GiB uses self-supervised latent features and Mahalanobis distance to filter erroneous subtasks from mixed-quality human demonstrations, improving robot policy learning in simulation and real-world tasks.
Towards effective utilization of mixed-quality demonstrations in robotic manipulation via segment- level selection and optimization
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.RO 2years
2026 2verdicts
UNVERDICTED 2roles
background 1polarities
background 1representative citing papers
TREAD augments robotics datasets via VLM-based sub-task generation, video segmentation, and linguistic diversity to improve policy generalization on novel tasks in LIBERO benchmarks.
citing papers explorer
-
Good in Bad (GiB): Sifting Through End-user Demonstrations for Learning a Better Policy
GiB uses self-supervised latent features and Mahalanobis distance to filter erroneous subtasks from mixed-quality human demonstrations, improving robot policy learning in simulation and real-world tasks.
-
Task Robustness via Re-Labelling Vision-Action Robot Data
TREAD augments robotics datasets via VLM-based sub-task generation, video segmentation, and linguistic diversity to improve policy generalization on novel tasks in LIBERO benchmarks.