Adaptive Mixing of Policies from Searching and Policies from Learning

Looped architectures scale computation by reusing the same parameters across recurrent steps, and recent work shows that they substantially improve deep reinforcement learning policies on long-horizon tasks. Since recurrent depth directly controls computation, one may expect looped policies to naturally support elastic inference across recurrent depths. Surprisingly, we find that pretrained looped policies exhibit severe recurrent-depth specialization: reliable decisions are concentrated near the full trained depth, tying deployment computation to this depth even when less computation may suffice. Achieving depth elasticity, i.e., reliable decisions across recurrent depths with adaptive computation at deployment, therefore remains a key challenge. To address this, we propose FlexLoop, a novel post-training framework that converts pretrained fixed-depth looped policies into depth-elastic policies. FlexLoop keeps training on the original RL objective to preserve full-depth capability while performing adjacent-depth policy distillation to progressively transfer decision quality from deeper to shallower recurrent steps. The resulting policy supports reliable inference across recurrent depths and enables state-wise adaptive inference through recurrent-depth consistency. Experiments on $30$ online and offline long-horizon goal-conditioned environments show that FlexLoop preserves full-depth performance while making shallower depths effective. Keeping competitive performance, FlexLoop reduces average recurrent depth by up to $\bf{43\%}$ and achieves up to $\bf{1.34\times}$ wall-clock speedup in a stress test.

Monte Carlo Tree Search (MCTS) is a widely used approach for policy improvement and action selection in Reinforcement Learning. Due to its sequential and deterministic nature, principled runtime-scaling of MCTS with parallel compute remains a major challenge. We introduce Particle MCTS (PMCTS), a parallel MCTS algorithm which is suited for neural network evaluations, designed for GPU-acceleration with batch-parallelization and retains MCTS's principled approximate policy improvement interpretation. Empirically, PMCTS scales well with parallel compute and consistently outperforms or compares well to the popular heuristic-based baselines across a range of popular discrete- and continuous-action benchmark domains, including Chess, 19x19 Go, 9x9 Go, Gardner Chess, Snake, classical control environments from Brax and LLM reasoning in Sokoban.

Deep search agents operate over trajectories spanning dozens of steps, yet standard reinforcement learning provides only a single outcome reward per trajectory, which is far too sparse for effective credit assignment. On-policy self-distillation (OPSD) addresses this by using the model's own logits as dense token-level teachers, but extending it to search agents introduces a fundamental tension: the teacher, having access to privileged information such as the correct answer, produces a distribution that differs systematically from the student's exploration-based reasoning, and naive distillation causes the student to inherit this information asymmetry rather than learn better search strategies. We resolve this tension through two contributions. First, we construct Evidence Anchors, which are concise, step-level evidence snippets extracted from the web, as privileged information that captures key reasoning steps without revealing the entire answer path. Second, we propose Step-Level Self-Distilled Policy Optimization (SSPO), which converts teacher-student disagreement into step-level advantage weights within GRPO, applied exclusively to incorrect trajectories. This design decouples what to update from how much to update: the outcome reward determines the direction of policy change, while the teacher modulates its magnitude at each step. Correct trajectories are left untouched, preserving their diversity. On Qwen3-8B, SSPO consistently outperforms GRPO across BrowseComp, GAIA, and FRAMES, surpassing or matching GRPO trained with twice as many gradient steps while adding only about 5 percent overhead per step from a single additional forward pass.