7 papers · 1 filter
Compositional Monte Carlo Tree Diffusion for Extendable Planning
Jaesik Yoon, Hyeonseo Cho, Sungjin Ahn
Monte Carlo Tree Diffusion (MCTD) integrates diffusion models with structured tree search to enable effective trajectory exploration through stepwise reasoning. However, MCTD remai…
Loopholing Discrete Diffusion: Deterministic Bypass of the Sampling Wall
Mingyu Jo, Jaesik Yoon, Justin Deschenaux +2
Discrete diffusion models offer a promising alternative to autoregressive generation through parallel decoding, but they suffer from a sampling wall: once categorical sampling occu…
CrafterDojo: A Suite of Foundation Models for Building Open-Ended Embodied Agents in Crafter
Junyeong Park, Hyeonseo Cho, Sungjin Ahn
Developing general-purpose embodied agents is a core challenge in AI. Minecraft provides rich complexity and internet-scale data, but its slow speed and engineering overhead make i…
Fast Monte Carlo Tree Diffusion: 100x Speedup via Parallel Sparse Planning
Jaesik Yoon, Hyeonseo Cho, Yoshua Bengio +1
Diffusion models have recently emerged as a powerful approach for trajectory planning. However, their inherently non-sequential nature limits their effectiveness in long-horizon re…
Adaptive Inference-Time Scaling via Cyclic Diffusion Search
Gyubin Lee, Truong Nhat Nguyen Bao, Jaesik Yoon +4
Diffusion models have demonstrated strong generative capabilities across domains ranging from image synthesis to complex reasoning tasks. However, most inference-time scaling metho…
Latent Veracity Inference for Identifying Errors in Stepwise Reasoning
Minsu Kim, Jean-Pierre Falet, Oliver E. Richardson +5
Chain-of-Thought (CoT) reasoning has advanced the capabilities and transparency of language models (LMs); however, reasoning chains can contain inaccurate statements that reduce pe…
Monte Carlo Tree Diffusion for System 2 Planning
Jaesik Yoon, Hyeonseo Cho, Doojin Baek +2
Diffusion models have recently emerged as a powerful tool for planning. However, unlike Monte Carlo Tree Search (MCTS)-whose performance naturally improves with inference-time comp…