To build an Artificial Intelligence system that can assist us in daily lives, the ability to understand the world around us through visual input is essential. Prior studies train visual perception models by defining concept vocabularies and annotate data against the fixed vocabulary. It is hard to define a comprehensive set of everything, and thus…

Highly-automated technologies are increasingly incorporated into existing systems, for instance in advanced car models. Although highly automated modes permit non-driving activities (e.g. internet browsing), drivers are expected to reassume control upon a 'take over' signal from the automation. To assess a person's readiness for takeover,…

Recent research shows that deaf signers show increased behavioral and neural sensitivity to certain types of movement, such as biological motion, human actions, and signing avatars. However, other work suggests that in deaf signers exposed to signed language before age five, the mirror mechanism has minimal involvement during the perception of…

In real-world settings, vision language models (VLMs) should robustly handle naturalistic, noisy visual content as well as domain-specific language and concepts. For example, K-12 educators using digital learning platforms may need to examine and provide feedback across many images of students' math work. To assess the potential of VLMs to support…

Information processing researchers have assumed that algebra symbol skills depend on mastery of the abstract rules presented in the curriculum (Matz, 1980; Sleeman, 1986). Thus, students' ubiquitous algebra errors have been taken as indicating the need to embed algebra in rich contextual settings (Kaput, 1995; National Council of Teachers of…