T2-VLM generates temporally consistent rewards for robot manipulation by tracking VLM-defined subgoal completion with a particle filter, improving reward accuracy and cutting VLM query time.
Manipulate- anything: Automating real-world robots using vision- language models
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.RO 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Training-free Generation of Temporally Consistent Rewards from VLMs
T2-VLM generates temporally consistent rewards for robot manipulation by tracking VLM-defined subgoal completion with a particle filter, improving reward accuracy and cutting VLM query time.