AI Research 技术调研

标签: self-supervised-pretraining

此标签下有2条笔记。

  • 2026年7月16日

    R3M: A Universal Visual Representation for Robot Manipulation

    • visual-representation-learning
    • ego4d
    • time-contrastive-learning
    • video-language-alignment
    • self-supervised-pretraining
    • frozen-visual-encoder
    • imitation-learning
    • resnet
    • robot-manipulation
  • 2026年7月16日

    Visual Point Cloud Forecasting enables Scalable Autonomous Driving (ViDAR)

    • world-model
    • autonomous-driving
    • self-supervised-pretraining
    • point-cloud-forecasting
    • bev
    • latent-rendering
    • nuscenes
    • cvpr2024

  • GitHub