Computing and Information Systems - Research Publications
Permanent URI for this collection
Search Results
1 - 10 of 61
-
ItemNo Preview AvailableDirective Explanations for Actionable Explainability in Machine Learning Applications(ASSOC COMPUTING MACHINERY, 2023-12)In this article, we show that explanations of decisions made by machine learning systems can be improved by not only explaining why a decision was made but also explaining how an individual could obtain their desired outcome. We formally define the concept of directive explanations (those that offer specific actions an individual could take to achieve their desired outcome), introduce two forms of directive explanations (directive-specific and directive-generic), and describe how these can be generated computationally. We investigate people’s preference for and perception toward directive explanations through two online studies, one quantitative and the other qualitative, each covering two domains (the credit scoring domain and the employee satisfaction domain). We find a significant preference for both forms of directive explanations compared to non-directive counterfactual explanations. However, we also find that preferences are affected by many aspects, including individual preferences and social factors. We conclude that deciding what type of explanation to provide requires information about the recipients and other contextual information. This reinforces the need for a human-centered and context-specific approach to explainable AI.
-
ItemSocio-technical context for insertable devices(FRONTIERS MEDIA SA, 2022-11-21)In this article, we show that voluntarily inserting devices inside the body is contested and seek to understand why. This article discusses insertables as a source of contestation. To describe and understand the social acceptability, reactions toward, and rhetoric surrounding insertable devices, we examine (i) the technical capabilities of insertable devices (the technical context), (ii) human reactions toward insertables (the social context), and (iii) the regulatory environment. The paper offers explanations to the misperceptions about insertables.
-
ItemPlaying blind: Revealing the world of gamers with visual impairment(ACM, 2019-05)Previous research on games for people with visual impairment (PVI) has focused on co-designing or evaluating specific games - mostly under controlled conditions. In this research, we follow a game-agnostic, "in-the-wild" approach, investigating the habits, opinions and concerns of PVI regarding digital games. To explore these issues, we conducted an online survey and follow-up interviews with gamers with VI (GVI). Dominant themes from our analysis include the particular appeal of digital games to GVI, the importance of social trajectories and histories of gameplay, the need to balance complexity and accessibility in both games targeted to PVI and mainstream games, opinions about the state of the gaming industry, and accessibility concerns around new and emerging technologies such as VR and AR. Our study gives voice to an underrepresented group in the gaming community. Understanding the practices, experiences and motivations of GVI provides a valuable foundation for informing development of more inclusive games.
-
ItemNot For Me: Older Adults Choosing Not to Participate in a Social Isolation Intervention(Association for Computing Machinery, 2016-05)This paper considers what we can learn from the experiences of people who choose not to participate in technology-based social interventions. We conducted ethnographically-informed field studies with socially isolated older adults, who used and evaluated a new iPad application designed to help build new social connections. In this paper we reflect on how the values and assumptions guiding the technological intervention were not always shared by those participating in the evaluation. Drawing on our field notes and interviews with the older adults who chose to discontinue participation, we use personas to illustrate the complexities and tensions involved in individual decisions to not participate. This analysis contributes to HCI research calling for a more critical perspective on technological interventions. We provide detailed examples highlighting the complex circumstances of our non-participants' lives, present a framework that outlines the socio-technical context of non-participation, and use our findings to promote reflective practice in HCI research that aims to address complex social issues.
-
ItemClassifying Attention Types with Thermal Imaging and Eye Tracking(Association for Computing Machinery (ACM), 2019-09-09)Despite the importance of attention in user performance, current methods for attention classification do not allow to discriminate between different attention types. We propose a novel method that combines thermal imaging and eye tracking to unobtrusively classify four types of attention: sustained, alternating, selective, and divided. We collected a data set in which we stimulate these four attention types in a user study (N = 22) using combinations of audio and visual stimuli while measuring users' facial temperature and eye movement. Using a Logistic Regression on features extracted from both sensing technologies, we can classify the four attention types with high AUC scores up to 75.7% for the user independent-condition independent, 87% for the user-independent-condition dependent, and 77.4% for the user-dependent prediction. Our findings not only demonstrate the potential of thermal imaging and eye tracking for unobtrusive classification of different attention types but also pave the way for novel applications for attentive user interfaces and attention-aware computing.
