Tag
A HEC Montréal/MILA paper proposes fair policy optimization for major-minor weakly coupled MDPs, replacing the utilitarian objective with monotone concave fairness functions and introducing a count-proportion-based deep RL approach with a priority-based sampler, validated on machine replacement and NYC taxi pricing/relocation tasks.
Introduces PrimeScientist, a method for strategic allocation of research effort in autonomous research agents using an adaptive MCTS-based policy to improve research quality and sample efficiency under resource constraints.
This paper resolves an open problem by showing that a single algorithm achieves optimal switching regret for every S against an oblivious adversary in multi-armed bandits.
This paper introduces Cros, a risk-constrained stopping layer for sequential clinical diagnosis agents that ensures finite-sample guarantees for diagnostic error and coverage, with evaluation on a benchmark derived from MIMIC.
This paper proposes ProSE-Plan, a Bayes-adaptive planner that improves AI assistance by considering user evaluability under bounded rationality, using proposals as probes to learn preferences and enhance decision-making.
The paper proposes RATTL, a framework that adjusts an agent's caution based on its Bayesian belief uncertainty using Wasserstein distance for safe sequential decision making, applicable to LLM-based systems.
This paper introduces a reinforcement learning approach for training agents to wait strategically in sequential decision-making tasks, balancing task performance with resource conservation. Experiments show significant waiting behaviors across household and continuous-state environments.
Introduces Rushes, a large-scale dataset of human engagement preferences in AI-generated branching narratives, revealing that current LLMs like GPT-5 fail to outperform simple baselines in predicting user choices, highlighting the need for personalized alignment.
This paper investigates whether induced emotions can bias the sequential decision-making of LLMs using the Iowa Gambling Task as a testbed. The authors find that while emotional induction does not significantly affect average decision dynamics, anger can reduce penalty sensitivity and early-stage exploration.
This paper studies stochastic linear bandits where the agent only observes a random subset of action coordinates, proving that sublinear regret is possible when actions have low intrinsic dimension, and proposes the TOFU-POV algorithm with theoretical guarantees.
Introduces a three-stage training paradigm to internalize world model planning in LLM agents, enabling future-aware decision-making. Outperforms baselines on search and mathematical reasoning tasks.
This paper introduces Agent-Authored World Modeling (AAWM), a training procedure that constructs world-model supervision based on the policy's own decision needs rather than next-observation prediction, aligning the learning objective with the dynamics required for effective decision-making.
This paper introduces Agentick, a unified benchmark for evaluating general sequential decision-making agents across RL, LLM, and VLM paradigms. It provides 37 procedurally generated tasks and reveals that no single approach currently dominates, highlighting significant room for improvement in agent autonomy.
This paper introduces PRISM, a framework that integrates Vision-Language Models and Large Language Models through a dynamic question-answering pipeline to improve sequential decision-making in embodied AI tasks.