The paper introduces DiBS, a novel diffusion model-guided approach for branch selection in solving Sudoku, a constraint satisfaction problem. It enhances symbolic solvers by using a diffusion model to guide branch ordering, ensuring completeness while mitigating long-tail search issues.
This content discusses Deep Dyna-Q, an approach that integrates planning for dialogue policy learning in conversational AI systems. The focus is on optimizing the task-completion process through spoken interaction with AI.
FREIA is a novel reinforcement learning algorithm designed to enhance LLMs for unsupervised reasoning, addressing the lack of adaptability in existing methods. It employs Free Energy-Driven Reward (FER) to balance consensus and exploration, and Adaptive Advantage Shaping (AAS) to adjust learning signals. FREIA outperforms unsupervised baselines across various reasoning tasks, particularly in mathematical reasoning.
This research introduces OP-Mix, a novel algorithm for efficient data mixing throughout the entire lifecycle of language model training. It addresses the challenge of combining diverse data sources for pretraining, continual learning, and adaptation, proposing a unified online decision-making solution.
This paper introduces novel rerooter designs for the $\sqrt{\text{LTS}}$ algorithm, addressing the scalability limitations of explicit subgoal generation in subgoal-based policy tree search. These designs implicitly decompose problems, enabling scalable allocation of search effort.
LightLDA is an innovative algorithm for efficiently building large topic models, even on modest compute clusters. It optimizes scalability and the processing of large data volumes, making advanced topic modeling more accessible.