arXiv:2607. 13319v1 Announce Type: cross Abstract: High-speed off-road autonomy requires precise closed-loop control for a target vehicle while remaining robust across changing terrains.
arXiv:2607. 10362v1 Announce Type: new Abstract: Latent world models are trained to predict future states in a learned representation and are then deployed inside a planner that selects actions by simulating them forward.
arXiv:2610.28619v1 Announce Type: new Abstract: Multi-token prediction (MTP) can accelerate language-model decoding by drafting several tokens and verifying them in a single target-model pass. Howeve...
arXiv:2608. 12959v1 Announce Type: cross Abstract: Latent world models are judged by how well they predict, so when planning fails at long horizons the natural reading is that the predictor degrades.
AD-WM is a new action‑discriminative joint‑embedding world model designed for counterfactual model predictive control. It augments residual latent dynamics with action‑recovery regularization based on inverse dynamics and conditional mutual information, while discarding auxiliary heads at test time so that MPC remains unchanged. Experiments on OGBench‑Cube and other simulation environments show substantial gains in hard‑start success and mean success, and zero‑shot transfer to a Franka robot improves pick‑and‑place success from 42.2% to 71.1%.
The paper investigates the problem of sharing a single critic across multiple parallel environments in reinforcement learning. It shows that when environments assign different expected returns to the same state, a shared critic must reconcile conflicting value targets, which can distort advantage estimates and misguide policy updates. The authors propose a simple fix—providing the critic with the environment index—demonstrating through bandit models and experiments on CartPole, MuJoCo, BipedalWalker, and 16 Procgen games that this conditional critic stabilizes learning and boosts returns, achieving a 40.8% improvement in aggregate normalized return on unseen levels.
arXiv:2609.36830v1 Announce Type: new Abstract: Fully asynchronous reinforcement learning (RL) improves resource utilization in large language model post-training by overlapping rollout generation wi...
The paper proposes a method for selectively querying language‑model advice in reinforcement learning by predicting the value of potential responses and only querying when the expected benefit outweighs the cost. It introduces a certified, response‑contingent metareasoning framework that guarantees near‑optimal advice usage under certain assumptions, and demonstrates that a calibrated controller with Qwen2.5 advisors can improve task performance while drastically reducing the number of advice calls on the BabyAI benchmark.
The paper investigates how causal modeling can improve the robustness of world models to physical parameter shifts. By freezing a dependency mask derived from paired interventions, the authors reduce out‑of‑distribution RMSE on a synthetic benchmark from 1.769 to 0.697, nearly matching an oracle mask. In a robotic link‑mass extrapolation task, incorporating explicit affine coefficients for mass effects and using a known PD controller yields a consistent predictive gain, raising median joint R² from 0.609 to 0.727 across all runs, while a learned‑torque model competes with the best recurrent world model when out‑of‑range masses are supplied.
arXiv:2608. 01130v1 Announce Type: new Abstract: A broad range of models face the mismatch where they are updated through trajectory losses but are evaluated by downstream task reward.
arXiv:2606. 24039v1 Announce Type: cross Abstract: Robotics increasingly relies on GPUs for parallel simulation, large-scale learning, and neural-network inference.
arXiv:2606. 30852v1 Announce Type: new Abstract: Reasoning models spend different amounts of useful computation across instances, but it remains unclear when a learned stopping rule improves over simple confidence or convergence thresholds.