update

Author: zhangwencong317

index.html

@@ -635,7 +635,7 @@ <h2>Abstract</h2>
     Despite the sustained scaling on model capacity and data acquisition, 
     Vision–Language–Action (VLA) models remain brittle in contact-rich and dynamic manipulation tasks, 
     where minor execution deviations can compound into failures. 
-    While reinforcement learning (RL) offers a principled path to robustness, on-policy RL in the physical world is constrained by safety risk, hardware cost, and environment reset. To bridge this gap, we present <strong class="highlight">RISE</strong>, a scalable framework of robotic reinforcement learning via imagination. At its core is a <strong class="highlight">Compositional World Model</strong> that (i) predicts multi-view future via a controllable dynamics model, and (ii) evaluates imagined outcomes with a progress value model, producing informative advantages for the policy improvement. Such compositional design allows state and value to be tailored by best-suited yet distinct architectures and objectives. These components are integrated into a closed-loop self-improving pipeline that continuously generates imaginary rollouts, estimates advantages, and updates the policy in imaginary space without costly physical interaction. Across three challenging real-world tasks, <strong class="highlight">RISE</strong> yields significant improvement over prior art, with more than <span class="number-highlight">+35%</span> absolute performance increase in dynamic brick sorting, <span class="number-highlight">+45%</span> for backpack packing, and <span class="number-highlight">+35%</span> for box closing, respectively.
+    While reinforcement learning (RL) offers a principled path to robustness, on-policy RL in the physical world is constrained by safety risk, hardware cost, and environment reset. To bridge this gap, we present RISE, a scalable framework of robotic reinforcement learning via imagination. At its core is a Compositional World Model that (i) predicts multi-view future via a controllable dynamics model, and (ii) evaluates imagined outcomes with a progress value model, producing informative advantages for the policy improvement. Such compositional design allows state and value to be tailored by best-suited yet distinct architectures and objectives. These components are integrated into a closed-loop self-improving pipeline that continuously generates imaginary rollouts, estimates advantages, and updates the policy in imaginary space without costly physical interaction. Across three challenging real-world tasks, RISE yields significant improvement over prior art, with more than +35% absolute performance increase in dynamic brick sorting, +45% for backpack packing, and +35% for box closing, respectively.
   </p>
 </section>