Consumption-based approaches in proactive detection for content moderation
Shahar Elisha, John N. Pougué-Biyong, Mariano Beguerisse-Díaz
The growing market of voice-enabled devices introduces new types of music search requests. As voice assistants can potentially support conversational requests, music requests can be more ambiguous than requests in typed search interfaces. However, these systems may not be able to fulfill ambiguous requests in a manner that matches the user need. In this work, we study an example of ambiguous requests which we term as non-specific queries (NSQs), such as “play music,” where users ask to stream content using a single utterance that does not specify what content they want to hear. To better understand user motivations for making NSQs, we conducted semi-structured qualitative interviews with voice users. We observed four themes that structure user perceptions of the benefits and shortcomings of making NSQs: the tradeoff between control and convenience, varying expectations for personalization, the effects of context on expectations, and learned user behaviors. We conclude with implications for how these themes can inform the interaction design of voice search systems in handling non-specific music requests in voice search systems.
Shahar Elisha, John N. Pougué-Biyong, Mariano Beguerisse-Díaz
A. Ghazimatin, E. Garmash, G. Penha, K. Sheets, M. Achenbach, O. Semerci, R. Galvez, M. Tannenberg, S. Mantravadi, D. Narayanan, O. Kalaydzhyan, D. Cole, B. Carterette, A. Clifton, P. N. Bennett, C. Hauff, M. Lalmas-Roelleke
Julie Jiang, Aditiya Ponnada, Ang Li, Ben Lacker, and Samuel F Way