Submit at EasyChair
Smart systems that apply complex reasoning to make decisions and plan behavior, such as decision support systems and personalized recommendations, are difficult for users to understand. Algorithms allow the exploitation of rich and varied data sources, in order to support human decision-making and/or taking direct actions; however, there are increasing concerns surrounding their transparency and accountability, as these processes are typically opaque to the user — e.g., because they are too technically complex to be explained or are protected trade secrets. The topics of transparency and accountability have attracted increasing interest to provide more effective system training, better reliability and improved usability. This workshop will provide a venue for exploring issues that arise in designing, developing and evaluating intelligent user interfaces that provide system transparency or explanations of their behavior. We will focus specifically on explaining systems and models toward ensuring fairness and social justice, such as approaches to detecting or mitigating algorithmic biases or discrimination (e.g., awareness, data provenance, and validation).
Suggested themes include, but are not limited to:
- What are explanations? What should they look like? What should be included in explanations and how (and to whom) should they be presented?
- Is transparency (or explainability) always a good idea? Can transparent algorithms or explanations “hurt” the user experience, and in what circumstances?
- How can we build (good) algorithmic systems, particularly those that demonstrate that they are fair, accountable, and unbiased?
- When are the optimal points at which explanations are needed for transparency?
- What are more transparent models that still have good performance in terms of speed and accuracy?
- What is important in user modeling for system transparency and explanations?
- What are possible metrics that can be used when evaluating transparent systems and explanations?
- How can we evaluate explanations and their ability to accurately explain underlying algorithms and overall systems’ behavior, especially for the goals of fairness and accountability?
- How can explanations allow human evaluators to select model(s) that are unbiased, such as by revealing traits or outcomes of the underlying learned system?
- What are important social aspects in interaction design for system transparency and explanations?
- How can we detect biases and discrimination in transparent systems?
- Through explanations, transparency, or other means, how can we raise stakeholders’ awareness of the potential risk for biases and social harms that could result from developing and using intelligent systems?
Researchers and practitioners in academia or industry who have an interest in these areas are invited to submit papers up to 8 pages (not including references) in ACM SIGCHI Paper Format. These submissions must be original and relevant contributions to the workshop theme. These submissions must be original and relevant contributions. Examples include, but not limited to, position papers summarizing authors’ existing research in this area and how it relates to the workshop theme, papers offering an industrial perspective on the workshop theme or a real-world approach to the workshop theme, papers that review the related literature and offer a new perspective, and papers that describe work-in-progress research projects.
The workshop will feature a keynote by Timnit Gebru who co-leads the Ethical Artificial Intelligence Team at Google. Paper authors will then present their work as part of thematic panels. The remainder of the workshop will consist of smaller group activities related to the workshop theme.
For further questions please contact the workshop organizers at email@example.com.