Introduces multimodal trajectory retrieval, constructs the GAE-Bench benchmark from five GUI agent datasets, and trains GAE-Retriever, which outperforms zero-shot retrieval baselines.
Positions are represented in relative coordinates within the range [0,1] on the observation screenshot
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Universal Retrieval for Multimodal Trajectory Modeling
Introduces multimodal trajectory retrieval, constructs the GAE-Bench benchmark from five GUI agent datasets, and trains GAE-Retriever, which outperforms zero-shot retrieval baselines.