人工智能在策略游戏中胜过人类(上)

Trust no one.

不要相信任何人。

AI outplays most humans in a game of negotiation and double-dealing.

人工智能在需要谈判和策略的游戏中胜过了大多数人类。

Backgammon was an easy win.

五子棋，轻松获胜。

Chess, harder.

国际象棋，艰难取胜。

Go, harder still.

围棋，更加艰辛。

But for some aficionados it is only now that artificial intelligence (AI) can truly say it has joined the game-playing club - for it has proved it can routinely beat humans at Diplomacy.

但对于一些人工智能（AI）的狂热爱好者来说，AI现在才算是真正加入了游戏俱乐部——因为它已被证明可以在《强权外交》游戏中击败人类。

For those unfamiliar with the game, its board is a map of Europe just before the first world war (except that, for no readily apparent reason, Montenegro is missing).

为不熟悉这个游戏的人解释一下，桌游《强权外交》的棋盘是一张第一次世界大战前的欧洲地图（由于某些原因，地图上没有标注黑山）。

Participants, seven ideally, each take on the role of one of the Great Powers: Austria, England, France, Germany, Italy, Russia and Turkey.

该游戏推荐七个人共同参与，参与者每人代表一个大国：奥地利、英国、法国、德国、意大利、俄罗斯和土耳其。

Each has armies and navies, and geographically based resources to support them, and can use its forces to capture the territory of neighbors, thus gaining the means to raise more forces while depriving others of the same.

每个国家都有自己的陆军和海军，以及用于支持本国的地理资源，每个国家都可以利用自己的军队占领邻国的领土，从而组建更庞大的军队，同时削弱其他国家的军队。

The trick is that, at least at the beginning, players will get nowhere without making agreements to collaborate - yet they are not bound by the game's rules to keep to these agreements.

游戏的诀窍在于，玩家至少要在游戏开始时与其他玩家达成合作协议，否则将无法取得进展——但他们是否必须遵守这些协议则不受游戏规则的约束。

Only when orders for the movement of troops and vessels, which have to be written down, are revealed, does a player discover who really is a friend, or an enemy.

只有玩家对部队和船只移动下达指令（必须要写在纸面上）时，其他玩家才会明白谁是真正的朋友，谁是真正的敌人。

Cicero, a program devised by a group of Mark Zuckerberg's employees who dub themselves the Meta Fundamental AI Research Diplomacy Team, proved an adept pupil.

由马克·扎克伯格的员工组成的“Meta基础人工智能研究强权外交游戏团队”设计了一个名为“Cicero”的项目，事实证明，Cicero是一个成熟的程序。

As the team describe in Science, when they entered their creation into an online Diplomacy league, in which it played 40 games, it emerged as one of the top 10% of players - and no one rumbled that it was not human.

正如该团队在《科学》杂志上所描述的那样，他们让Cicero参加了一个网络《强权外交》游戏联赛，进行了40场比赛，排名玩家榜前10%——没有人发现对手不是人类。

In all past AI game-playing projects the program has learned by reinforcement.

以前所有的人工智能游戏项目都是通过强化学习进行训练的。

Playing repeatedly against itself or another version of itself, it acts first at random, then more selectively.

强化学习要求程序反复与自身或该程序的其它版本进行对抗，先是随机采取行动，然后更有选择性地行动。

Eventually, it learns how to achieve the desired goal.

最终，程序将学会实现预期目标的方法。

Cicero was taught this way, too.

Cicero也接受了同样的训练。

But that was only part of its training.

但这只是其训练的一部分。

Besides having the reasoning to plan a winning strategy, a successful Diplomacy player must also possess the communicative ability to implement it.

一个优秀的《强权外交》玩家除了需要制定制胜战略的推理能力，还必须具备实施该战略的沟通能力。