AI 中文总结
本研究推出 DSLE 平台，将《黑暗之魂》22 场 Boss 战作为智能体基准，定义 DSLE-5 子集并评估多种智能体，发现多数方法难以击败多数 Boss，仅少数早期 Boss 可被击败。
AI 中文摘要
我们推出《黑暗之魂》学习环境（DSLE），这是一个容器化平台，通过 Gymnasium 风格的界面将《黑暗之魂：重制版》的全部 22 场 Boss 遭遇战作为游戏智能体基准任务呈现。DSLE 结合了实时战斗、高维视觉输入和稀疏的终止奖励，每一个环境步骤都是针对运行中的游戏执行的真实动作。为支持可控比较，我们定义了 DSLE-5，这是一个具有代表性的 5 场 Boss 子集，涵盖近战战斗、空间受限竞技场、环境危害战斗、多目标战斗以及快速的最终 Boss 战斗，我们推荐该子集作为基于 DSLE 构建的智能体的起始套件。在 DSLE-5 上，我们评估了随机策略、专家系统、进化基线以及从视觉输入训练的 PPO 和 DQN 智能体。专家系统和进化基线分别击败了游戏的教程 Boss asylum 恶魔，峰值胜率为 63% 和 43%，但这五种方法均未击败 DSLE-5 的其他四个 Boss；在每次运行已耗时数十个挂钟小时的预算内，PPO 和 DQN 未表现出可测量的学习效果（在教程 Boss 上的胜率最高为 0.33%，其他地方为 0%）。更广泛的研究在所有 22 场遭遇战中对进化基线运行，且所有属性均为 50 级优势属性，仅在少数额外的早期游戏 Boss 上获胜，其余均未获胜。失败案例包括在狭窄的多目标遭遇战中 10 秒内死亡，以及长达数分钟的僵局，几乎未造成任何伤害，我们不仅通过胜率，还通过生存时间和造成的伤害来报告这些失败案例。
英文摘要
We introduce the Dark Souls Learning Environment (DSLE), a containerized platform that presents all 22 boss encounters of Dark Souls: Remastered as game-playing agent benchmarks through a Gymnasium-style interface. DSLE combines real-time combat, high-dimensional visual input, and sparse terminal rewards, with each environment step being a real action executed against the running game. To support controlled comparison, we define DSLE-5, a representative five-boss subset, spanning a melee fight, a spatially constrained arena, an environmental-hazard fight, a multi-target fight, and a fast final-boss fight, that we recommend as the starting suite for agents built on DSLE. On DSLE-5 we evaluate a random policy, an expert system, an evolutionary baseline, and PPO and DQN agents trained from visual input. The expert system and the evolutionary baseline each defeat the Asylum Demon, the game's tutorial boss (63% and 43% peak win rates), but none of the five methods defeats the other four DSLE-5 bosses; PPO and DQN show no measurable learning (at most 0.33% win rate on the tutorial boss, 0% elsewhere) within a budget that already costs tens of wall-clock hours per run. A broader study running the evolutionary baseline across all 22 encounters under advantaged all level-50 stats yields wins on only a handful of additional early-game bosses and leaves the rest unwon. The failure cases range from sub-10-second deaths in cramped, multi-target encounters to minute-long stalemates that inflict almost no damage, and we report them through survival time and damage dealt rather than win rate alone.
CommentsAAAI Conference on Artificial Intelligence and Interactive Digital Entertainment, 2026
