with its internal workings abstracted from the...

This paper presents the results of recent modification in the Constructive Adaptive User Interface (CAUI) that induces a model of emotional impressions towards certain musical piece structures and improvises a piece based on the model. The CAUI previously employed a ready-made melody generating module with its internal workings abstracted from the...

To have an instructional plan guide the learning process is significant to various teaching styles and an important task in an ITS. Though various approaches have been used to tackle this task, the compelling need is for an ITS to improve on its own the plans established in a dynamic way. We hypothesize that the use of knowledge derived from studen...

This paper discusses a predictive modeling framework actualized in a learning agent that uses logged tutorial interactions to discover predictive characteristics of students. The agent automatically forms cluster models that are described in terms of student–system interaction attributes, i.e., in terms of the student’s knowledge state and behaviou...

A conceptual clustering algorithm can search through huge amounts of data looking for multi-dimensional structures, where each structure or cluster represents a relevant concept in the problem-solving domain. We investigated on the effect of cluster knowledge for a learning agent to improve its prediction of higher level student response aspects. O...

Though various approaches have been used to tackle the task of instructional planning, the compelling need is for ITSs to improve their own plans dynamically. We have developed a Category-based Self-improving Planning Module (CSPM) for a tutor agent that utilizes the knowledge learned from automatically derived student categories to support efficie...

We have conceived of a Multi-agent Self-improving Planner (MSIP) within the tutor module of an intelligent tutoring system (ITS). It embodies a learning process that utilizes knowledge about different student categories to adapt and improve its instructional plans on the level of these categories. In this sense, the categories become recipients and...

This paper describes a learning process for the tutor of an intelligent tutoring system (ITS) to automatically learn models of student categories and self-improve its instructional plans on the level of these categories. Using real-world teaching scenarios as experiment data, we empirically show that for every category the tutor is able to efficien...

This paper discusses a machine learning framework that uses extraction, classification, and generalization techniques to classify students according to their cognitive and behavioral learning patterns and to categorize tutoring strategies of expert human tutors. A great deal of the discussion focuses on the use of reinforcement learning techniques,...

We define our empathy learning problem as determining the extent by which a system can perceive user affect and intention as well as ambient context, construct models of these perceptions and of its interaction behavior with the user, and incrementally improve on its own its models in order to effectively provide empathic responses that change the...

We propose a machine learning framework for an adaptive and self-improving tutoring system that performs efficient and effective teaching strategies, and learns these strategies efficiently while interacting on-line with students belonging to specific learning categories. Categories that may be learned via a conceptual clustering algorithm serves a...

As more and more information find their way to the internet, people are able to do more at their own desk than ever before, all in the comfort of a private environment. But as more activities, especially learning, are able to be done through the personal desktop space, the question is then raised of whether or not one is really engaged and/or learn...

Knowledge about position of the participants is commonly used in location-based services and applications in smart environment, which need to know an approximated location of the users to provide a proper service. Furthermore, when users are moving in an environment doing tasks, knowledge of the next location or destination of those movements can b...

Current music recommender systems only use basic information for recommending music to its listeners. These usually include artist, album, genre, tempo and other song information. Online recommender systems would include ratings and annotation tags by other people as well. We propose a recommender system that recommends music depending on how the l...