The paper proposes a method for allocating a limited budget of expert annotations to optimize the accuracy of off-policy evaluation in settings where rewards are missing or noisy. By deriving variance‑optimal annotation probabilities for sequential, forward‑monotone protocols, the authors provide a batch‑adaptive implementation that can be applied to real data. Experiments on casenotes from a homelessness services nonprofit and on human‑preference votes from LMArena demonstrate substantial reductions in RMSE—up to 65% for housing placement and 68% for progress toward a housing application—when using only 40% or more of the full annotation budget.
By Woojin Chae, Ezinne Nwankwo, Haitong Qin, Angela Zhou
arXiv:2502. 10605v4 Announce Type: replace-cross Abstract: Problem definition: Estimating causal effects of interventions is central to policy and operations, but outcome data are often missing or costly to obtain.
By Ezinne Nwankwo, Lauri Goldkind, Angela Zhou
arXiv:2609.38860v1 Announce Type: cross
Abstract: Learning from human preferences is central to large language model (LLM) alignment, but human preference annotation is costly. Active preference lear...
By Zhongman Du, Huiming Zhang, Haodong Zhu, Baochang Zhang
The paper explores how data from fixed A/B tests can guide the deployment of adaptive experiments using contextual bandits. By combining off‑policy evaluation with a controlled warm‑start simulation, the authors rank pre‑specified adaptive and non‑adaptive policies using doubly robust estimators. Experiments on synthetic trials and real benchmarks show that adaptive, context‑aware policies outperform fixed allocations when heterogeneity exists, but offer little advantage otherwise.
By Jo\~ao Victor Ferreira Alves, Eduardo Rocha Laurentino, Gustavo de Oliveira Kanno, Thiago Costa Rizuti da Rocha
MInTRL (Minimal Intervention Reinforcement Learning) expands exploration in on-policy reinforcement learning by inserting sparse, local corrections into rollouts via a judge-intervention policy. These interventions replace erroneous suffixes and immediately return control to the main policy, allowing the agent to explore beyond its natural trajectory while maintaining on-policy data. The method uses a sequence-level advantage-regression objective, avoiding importance sampling, and demonstrates superior performance on math and code benchmarks compared to standard on-policy and off-policy baselines.
By Mingyu Chen, Yefan Tao, Gerald Friedland, Xuezhou Zhang, Chris Kong
arXiv:2606. 20206v1 Announce Type: cross Abstract: In offline Reinforcement Learning, immediate rewards in logged batch data are often unobserved due to sparse or irregular record-keeping, or censored beyond certain reward values.
By Ziheng Wei, Annie Qu, Rui Miao
arXiv:2607. 20083v1 Announce Type: cross Abstract: Post-training with evaluator feedback on policy-induced samples serves as a major mechanism for improving large language models.
By Beining Wang, Weihang Su, Hongtao Tian, Hao Kong, Tao Yang, Ting Yao, Qingyi Pan, Yueyue Wu, Qingyao Ai, Min Zhang, Yiqun Liu
The paper introduces a framework for hypothesis testing that combines inexpensive AI judgments with selective human verification to control type‑I and type‑II errors while minimizing cost. It derives an information‑theoretic lower bound on the minimum cost and proposes the SCALE policy, a sequential, cost‑aware strategy that adapts AI scoring and human escalation. SCALE is proven valid for finite samples and asymptotically matches the lower bound, achieving significant savings when both AI and human inputs are valuable.
By Dae Woong (David), Ham, Xuejun Zhao, Stefanus Jasin, Fenghua Yang
arXiv:2606. 18327v1 Announce Type: cross Abstract: Language models (LMs) that faithfully describe their own behavior can more easily be audited, understood, and trusted by users.
By Itamar Pres, Laura Ruis, Melat Ghebreselassie, Belinda Z. Li, Jacob Andreas
arXiv:2605. 28882v2 Announce Type: replace-cross Abstract: With the rapid advancement of large language models, evaluating human-likeness in open-ended conversation has become increasingly important.
By Yihang Lin, Yunze Gao, Zeyang Lin, Dongbo Li, Kun Peng, Yue Liu
arXiv:2607. 10694v1 Announce Type: cross Abstract: We study the problem of optimal continual fine-tuning for a pre-trained Foundation Model deployed at a resource-limited device.
By Thomas Tsouparopoulos, Iordanis Koutsopoulos
The paper introduces ProSE, a framework for AI assistants that generate proposals while considering users’ bounded rationality and evaluability constraints. It proposes a KL‑regularised bounded‑rational binary response model and a depth‑2 Bayes‑adaptive planner, “ProSE‑Plan,” which scores proposals by expected responses and resulting belief updates. Experiments on graph simulations show that “ProSE‑Plan” outperforms evaluability‑unaware and myopic baselines, especially when evaluation cost is high, and that informative probes are crucial for effective assistance.
By Yifan Zhu, Sammie Katt, Samuel Kaski