arXiv:2607. 08335v1 Announce Type: cross Abstract: Policy-based approaches to Bayesian experimental design (BED) allow the learning of deep policy networks that adaptively make intelligent design decisions based on previously collected data.
arXiv:2608.16466v2 Announce Type: replace-cross Abstract: Bayesian optimal experimental design (BOED) aims to collect informative data by optimizing an expected utility reflecting the goals of an exp...
arXiv:2607. 28408v1 Announce Type: new Abstract: This thesis studies policy learning in interactive systems where an agent observes a context, selects an action from a very large set, and receives partial feedback.
The paper investigates how reusing past samples can improve the sample efficiency of Proximal Policy Optimization (PPO). Two variants, wPPO-U and wPPO-BH, are introduced within a multiple importance weighting framework, each reusing data from recent iterations while preserving core PPO mechanics. The authors derive theoretical policy improvement bounds for both variants and empirically evaluate their impact on continuous control tasks.
The paper investigates how policy learning algorithms should balance expected welfare against sampling risk in evidence-based policymaking. It demonstrates that algorithmic stability—specifically, a policy’s insensitivity to the replacement of a single experimental unit—limits sampling risk. The authors introduce policy‑vote bagging, which trains on many subsamples and averages their votes, preserving expected welfare while improving expected utility for risk‑averse researchers, and provide sharp bounds linking estimation accuracy, subsample size, and welfare variation, including an exact guarantee under CARA utility.
arXiv:2509. 03456v2 Announce Type: replace-cross Abstract: Off-policy evaluation (OPE) and off-policy learning (OPL) are foundational for decision-making in offline contextual bandits.
arXiv:2607. 03385v1 Announce Type: cross Abstract: Policy learning has received substantial attention with the goal of learning policies from observational data for decision-making.
The paper introduces a reinforcement learning framework that selects among a portfolio of gradient‑based and derivative‑free optimizers during a run. At each decision point a recurrent policy reads the current run state and chooses both the next optimizer and its usage duration, passing the best solution and step size forward. The method is trained with a decoupled actor‑critic using the same runtime distribution metric as evaluation, and on unseen problems it outperforms all individual portfolio optimizers except at the smallest budgets, remaining robust to distribution shift.
arXiv:2607. 07769v1 Announce Type: cross Abstract: Starting from the utilization of deep neural networks to approximate the state-action value function that led to winning one of the most challenging games, to algorithmic advancements that allowed solving problems without even explicitly stating the rules of the challenge at hand, reinforcement learning research has been the center of remarkable scientific progress for the past decade.
arXiv:2602. 23672v2 Announce Type: replace-cross Abstract: This study proposes a General Bayes framework for policy learning.
arXiv:2606. 25197v1 Announce Type: new Abstract: Learning effective policies for adaptive data acquisition remains challenging: posterior-based methods rely on surrogate models and posterior approximations that can be misspecified or biased, while direct policy-learning methods map from historical observations and fail to exploit available model representations, making learning harder.
The paper investigates how reinforcement learning can be effectively applied to diffusion models for visual tasks, focusing on the role of likelihood estimation. By systematically separating policy‑gradient objectives, likelihood estimators, and rollout sampling schemes, the authors find that using an evidence lower bound (ELBO) based likelihood estimator computed from the final generated sample is the key factor for stable and efficient RL optimization, outweighing the choice of loss function. Experiments on SD 3.5 Medium across multiple reward benchmarks confirm that this approach improves GenEval scores from 0.24 to 0.95 in 90 GPU hours, outperforming existing methods such as FlowGRPO and the current state‑of‑the‑art without reward hacking.
arXiv:2602. 05999v3 Announce Type: replace Abstract: How does the amount of compute available to a reinforcement learning (RL) policy affect its learning?