-
ItemEvaluating Real-Time Gaze Representations to Infer Intentions in Competitive Turn-Based Strategy Games(Association for Computing Machinery, 2017)In this paper, we investigate nine different visual representations of gaze in a competitive digital game setting. We evaluate the ability of spectators to infer a player's intentions in the game for each visual representation. Our results show that spectators have a remarkable ability to infer intent accurately using all nine visualizations, but that visualizations with certain characteristics were more comprehensible and more readily revealed the player's intent. The real-time Heatmap visualization was the most highly preferred by participants and the most effective in revealing intent, due to its ability to balance real-time gaze information with a persistent summary of recent gaze behaviour. Our findings show that eye-tracking visualization can enable playful interactions in competitive games based on players' ability to interpret opponents' attention and intention through gaze information.
-
ItemNo Preview AvailableNonverbal Communication in Human-AI Interaction: Opportunities and Challenges(Cardiff University Press, 2020-05-07)
-
ItemNo Preview AvailableExplainable Reinforcement Learning through a Causal Lens(ASSOC ADVANCEMENT ARTIFICIAL INTELLIGENCE, 2020)Prominent theories in cognitive science propose that humans understand and represent the knowledge of the world through causal relationships. In making sense of the world, we build causal models in our mind to encode cause-effect relations of events and use these to explain why new events happen by referring to counterfactuals — things that did not happen. In this paper, we use causal models to derive causal explanations of the behaviour of model-free reinforcement learning agents. We present an approach that learns a structural causal model during reinforcement learning and encodes causal relationships between variables of interest. This model is then used to generate explanations of behaviour based on counterfactual analysis of the causal model. We computationally evaluate the model in 6 domains and measure performance and task prediction accuracy. We report on a study with 120 participants who observe agents playing a real-time strategy game (Starcraft II) and then receive explanations of the agents' behaviour. We investigate: 1) participants' understanding gained by explanations through task prediction; 2) explanation satisfaction and 3) trust. Our results show that causal model explanations perform better on these measures compared to two other baseline explanation models.
-
ItemNo Preview AvailableIntroducing the Gamer Information-Control Framework(ACM, 2020-04-21)In this paper, we present a foundation for understanding the elements that enable people with visual impairment to engage with digital games. This is defined by the gamer's relation- ships with information and with elements of control provided by the game, and is mediated through in-game metaphors and affordances when gamers interact as users or creators. This work complements previous research exploring the points of view of gamers with visual impairment by focusing on the games they play and prioritising the relationships between the key enablers of access to digital games. Using the framework to examine existing and missing components will enable de- signers to consider broader aspects of accessibility in game design.
-
ItemAvatar-Mediated Communication in Social VR: An In-depth Exploration of Older Adult Interaction in an Emerging Communication Platform(Association for Computing Machinery (ACM), 2021-05-08)While HCI researchers have begun designing personalised VR experiences for older adults, there has been limited research examining the use of social VR - where users interact via avatars in a virtual environment. Avatar-mediated communication (AMC) is a crucial component of the social VR experience, but older users’ experience with AMC is poorly understood. We conducted a five-month study with 16 older adults evaluating a co-designed social VR prototype. Results show that AMC in social VR was seen as medium that supported introverted users to express themselves and was viewed as offering advantages when discussing sensitive topics. Our study provides new insights into how older adults view AMC in social VR as a communication medium and we contribute six design reflections, based on our results, that highlight the steps that can be taken to ensure that AMC in social VR can meet the communication needs of older users.