Kuk-Jin Yoon

KAIST

収録論文 13本 ・ フィジカルAI/ロボット学習

※arXiv著者名で収集。同姓同名の別人の論文が含まれる場合があります。

論文

• 検索前に再利用せよ：具現化マルチモーダルポリシーのテスト時拡張における余裕と相補性の診断
  VLA/テスト時拡張2026/8/18

  凍結された視覚-言語-行動(VLA)ポリシーをテスト時に改善する際、追加サンプリングと外部デモ検索のどちらが有効かを、回復可能な余裕と検索相補性という2つの指標で診断し、選択器が性能を向上させることを示した。

• Unified Prediction and Planning via Conflict-Aware Disjoint Parameter Training
  2026/7/1
• HEAT: Heterogeneous End-to-End Autonomous Driving via Trajectory-Guided World Models
  2026/5/1
• Cooperative Informative Sensing for Monitoring Dynamic Indoor Environments via Multi-Agent Reinforcement Learning
  2026/4/1
• DeLTa: Demonstration and Language-Guided Novel Transparent Object Manipulation
  2025/10/1
• Interaction-Merged Motion Planning: Effectively Leveraging Diverse Motion Datasets for Robust Planning
  2025/7/1
• Non-differentiable Reward Optimization for Diffusion-based Autonomous Motion Planning
  2025/7/1
• Any6D: Model-free 6D Pose Estimation of Novel Objects
  2025/3/1
• Cognitive TransFuser: Semantics-guided Transformer-based Sensor Fusion for Improved Waypoint Prediction
  2023/8/1
• One-Shot Neural Fields for 3D Object Understanding
  2022/10/1
• SiamEvent: Event-based Object Tracking via Edge-aware Similarity Learning with Siamese Networks
  2021/9/1
• Deceiving Image-to-Image Translation Networks for Autonomous Driving with Adversarial Perturbations
  2020/1/1
• Exploiting Feature Confidence for Forward Motion Estimation
  2017/4/1