Design research is important for understanding and interrogating how emerging technologies shape human experience. However, design research with Machine Learning (ML) is relatively underdeveloped. Crucially, designers have not found a grasp on ML uncertainty as a design opportunity rather than an obstacle. The technical literature points to data and model uncertainties as two main properties of ML. Through post-phenomenology, we position uncertainty as one defining material attribute of ML processes which mediate human experience. To understand ML uncertainty as a design material, we investigate four design research case studies involving ML. We derive three provocative concepts: thingly uncertainty: ML-driven artefacts have uncertain, variable relations to their environments; pattern leakage: ML uncertainty can lead to patterns shaping the world they are meant to represent; and futures creep: ML technologies texture human relations to time with uncertainty. Finally, we outline design research trajectories and sketch a post-phenomenological approach to human-ML relations. CCS CONCEPTS• Human-centered computing → HCI theory, concepts and models.

No abstract

Design futuring approaches, such as speculative design, design fiction and others, seek to (re)envision futures and explore alternatives. As design futuring becomes established in HCI design research, there is an opportunity to expand and develop these approaches. To that end, by reflecting on our own research and examining related work, we contribute five modes of reflection. These modes concern formgiving, temporality, researcher positionality, real-world engagement, and knowledge production. We illustrate the value of each mode through careful analysis of selected design exemplars and provide questions to interrogate the practice of design futuring. Each reflective mode offers productive resources for design practitioners and researchers to articulate their work, generate new directions for their work, and analyze their own and others' work.

No abstract

Personal and wearable computing are moving toward smaller and more seamless devices. We explore how this trend could be mirrored in an authentication scheme based on electroencephalography (EEG) signals collected from the ear. We evaluate this model using a low cost, single-channel, consumer grade device for data collection. Using data from 12 study participants who performed a set of 5 mental tasks, we achieve a 44% reduction in half total error rate (HTER) compared with a random classifier, corresponding to a 72% authentication accuracy in within-participants analyses and a 60% reduction and 80% accuracy in between-participant analyses. Given our results and those of previous research, we conclude that earEEG shows potential as a uniquely convenient authentication method as it is integrable into devices like earbud headphones already commonly worn in the ear, and the mental gestures generating the signal are invisible to would-be eavesdroppers.