AI Research 技术调研

标签: continuous-control

此标签下有6条笔记。

  • 2026年7月16日

    Dream to Control: Learning Behaviors by Latent Imagination (Dreamer)

    • world-model
    • model-based-rl
    • rssm
    • actor-critic
    • latent-imagination
    • value-gradient
    • dm-control
    • continuous-control
  • 2026年7月16日

    Learning and Planning in Complex Action Spaces (Sampled MuZero)

    • world-model
    • model-based-rl
    • mcts
    • sampled-actions
    • continuous-control
    • muzero
    • dm-control
    • policy-iteration
  • 2026年7月16日

    Temporal Difference Learning for Model Predictive Control

    • world-model
    • model-based-rl
    • mpc
    • latent-planning
    • td-learning
    • continuous-control
    • mppi
    • dmcontrol
    • meta-world
  • 2026年7月16日

    TD-MPC2: Scalable, Robust World Models for Continuous Control

    • world-model
    • model-based-rl
    • mpc
    • latent-planning
    • td-learning
    • continuous-control
    • multitask
    • simnorm
    • mppi
    • enc-layers
    • Q
    • enc层
  • 2026年7月16日

    EfficientZero V2: Mastering Discrete and Continuous Control with Limited Data

    • world-model
    • model-based-rl
    • mcts
    • gumbel-search
    • sample-efficiency
    • continuous-control
    • atari-100k
    • dmcontrol
    • off-policy-correction
  • 2026年7月16日

    PWM: Policy Learning with Multi-Task World Models

    • world-model
    • model-based-rl
    • first-order-gradient
    • differentiable-simulation
    • multitask
    • td-mpc2
    • continuous-control
    • policy-extraction

  • GitHub