SGN-CIRL combines SAC reinforcement learning, imitation learning, and curriculum learning with a CLIP-pooled 3D scene graph, reporting higher navigation success in Isaac Sim when the graph is used.
Graph attention networks,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.RO 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
SGN-CIRL: Scene Graph-based Navigation with Curriculum, Imitation, and Reinforcement Learning
SGN-CIRL combines SAC reinforcement learning, imitation learning, and curriculum learning with a CLIP-pooled 3D scene graph, reporting higher navigation success in Isaac Sim when the graph is used.