September 10, 2025SensorsOpen Access

Which AI Sees Like Us? Investigating the Cognitive Plausibility of Language and Vision Models via Eye-Tracking in Human-Robot Interaction

View Full Paper

Ask AI

Bookmark

Authors

KGKhashayar Ghamati MDMaryam Banitalebi Dehkordi AZAbolfazl ZarakiUniversity of Hertfordshire

Discussion

Member takes

Implication

Observational analysis explores cognitive processes in language and vision models during human-robot interaction, suggesting short-term memory enhances saliency prediction.

Key Points

Stateless LLaVA closely replicates human gaze patterns, showcasing the potential of language models in cognitive tasks.
Eye-tracking data from social human interactions provided insights into visual attention, establishing a benchmark for model evaluation.
State-of-the-art vision-language models were analyzed under different memory conditions, revealing mixed performance outcomes.
Findings highlight the importance of short-term memory in enhancing human-like visual attention in generative models.

Cite This Study

Ghamati et al. (2025) studied this question.

synapsesocial.com/papers/68c2404eb210217d6479a5fb https://doi.org/10.3390/s25154687

View Full Paper

Ask AI

Bookmark

Implication

Observational analysis explores cognitive processes in language and vision models during human-robot interaction, suggesting short-term memory enhances saliency prediction.

Which AI Sees Like Us? Investigating the Cognitive Plausibility of Language and Vision Models via Eye-Tracking in Human-Robot Interaction

Discussion

Member takes

Implication

Key Points

Cite This Study

Which AI Sees Like Us? Investigating the Cognitive Plausibility of Language and Vision Models via Eye-Tracking in Human-Robot Interaction

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider