Eliciting User Preferences for Personalized Explanations for Video Summaries

Abstract

Video summaries or highlights are a compelling alternative for exploring and contextualizing unprecedented amounts of video material. However, the summarization process is commonly automatic, non-transparent and potentially biased towards particular aspects depicted in the original video. Therefore, our aim is to help users like archivists or collection managers to quickly understand which summaries are the most representative for an original video. In this paper, we present empirical results on the utility of different types of visual explanations to achieve transparency for end users on how representative video summaries are, with respect to the original video. We consider four types of video summary explanations, which use in different ways the concepts extracted from the original video subtitles and the video stream, and their prominence. The explanations are generated to meet target user preferences and express different dimensions of transparency: concept prominence, semantic coverage, distance and quantity of coverage.In two user studies we evaluate the utility of the visual explanations for achieving transparency for end users. Our results show that explanations representing all of the dimensions have the highest utility for transparency, and consequently, for understanding the representativeness of video summaries. CCS CONCEPTS• Human-centered computing → User studies; Empirical studies in HCI.

Design Implications for Explanations: A Case Study on Supporting Reflective Assessment of Potentially Misleading Videos

“…While video summaries offer quick and concise video overviews, they could potentially miss relevant and important information and thus lead to misinformation and deception. Graphical explanations can support users in better understanding how representative video summaries are for the original video ( Inel et al, 2020 ). With our reflection triggers, consumers could also actively reflect on the content of the videos, while watching a summary.…”

Measuring the benefit of increased transparency and control in news recommendation

“…Finally, we note that there is immense value in qualitative studies of news recommender studies, as such studies can provide deep, qualitative insights into the effect of transparency and control on users perspective and consumption behaviors. For example, one can conduct Grounded Theory studies to establish theories of the effects of support mechanisms on users' psychology (Charmaz 2006), or Contextual Design studies to gain a thorough understanding of users' experiences and their system needs (Holtzblatt and Beyer 2011). Such studies are particularly useful when investigating evaluation targets that are highly contextdependent and/or not yet very well understood, such as fairness, serendipity or surprise.…”

“…Qualitative studies vary from purely observational studies to in-depth user interviews, and from single sessions to long-running studies where the researcher follows a user over a longer period of time (Holtzblatt and Beyer 2011). Note, though, that the collection and analysis of qualitative data is particularly labor-intensive, especially when it must integrate into a larger real world research infrastructure.…”

Fine-tuning machine confidence with human relevance for video discovery

“…The popularity of video on the Web is reaching an unimaginable scale-a person would need more than 5 million years to watch all videos available online up to 2020 [1]. This scale poses an immense challenge to the discovery of relevant videos.…”