The Discontent with Intent Estimation In-the-Wild
The Case for Unrealized Intentions
Hayley Hung (TU Delft - Pattern Recognition and Bioinformatics)
Litian Li (Student TU Delft)
Jord Molhoek (Student TU Delft)
Jing Zhou (Student TU Delft)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
The future of socially intelligent systems depends on developing abilities to anticipate and empathize with users. Whilst great strides have been made on developing systems for future behavior forecasting that sometimes also claim to do intention estimation, we argue that the predominant state-of-the-art treatment of these problems leads to a significant misunderstanding about this topic. This paper revisits intention estimation, describing the "intention by outcome" problem and how it severely limits a deeper understanding of the nature of the problem. We argue that without a deeper more nuanced understanding of how to develop intention estimation systems, we head into a severely biased world where intentions would only be considered valid by intelligent systems if they came true. Through a case study on estimating unrealized intentions to speak in-the-wild, we highlight open challenges of this largely unexplored topic.