Explainable Artificial Intelligence (XAI) aims to provide insights into the inner workings and the outputs of AI systems. Recently, there’s been growing recognition that explainability is inherently human-centric, tied to how people perceive explanations. Despite this, there is no consensus in the research community on whether user evaluation is crucial in XAI, and if so, what exactly needs to be evaluated and how. This systematic literature review addresses this gap by providing a detailed overview of the current state of affairs in human-centered XAI evaluation. We reviewed 73 papers across various domains where XAI was evaluated with users. These studies assessed what makes an explanation “good” from a user’s perspective, i.e., what makes an explanation meaningful to a user of an AI system. We identified 30 components of meaningful explanations that were evaluated in the reviewed papers and categorized them into a taxonomy of human-centered XAI evaluation, based on: (a) the contextualized quality of the explanation, (b) the contribution of the explanation to human-AI interaction, and (c) the contribution of the explanation to human- AI performance. Our analysis also revealed a lack of standardization in the methodologies applied in XAI user studies, with only 19 of the 73 papers applying an evaluation framework used by at least one other study in the sample. These inconsistencies hinder cross-study comparisons and broader insights. Our findings contribute to understanding what makes explanations meaningful to users and how to measure this, guiding the XAI community toward a more unified approach in human-centered explainability.
MULTIFILE
In this paper, the authors address a literature gap with regard to sponsorship outcomes of mega-events and their host countries. This paper is about research that investigates the interrelatedness of three important images-host country, mega-event, and sponsor images-from the perspective of a cameo appearance building on the sponsorship and brand placement literature. It is based on the premise that the host city makes a cameo appearance during a mega-event for sport tourists while the event itself makes a cameo appearance for residents of the host country. The results indicate that mega-events can have a transitory influence, and that cameo effects exist, but that the patterns of relationships are different for sport tourists and residents.
Aim: To evaluate healthcare professionals' performance and treatment fidelity in the Cardiac Care Bridge (CCB) nurse-coordinated transitional care intervention in older cardiac patients to understand and interpret the study results. Design: A mixed-methods process evaluation based on the Medical Research Council Process Evaluation framework. Methods: Quantitative data on intervention key elements were collected from 153 logbooks of all intervention patients. Qualitative data were collected using semi-structured interviews with 19 CCB professionals (cardiac nurses, community nurses and primary care physical therapists), from June 2017 until October 2018. Qualitative data-analysis is based on thematic analysis and integrated with quantitative key element outcomes. The analysis was blinded to trial outcomes. Fidelity was defined as the level of intervention adherence. Results: The overall intervention fidelity was 67%, ranging from severely low fidelity in the consultation of in-hospital geriatric teams (17%) to maximum fidelity in the comprehensive geriatric assessment (100%). Main themes of influence in the intervention performance that emerged from the interviews are interdisciplinary collaboration, organizational preconditions, confidence in the programme, time management and patient characteristics. In addition to practical issues, the patient's frailty status and limited motivation were barriers to the intervention. Conclusion: Although involved healthcare professionals expressed their confidence in the intervention, the fidelity rate was suboptimal. This could have influenced the non-significant effect of the CCB intervention on the primary composite outcome of readmission and mortality 6 months after randomization. Feasibility of intervention key elements should be reconsidered in relation to experienced barriers and the population. Impact: In addition to insight in effectiveness, insight in intervention fidelity and performance is necessary to understand the mechanism of impact. This study demonstrates that the suboptimal fidelity was subject to a complex interplay of organizational, professionals' and patients' issues. The results support intervention redesign and inform future development of transitional care interventions in older cardiac patients.
In the last decade, the automotive industry has seen significant advancements in technology (Advanced Driver Assistance Systems (ADAS) and autonomous vehicles) that presents the opportunity to improve traffic safety, efficiency, and comfort. However, the lack of drivers’ knowledge (such as risks, benefits, capabilities, limitations, and components) and confusion (i.e., multiple systems that have similar but not identical functions with different names) concerning the vehicle technology still prevails and thus, limiting the safety potential. The usual sources (such as the owner’s manual, instructions from a sales representative, online forums, and post-purchase training) do not provide adequate and sustainable knowledge to drivers concerning ADAS. Additionally, existing driving training and examinations focus mainly on unassisted driving and are practically unchanged for 30 years. Therefore, where and how drivers should obtain the necessary skills and knowledge for safely and effectively using ADAS? The proposed KIEM project AMIGO aims to create a training framework for learner drivers by combining classroom, online/virtual, and on-the-road training modules for imparting adequate knowledge and skills (such as risk assessment, handling in safety-critical and take-over transitions, and self-evaluation). AMIGO will also develop an assessment procedure to evaluate the impact of ADAS training on drivers’ skills and knowledge by defining key performance indicators (KPIs) using in-vehicle data, eye-tracking data, and subjective measures. For practical reasons, AMIGO will focus on either lane-keeping assistance (LKA) or adaptive cruise control (ACC) for framework development and testing, depending on the system availability. The insights obtained from this project will serve as a foundation for a subsequent research project, which will expand the AMIGO framework to other ADAS systems (e.g., mandatory ADAS systems in new cars from 2020 onwards) and specific driver target groups, such as the elderly and novice.
