HyperMCTS: Hypergraph-Augmented MCTS for Long-Horizon LLM Agents
This work proposes HyperMCTS, a training-free method that augments an ordered MCTS tree with a cross-trajectory hypergraph, allowing outcomes collected under one prefix to inform selection under another while preserving execution histories in the tree.
Ting-Song Xiao, Nithish Balachandar Moudhgalya, Chandrayee Basu et al.
· 0 citations