Search:
Match:
2 results
Paper#Cosmology🔬 ResearchAnalyzed: Jan 3, 2026 18:28

Cosmic String Loop Clustering in a Milky Way Halo

Published:Dec 29, 2025 19:14
1 min read
ArXiv

Analysis

This paper investigates the capture and distribution of cosmic string loops within a Milky Way-like halo, considering the 'rocket effect' caused by anisotropic gravitational radiation. It uses N-body simulations to model loop behavior and explores how the rocket force and loop size influence their distribution. The findings provide insights into the abundance and spatial concentration of these loops within galaxies, which is important for understanding the potential observational signatures of cosmic strings.
Reference

The number of captured loops exhibits a pronounced peak at $ξ_{\textrm{peak}}≈ 12.5$, arising from the competition between rocket-driven ejection at small $ξ$ and the declining intrinsic loop abundance at large $ξ$.

Analysis

This paper addresses the challenge of off-policy mismatch in long-horizon LLM reinforcement learning, a critical issue due to implementation divergence and other factors. It derives tighter trust region bounds and introduces Trust Region Masking (TRM) to provide monotonic improvement guarantees, a significant advancement for long-horizon tasks.
Reference

The paper proposes Trust Region Masking (TRM), which excludes entire sequences from gradient computation if any token violates the trust region, providing the first non-vacuous monotonic improvement guarantees for long-horizon LLM-RL.