5 papers
Information-Theoretic Policy Pre-Training with Empowerment
Moritz Schneider, Robert Krug, Narunas Vaskevicius +3
Empowerment, an information-theoretic measure of an agent's potential influence on its environment, has emerged as a powerful intrinsic motivation and exploration framework for rei…
GraphEQA: Using 3D Semantic Scene Graphs for Real-time Embodied Question Answering
Saumya Saxena, Blake Buchanan, Chris Paxton +6
In Embodied Question Answering (EQA), agents must explore and develop a semantic understanding of an unseen environment to answer a situated question with confidence. This problem…
OpenSplat3D: Open-Vocabulary 3D Instance Segmentation using Gaussian Splatting
Jens Piekenbrinck, Christian Schmidt, Alexander Hermans +3
3D Gaussian Splatting (3DGS) has emerged as a powerful representation for neural scene reconstruction, offering high-quality novel view synthesis while maintaining computational ef…
RelationField: Relate Anything in Radiance Fields
Sebastian Koch, Johanna Wald, Mirco Colosi +4
Neural radiance fields are an emerging 3D scene representation and recently even been extended to learn features for scene understanding by distilling open-vocabulary features from…
The Surprising Ineffectiveness of Pre-Trained Visual Representations for Model-Based Reinforcement Learning
Moritz Schneider, Robert Krug, Narunas Vaskevicius +2
Visual Reinforcement Learning (RL) methods often require extensive amounts of data. As opposed to model-free RL, model-based RL (MBRL) offers a potential solution with efficient da…