Dr. Guilliem Alenyà
Title: Do explainability and semantic understanding play a role in personalisation?
Abstract: Understanding humans and how to adapt to everyone in a different manner is a complex task that can happen at very different levels. Assistive robotics research should envisage robots as tools for caregivers, professionals or not. In this context, we will explore how explainability and semantic understanding can play an important role on personalisation. We will also discuss about: who is in charge of personalising?
Prof. Shelly Levy-Tzedek
Title: Personalization of human-robot interactions in rehabilitation: where we need to go, and why it's hard to get there
Abstract: Socially assistive robots (SARs) can help meet the growing need for rehabilitation assistance. In my talk, I will argue that, as opposed to a single setting which might suffice in other circumstances, personalization of HRI in rehabilitation is multi-layered and requires frequent updates. Personalization in the context of rehabilitation is not only important for the purpose of engaging patients, but it is also essential for facilitating motor and cognitive recovery.
Prof. Mohamed Chetouani
Title: Human-Centered Robot Learning
Abstract: In this talk, we focus on main methods and models enabling humans to teach embodied social agents such as social robots, using natural interaction. Humans guide the learning process of such agents by providing various teaching signals, which could take the form of feedback, demonstrations and instructions. This overview describes how human teaching strategies are incorporated within machine learning models. We detail the approaches by providing definitions, technical descriptions, examples and discussions on limitations. We also address natural human biases during teaching. We discuss how personalization impacts task learning.
Prof. Andrea Bertolini
Title: The subtle line between personalization and user manipulation in a European regulatory perspective. A proposal for a technology assessment methodology for Artificial Intelligence Systems.
Abstract: The talk will address those aspect of HRI research that could give rise to concerns in a ethical and regulatory perspective in light of emerging EU regulation (the AI Act) and European legal principles. It will then introduce a methodology to assess potential concerns in HRI research, already in the design phase, also to allow strategies to tackle it, so as to minimize risks of facing legal bans and limitations.
Robot Behaviours for Transparent Human-Robot Interaction
Georgios Angelopoulos, Alessandra Rossi, and Silvia Rossi
Abstract: The lack of transparency in robot processes poses a significant challenge to effective human-robot collaboration, specifically due to the increasing demand for autonomous human-robot interaction. This is particularly relevant in non-industrial settings because it prevents humans from adequately comprehending a robot's intentions, progress, and decision-making rationale, which is essential for seamless interaction. To address this issue, this work proposes two methods for integrating transparent behaviours into social robots: a by-design and a by-learning approach. Our first step was to explore the by-design approach by endowing the robot with non-verbal behaviours in a learning scenario. Our results showed that incorporating non-verbal and verbal cues significantly enhanced the robot's transparency, leading to a more engaging interaction for human participants. Expressing emotions and using an inner dialogue (the robot explaining its thought process out loud) were identified as significant factors in making the robot's behaviour more transparent during learning. The next step was to investigate the by-learning approach, where we propose an RL mechanism that includes human preferences aiming at improving the transparency of the robot's behaviours. These findings highlight the need for effectual transparent behaviours in future social robotic applications.
Design of a Music Recommendation Robot Using Brainwave Analysis Based on Music Tempo
Dong-Ki Jeong, Jin-Young Kim, and Hyoung-Gook Kim
Abstract: This paper proposes a robot system that can automatically classify music by learning the characteristics of brain waves generated when listening to music at various tempos and recommend music using these functions. In the proposed robotic system, the correlation between the tempo features extracted from the music signal and the brainwave features extracted from the brainwave signal is learned via a regression deep neural network. Indeed, based on this learned regression model, the proposed robotic system automatically generates EEG signal features mapped to the auditory tempo characteristics of input music and applies these features to attention-based deep neural networks to automatically classify music. Initial experimental results present the possibility of the proposed tempo-based music classification framework.
Personalised multi-modal communication for HRI
Suna Bensch, Jiangeng Sun, Juan Pedro Bandera Rubio, Adrian Romero-Garces and Thomas Hellstroem
Abstract: One important aspect when designing understandable robots is how robots should communicate with a human user to be understood in the best way. In elder care applications this is particularly important, and also difficult since many older adults suffer from various kinds of impairments. In this paper we present a solution where communication modality and communication parameters are adapted to fit both a user profile and an environment model comprising information about light and sound conditions that may affect communication. The RASA dialogue manager is extended with necessary functionality and the operation is verified with a Pepper robot interacting with several personas with impaired vision, hearing, and cognition. Several relevant ethical questions are identified and briefly discussed, as a contribution to the WARN workshop.
All times are reported in Korean Standard Time (GMT + 9)
13:10 - 13:15 OPENING Session
13:15 - 13:45 Talk by: Dr. Guilliem Alenyà - Do explainability and semantic understanding play a role in personalisation? {live}
13:45 - 13:55 Paper presentation: Angelopoulos et al. - Robot Behaviours for Transparent Human-Robot Interaction
13:55 - 14:05 Paper presentation: Jeong et al. - Design of a Music Recommendation Robot Using Brainwave Analysis Based on Music Tempo
14:05 - 14:35 Talk by: Prof. Andrea Bertolini - The subtle line between personalization and user manipulation in a European regulatory perspective. A proposal for a technology assessment methodology for Artificial Intelligence Systems. {live}
14:35 - 15:00 Coffee Break ☕
15:00 - 15:30 Talk by: Prof. Mohamed Chetouani - Human-Centered Robot Learning {remote}
15:30 - 15:40 Paper presentation: Bensch et al. - Personalised multi-modal communication for HRI
15:40 - 16:10 Talk by: Prof. Shelly Levy-Tzedek - Personalization of human-robot interactions in rehabilitation: where we need to go, and why it's hard to get there {remote}
16:10 - 16:15 Closing Session
Invited Talks X (25+5)
Tentative program subject to changes