A High-Throughput Compute-Efficient POMDP Hide-And-Seek-Engine (HASE) for Multi-Agent Operations

Timothy Flavin, Sandip Sen · 2026

Read A High-Throughput Compute-Efficient POMDP Hide-And-Seek-Engine (HASE) for Multi-Agent Operations by Timothy Flavin, Sandip Sen (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-27162.txt.

Description

Flavin and Sen build a fast C++ engine for hide-and-seek games played by many agents under partial observability. Such Dec-POMDP settings are costly for reinforcement learning, which needs many samples. Their Hide-And-Seek-Engine uses data-oriented design, 64-byte cache-line alignment to cut false sharing, and a zero-copy bridge to PyTorch with pinned memory. The goal is cheap, high-throughput steps for human-AI joint operations.

Catalog record

CREATORTimothy Flavin, Sandip Sen
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE43241
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.27162

Browse the full library · About the Tower of Babel