DeepMind's newest report reveals that it has developed a programme called 'MuZero' that can play and win complex games without knowing the rules, thereby requiring less data-training. 

The subsidiary of Google, DeepMind is famous for creating groundbreaking advancements using reinforcement learning to teach artificial intelligence (AI) systems to master popular games such as chess and Atari video games. However, in all previous instances, the algorithms were 'taught' the rule of the games. However, MuZero has mastered these games, and in some cases, even performed better - without being 'taught' the rules. 

MuZero's performance has been enhanced because of a principle called "look-ahead search." In this approach, MuZero assesses the possible potential moves, learning from how the opponent responds. For example in chess, by prioritising the most relevant and most common manoeuvres by retaining lessons from successful gambits and eliminating the moves that failed, MuZero eliminates the hundred other potential moves. 

"For the first time, we actually have a system that is able to build its own understanding of how the world works and use that understanding to do this kind of sophisticated look-ahead planning that you've previously seen for games like chess," said DeepMind's principal research scientist David Silver. MuZero can "start from nothing, and just through trial and error, both discover the rules of the world and use those rules to achieve kind of superhuman performance."

The potential of MuZero lies beyond just games, according to Silver. They're already working on video compression, a daunting task that has varying video formats and various compression methods. The team has seen the progress of 5% improvement. This means a lot to the company as it handles cache from YouTube, where billions of hours' worth content are updated and viewed daily. Further, the algorithm's creators are also exploring possibilities into robot programming and protein architecture design, which holds promise for personalized production of drugs.