General game testing relies on the use of human play testers, play test
scripting, and prior knowledge of areas of interest to produce relevant test
data. Using deep reinforcement learning (DRL), we introduce a self-learning
mechanism to the game testing framework. With DRL, the framework is capable of
exploring and/or exploiting the game mechanics based on a user-defined,
reinforcing reward signal. As a result, test coverage is increased and
unintended game play mechanics, exploits and bugs are discovered in a multitude
of game types. In this paper, we show that DRL can be used to increase test
coverage, find exploits, test map difficulty, and to detect common problems
that arise in the testing of first-person shooter (FPS) games.

本研究采用深度强化学习（DRL）技术来实现自学习机制的游戏测试框架，可提高测试覆盖率，发现并解决游戏内的漏洞及其他问题，特别适用于第一人称射击类型的游戏。

使用深度强化学习增强自动游戏测试

Augmenting Automated Game Testing with Deep Reinforcement Learning

We introduce a new simulation platform called Insight, created to design and
simulate cyber-attacks against large arbitrary target scenarios. Insight has
surprisingly low hardware and configuration requirements, while making the
simulation a realistic experience from the attacker's standpoint. The scenarios
include a crowd of simulated actors: network devices, hardware devices,
software applications, protocols, users, etc. A novel characteristic of this
tool is to simulate vulnerabilities (including 0-days) and exploits, allowing
an attacker to compromise machines and use them as pivoting stones to continue
the attack. A user can test and modify complex scenarios, with several
interconnected networks, where the attacker has no initial connectivity with
the objective of the attack. We give a concise description of this new
technology, and its possible uses in the security research field, such as
pentesting training, study of the impact of 0-days vulnerabilities, evaluation
of security countermeasures, and risk assessment tool.

介绍了一个新的模拟平台 Insight，用于设计和模拟针对大规模任意目标场景的网络攻击，包括模拟漏洞和利用，可应用于渗透测试、漏洞评估和风险评估等安全研究领域。