Action Selection in Multiplayer Hidden-Information Games with Expected Target Entropy
Expected-target-entropy action selection transfers well to target-identification games and exposes when task-progress costs are needed.
Game AI
Efficient Multi-Guard Patrol through Extended Composite Potential Fields
Lazy Field Evaluation, Voronoi Territorial Coordination, and Shared Novelty Patrol extend CPF for efficient coordinated stealth-game guard patrols.
Generate Diverse Skills with Large Language Models
A browser-based roguelike auto-battler and evaluation framework for intent-aware procedural skill generation with local LLMs.
Path2Patrol: Path-Conditioned Guard Generation in Stealth Games through MAP-Elites
A designer-centric quality-diversity framework for generating valid, diverse, and controllable stealth-game guard patrols.
(Perlin) Noise as AI coordinator
Noise fields as a practical AI coordinator for smooth local behavior and globally coordinated variety across space and time.
High Dimensional Procedural Content Generation
A general HDPCG framework that augments geometry with gameplay dimensions (layer/time) for controllable generation and high-dimensional validation.
How Far Can We Go with Pixels Alone? A Pilot Study on Screen-Only Navigation in Commercial 3D ARPGs
A vision-only navigation baseline for commercial 3D ARPGs using visual affordances and a simple finite-state controller.