This paper proposes the Bayesian Expected Uncertainty Reduction (B-EUR) model, which formalizes the value of trying a candidate design action as its expected reduction of epistemic uncertainty about action--outcome relations. The model addresses one part of the Uncertainty Driven Action (UDA) model's open question concerning how changes in uncertainty perception determine action selection.
The paper introduces ACTION-BED, a Bayesian experimental design framework that replaces traditional uncertainty‑reduction objectives with expected future loss (EFL) on downstream actions. By reformulating EFLs into singly intractable objectives, ACTION-BED enables joint optimisation of design and action policies using stochastic gradients without explicit posterior or marginal likelihood estimation. This approach simplifies learning of design policies and allows easy customization to various downstream tasks and loss functions.
By Tom Rossa, Angus Phillips, Tom Rainforth
arXiv:2606. 19353v1 Announce Type: cross Abstract: In-Context Learning (ICL) allows LLMs to adapt to new tasks from a few demonstrations, but its reliability remains a concern: predictions are highly sensitive to both prompt design and the model's ability to understand the context, obscuring whether failures arise from data properties or model limitations.
By Jinseok Chung, Minkyoung Song, Hyunji Jung, Namhoon Lee
arXiv:2607. 14817v1 Announce Type: cross Abstract: Current evaluation of epistemic uncertainty relies on tasks such as out-ofdistribution detection and active learning.
By Jakub Paplh\'am, Willem Waegeman, Eyke H\"ullermeier, Vojt\v{e}ch Franc
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...
By David Chen, Michael Evans, Xinwei Li, Prateek Bansal, David J. Nott
arXiv:2606. 19328v1 Announce Type: cross Abstract: Preference-based RL provides an approach to learning reward models from pairwise comparisons of behaviors, bypassing the need for explicit reward design.
By Mohamed Nabail, Leo Cheng, Jingmin Wang, Nicholas Rhinehart
The paper introduces Generalised Bayesian Optimal Experimental Design (GBOED), extending generalised Bayesian inference to experimental design. It replaces the likelihood in Bayesian updates with a loss function, deriving a new acquisition function called Gibbs expected information gain. Empirical results show that GBOED improves robustness against outliers and mis-specified outcome noise distributions.
By Yasir Zubayr Barlas, Sabina J. Sloman, Samuel Kaski
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
arXiv:2607. 26845v1 Announce Type: new Abstract: Inference-time thinking improves the performance of large language models, but aggregate outcomes do not reveal whether models use available evidence more effectively or seek information that could improve future decisions.
By Hua-Dong Xiong, Xinyuan Yan, Ji-An Li, Jingming Xue, Marcelo G. Mattar, Robert C. Wilson
arXiv:2610.32514v1 Announce Type: new
Abstract: Thompson sampling is a widely used strategy for sequential decision-making under uncertainty, offering a principled way to balance exploration and expl...
By Michal Kobiela, Michael U. Gutmann
Inference-time thinking improves the performance of large language models, but aggregate outcomes do not reveal whether models use available evidence more effectively or seek information that could improve future decisions. We distinguish these responses by measuring action preference, thinking length, and reported confidence under matched uncertainty.
The paper introduces Uncertainty‑Normalized Margin DPO (UNM‑DPO), a method that augments direct preference optimization with strength‑dependent margins and a learned prompt scale derived from a heteroskedastic Bradley‑Terry model. Two training objectives—advantage‑only (AO) and whole‑residual (WR)—compare implicit rewards from response log‑probability ratios, with WR enabling identification of the prompt scale and a practical learning procedure. Building on WR, ULNM‑DPO‑WR further normalizes rewards by response length, and experiments on HelpSteer2/3 and AlpacaEval show higher win rates, mean rewards, and shorter responses compared to baseline DPO and SimPO.
By Sadegh Khorasani, Petrus Mikkola, Matthias Grossglauser