266 woorden Op school kan de situatie zich voordoen dat de leerkracht onvoldoende tegemoet kan komen aan de extra ondersteuning die leerlingen met autisme nodig hebben. De klas kan te groot zijn, de leerkracht kan handelingsverlegen zijn, etc.. In dit projectplan wordt onderbouwd wat de relevantie is voor de dagelijkse praktijk van de leerkracht en de leerling met autisme en daaraan gerelateerde problemen. Tevens wordt onderbouwd waarom beeldende therapie theoretisch en empirisch kan bijdragen als creatieve oplossing voor kinderen met aan autisme gerelateerde problemen die in de klas extra aandacht vragen. Deze kinderen hebben een andere manier van informatie verwerken, kunnen zich vaak verbaal moeilijk uiten en hebben vaak sociale problemen. Deze kinderen lopen risico op verslavingsproblematiek (33%) en eenzaamheid, angst en depressie op volwassen leeftijd (80%). Kunstvormen in een leeromgeving bieden andere mogelijkheden voor kinderen om zich te uiten en om samen te werken. In dit projectplan wordt beschreven waarom het zinvol is te onderzoeken wat de effectiviteit is van beeldende therapie voor kinderen met autisme in primair (speciaal) onderwijs, ter preventie van risicogedrag. Het behandelprogramma ‘Zelf in beeld, beeldende therapie voor kinderen met autisme (bijlage 1) lijkt veelbelovende resultaten op te leveren (Schweizer, 2020). Om een indruk van de resultaten van praktijkgericht onderzoek naar ‘Zelf in beeld’ te krijgen kunt u de korte animatie bekijken (3 min): https://youtu.be/cVAAzRHZnb0 In dit vervolgproject wordt verkend in hoeverre ‘Zelf in beeld’ van toegevoegde waarde van kan zijn voor kind, leerkracht en ouders, binnen de setting van Speciaal Onderwijs. Dit project heeft een innovatief karakter omdat er een nieuwe vorm van (preventief) werken binnen passend onderwijs wordt toegepast en onderzocht.
Today, embedded devices such as banking/transportation cards, car keys, and mobile phones use cryptographic techniques to protect personal information and communication. Such devices are increasingly becoming the targets of attacks trying to capture the underlying secret information, e.g., cryptographic keys. Attacks not targeting the cryptographic algorithm but its implementation are especially devastating and the best-known examples are so-called side-channel and fault injection attacks. Such attacks, often jointly coined as physical (implementation) attacks, are difficult to preclude and if the key (or other data) is recovered the device is useless. To mitigate such attacks, security evaluators use the same techniques as attackers and look for possible weaknesses in order to “fix” them before deployment. Unfortunately, the attackers’ resourcefulness on the one hand and usually a short amount of time the security evaluators have (and human errors factor) on the other hand, makes this not a fair race. Consequently, researchers are looking into possible ways of making security evaluations more reliable and faster. To that end, machine learning techniques showed to be a viable candidate although the challenge is far from solved. Our project aims at the development of automatic frameworks able to assess various potential side-channel and fault injection threats coming from diverse sources. Such systems will enable security evaluators, and above all companies producing chips for security applications, an option to find the potential weaknesses early and to assess the trade-off between making the product more secure versus making the product more implementation-friendly. To this end, we plan to use machine learning techniques coupled with novel techniques not explored before for side-channel and fault analysis. In addition, we will design new techniques specially tailored to improve the performance of this evaluation process. Our research fills the gap between what is known in academia on physical attacks and what is needed in the industry to prevent such attacks. In the end, once our frameworks become operational, they could be also a useful tool for mitigating other types of threats like ransomware or rootkits.
