arXiv 2510.24698

ParallelMuse: Agentic Parallel Thinking for Deep Information Seeking

By Baixuan Li, Dingchu Zhang, et al.

Published 2025-10-28

Wiki summary

Explore the paper's summary, context, and related research on Papiers.

Parallel thinking expands exploration breadth, complementing the deep exploration of information-seeking (IS) agents to further enhance problem-solving capability. However, conventional parallel thinking faces two key challenges in this setting: inefficiency from repeatedly rolling out from scratch, and difficulty in integrating long-horizon reasoning trajectories during answer generation, as limited context capacit…

View the original paper on arXiv