Bellman
PulseAugur coverage of Bellman — every cluster mentioning Bellman across labs, papers, and developer communities, ranked by signal.
2 day(s) with sentiment data
-
New algorithm improves zero-sum POSG solving with conservative policy updates
Researchers have developed a new approximate policy-iteration algorithm designed for finite-horizon zero-sum partially observable stochastic games (zs-POSGs). This method maintains a policy for each player and updates t…
-
New RL method tackles diffusion control problems with log-regret bounds
A new research paper introduces a novel approach to reinforcement learning for controlled diffusions, specifically addressing issues with time discretization in high-frequency control problems. The proposed method, term…
-
New TD framework calibrates black-box simulators using Bellman orthogonality
Researchers have introduced a new temporal-difference (TD) framework designed to calibrate black-box simulators. This framework utilizes one-step transitions from a target system and noisy simulator-generated continuati…
-
New FSPO controller enhances LLM reinforcement learning post-training
Researchers have developed FSPO, a novel feedback-state controller designed to improve the post-training reinforcement learning process for large language models (LLMs). FSPO addresses key challenges in adapting LLM beh…
-
New Bellman-Certified Rounding method improves sparse policy deployment in MDPs
Researchers have developed a new method called Bellman-Certified Rounding to address the challenge of deploying sparse policies in Markov decision processes (MDPs). This technique aims to retain a significant portion of…
-
New AI agent optimizes Alzheimer's diagnosis costs
Researchers have developed SCOPE-AD, a novel sequential planning agent designed for cost-aware diagnosis of Alzheimer's disease. This agent utilizes an ordinal-belief model to represent uncertainty across the spectrum f…
-
New AI Framework Extends Maximum Occupancy Principle for Uncertain Environments
A new research paper introduces a belief-based extension to the Maximum Occupancy Principle (MOP), a framework for intrinsic motivation in AI agents. This extension allows agents to operate in partially observable envir…
-
New research tackles catastrophic forgetting in continual learning models · 8 sources tracked
Multiple research papers explore the complex dynamics of continual learning in neural networks, focusing on how models retain or forget information over time. Studies investigate methods to mitigate catastrophic forgett…
-
New research explores advanced reinforcement learning techniques and agent frameworks · 10 sources tracked
Multiple research papers released on arXiv introduce novel frameworks and techniques for advancing reinforcement learning (RL) and agentic AI. AgentFly proposes a unified resource system to scale agentic RL by treating …
-
New theory unifies AI communication, control, and decision-making
A new paper proposes a mathematical framework for pragmatic information theory, aiming to unify communication, control, and decision-making. The theory introduces the concept of isoteleia, which formalizes the idea that…
-
New inference method improves temporal-difference learning accuracy
Researchers have developed a new method called Self-Normalized Inference for Constant-Stepsize Temporal-Difference Learning. This technique allows for more accurate inference from single Markov trajectories by accountin…
-
New Wasserstein Policy Gradient Method for LQ Control Problems
Researchers have developed a Wasserstein policy gradient (WPG) method for entropy-regularized linear-quadratic (LQ) control problems. This approach leverages the fact that unrestricted LQ control problems have linear-Ga…
-
New research tackles credit assignment for LLM agents · 2 sources tracked
Two new research papers from arXiv explore advanced credit assignment techniques for large language model agents. The first paper, "From Reasoning to Agentic: Credit Assignment in Reinforcement Learning for Large Langua…
-
New auction system optimizes ad timing in LLM conversations
Researchers have developed LLM-OSDA, a novel dynamic auction mechanism for native advertising within multi-turn conversations. This system integrates optimal stopping theory with an auction framework to determine both t…
-
New robust Q-learning algorithm tackles mean-field control with Wasserstein uncertainty
Researchers have developed a new robust Q-learning algorithm designed for mean-field control problems. This algorithm addresses challenges posed by Wasserstein uncertainty in common noise laws by integrating a quantizat…
-
Withdrawn paper details novel continuous-time policy evaluation method
A research paper, now withdrawn, proposed a novel method for continuous-time policy evaluation called High-Order Generator Regression. This technique aims to improve upon the standard Bellman baseline by using multi-ste…
-
Yann LeCun clarifies technical definition of 'world models' in AI
Yann LeCun shared a technical discussion regarding the term "world models" in AI. He clarified that in control theory and the context of Markov Decision Processes (MDPs), "world models" specifically refers to transition…