谷歌DeepMind近日推出了一款名为SIMA（可扩展、可指导、多世界智能体）的新型人工智能系统，该系统旨在通过训练学习游戏技能，使其游戏表现更接近人类玩家。SIMA的开发标志着人工智能在模仿人类行为和适应复杂环境方面迈出了新的一步。
据The Verge报道，SIMA目前仍处于研究阶段，但已经在《Valheim：英灵神殿》、《无人深空》和《模拟山羊》等多款不同类型的游戏上进行了训练。它的设计目标是能够适应各种游戏，包括那些没有明确结束线或拥有广阔开放世界的挑战性游戏。
不同于传统的人工智能，SIMA强调学习和适应能力，能够在游戏过程中不断学习新策略，而不是仅仅执行预设的指令。这种智能体的灵活性和学习能力为未来人工智能在更广泛领域的应用打开了新的可能，例如在解决复杂问题或应对不确定环境时，能够展现出更接近人类的决策过程。
谷歌DeepMind的这一创新举措再次证明了他们在人工智能研究领域的领先地位，同时也引发了关于人工智能与人类智能融合的深入思考。随着SIMA等项目的进展，我们有望见证AI在游戏甚至其他领域的表现越来越接近人类，为科技进步带来新的突破。
英语如下：
**News Title:** “Google DeepMind Unveils SIMA: A Breakthrough in AI, Training to Play Games Like Humans”
**Keywords:** Google DeepMind, SIMA Agent, Game Training
**News Content:** Google DeepMind has recently introduced a novel artificial intelligence system named SIMA (Scalable, Instructable, Multi-World Agent), designed to learn game skills through training, mimicking human players’ performance. This development signifies a new milestone in AI’s ability to replicate human behavior and adapt to complex environments.
According to The Verge, SIMA is still in the research phase but has already been trained on a variety of games, including ‘Valheim:英灵神殿’, ‘No Man’s Sky’, and ‘ Goat Simulator’. Its objective is to adapt to diverse games, including those with no clear endgame or vast open-world challenges.
Distinguishing itself from conventional AI, SIMA emphasizes learning and adaptability. It continually learns new strategies during gameplay rather than simply executing pre-programmed instructions. This flexibility and learning capacity open up new possibilities for AI applications in broader domains, potentially enabling more human-like decision-making processes when dealing with complex problems or uncertain environments.
Google DeepMind’s innovative move reaffirms their prominence in AI research and sparks deeper contemplation on the integration of AI and human intelligence. As projects like SIMA progress, we can anticipate AI’s performance in gaming and potentially other sectors becoming increasingly human-like, ushering in new breakthroughs in technological advancement.
【来源】https://www.theverge.com/2024/3/13/24099024/google-deepmind-ai-agent-sima-video-games
Views: 1