What is the name of the AI system from DeepMind that achieved superhuman performance in games such as chess and 'Go'?

The reference to the AI system from DeepMind that achieved superhuman performance in games such as chess and 'Go' pertains to AlphaZero. This powerful AI program was designed to master games through reinforcement learning, starting from random play and learning from its own experiences, without relying on pre-existing human knowledge of the games.

AlphaZero represents a significant advancement in AI as it demonstrated the ability to learn strategies that surpassed human capabilities in these complex games. Specifically, it was able to learn and outperform state-of-the-art chess engines and the best versions of AlphaGo, which was specifically tailored for the game of 'Go'. While AlphaGo is the system that first showcased DeepMind’s prowess specifically in 'Go', AlphaZero generalizes this approach to other board games, making it a broader representation of DeepMind's advancements in AI. This distinction highlights AlphaZero's capability to learn and excel beyond what was previously achieved specifically in chess and 'Go'.

The other options refer to different aspects or iterations related to DeepMind's work in gaming and AI, but AlphaZero is the one that truly encapsulates the innovative approach and superhuman performance in multiple games.