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…

Infants appear to progress from universal to language-specific event perception. In Japanese, two different verbs describe a person crossing a "bounded ground" (e.g., street) versus an unbounded ground (e.g., field) while in English, the same verb -- "crossing" -- describes both events. Interestingly, Japanese "and"…