AI分析发现了奥斯卡赢家的秘诀？

On the surface, The Godfather, The Sixth Sense and Little Miss Sunshine appear to have little in common.

从表面上看，电影《教父》（The Godfather）、《灵异第六感》（The Sixth Sense）和《阳光小美女》（Little Miss Sunshine）似乎并没有什么共同之处。

But even though these films belong to different genres and have very different plots, technically they have the same "emotional arc" - a journey of highs and lows.

不过，尽管这些电影属于不同的类型，情节也大相径庭，但实际上他们有着相同的“情感弧”——一段跌宕起伏的历程。

Using artificial intelligence, we analysed more than 6,000 scripts from the past 80 years and discovered all films fall within six emotional arcs.

利用人工智能，我们分析了过去80年间的6000多个剧本，发现所有的电影不外乎有六种情感弧类别。

These include the emotional rise of "rags to riches" films such as The Shawshank Redemption and the rise and fall of "man in a hole" films such as Who Framed Roger Rabbit.

其中包括了“麻雀变凤凰”类电影中先抑后扬的情感强化与升华，如《肖申克的救赎》（The Shawshank Redemption），还有“绝处逢生”的励志片中包含的情感起伏，如 《谁陷害了兔子罗杰》（Who Framed Roger Rabbit）。

But which of these are the most successful, critically and commercially?

但是，从商业眼光和业界评论来看，哪种类型是最成功的呢？

Tragedies, which depict a continuing emotional fall, appear to receive the highest number of Oscar nominations per film.

悲剧——凭借对持续跌落的情感走向的描绘，似乎斩获了最多次奥斯卡提名。

This new technology means scientists may be able to give the film industry the tools to understand their viewers and work out what they really want to see on screen.

人工智能这项新技术的运用意味着科学家们或许能够为电影业提供一些帮助，让行业人员了解观众的口味，明白他们真正想在荧幕上看到什么。

Critical acclaim

广受好评

The tragedies we studied won an average of 2.14 Oscars per film.

在我们的研究中，平均每部悲剧电影就有2.14项奥斯卡奖加身。

For example, political thriller The Constant Gardener was nominated for four Oscars, winning the best supporting actress award for Rachel Weisz, in 2006.

比方说，2006年，政治惊悚片《不朽的园丁》（The Constant Gardener）就获得了四项奥斯卡奖提名， 演员Rachel Weisz也因此片一举夺得最佳女配角奖。

Telling the story of a British diplomat in Kenya uncovering the mystery behind his wife's death, it follows a downward emotional arc.

该片讲述了一名英国外交官在肯尼亚揭开妻子死亡之谜的故事。它就遵照了下向的情感弧模式。

But while they may be critically acclaimed, these "riches to rags" films often fail to set the box office alight.

不过，尽管这类“凤凰变麻雀”的电影可能会广受业界好评，却常常票房惨淡。

Consider Lawrence of Arabia, one of the most critically acclaimed films of all time, which received seven Oscars in 1963, including the best picture award. It initially flopped at the box office, making close to $6m (£4.7m) compared with a budget of $15m (£11.8m).

以《阿拉伯的劳伦斯》（Lawrence of Arabia）为例，它在1963年斩获了包括最佳影片奖在内的7项奥斯卡大奖，成为有史以来最受好评的电影之一。然而，与1500万美元（约合1千180万英镑）的预算相比，它在初上映时只收获了近6百万美元（约合470万英镑）的惨淡票房。

Another example is Girl, Interrupted which earned Angelia Jolie the best supporting actress win in 2000 but failed to achieve box office returns.

另一个例证是《移魂女郎》（Girl, Interrupted），安吉丽娜·朱莉凭借此片赢得了2000年的奥斯卡最佳女配角奖，但是该片未能获得相应的票房回报。

When it comes to financial success, it is actually "man in a hole" films - an emotional fall followed by a rise - that come out on top, regardless of genre or production budget.

谈到卖座电影，当属“绝处逢生”的“励志片”（man in a hole），情感上的先落后起，带来的是票房的脱颖而出，无论这些电影的类型如何、制作预算多少。

How it works

运作原理

Using a filtering process, we selected 6,147 films released between 1935 and 2018 and added up Oscar nominations in all categories.

通过过滤处理，我们挑选出了1935年至2018年间上映的6147部电影，涵盖了奥斯卡提名的所有类别奖项。

Our machine learning algorithm then split each film script into sentences.

然后，我们的机器学习算法将每个电影剧本分解成了单个的句子。

Each sentence was given a score calculated by averaging the sentimental value of each word.

通过将句子中各单词的情感数值取平均值，每个句子都会得到一个相应的分数。

An emotionally negative term scored minus one, an emotionally neutral term scored nought, and an emotionally positive term scored one point.

消极情绪得-1分，中性情绪0分，积极情绪得1分。

The total sentiment for each film was accumulated and overlaid on to the span of the film, to map the emotional arc.

各种情绪在影片中不断累积，贯穿整部电影，由此形成了“情感弧”。

Films were then grouped based on the similarity of their emotional arc. By doing this, scientists discovered that films, like novels, all fitted within six standard emotional journeys:

根据“情感弧”的相似性，我们对电影进行了分类。由此科学家们发现，与小说类似，电影也符合六种情感历程标准：

• Rags to riches: a continuing emotional rise (egThe Shawshank Redemption, Groundhog Day, The Nightmare Before Christmas)

• “麻雀变凤凰”型：情绪不断高涨，如《肖申克的救赎》（The Shawshank Redemption），《土拨鼠之日》（Groundhog Day），《圣诞夜惊魂》（The Nightmare Before Christmas）；

  
  

• Riches to rags: a continuing emotional fall (egPsycho, Monty Python and the Holy Grail, Toy Story 3)

• “凤凰变麻雀”型：情绪不断跌落，如《惊魂记》（Psycho），《巨蟒与圣杯》（Monty Python and the Holy Grail），《玩具总动员3》（Toy Story 3）；

  
  

• Man in a hole: a fall followed by a rise (egThe Godfather, The Lord of the Rings: The Fellowship of the Ring, The Sixth Sense)

• “绝处逢生”型：情绪先抑后扬，如《教父》（The Godfather），《指环王：护戒使者》（The Lord of the Rings: The Fellowship of the Ring），《灵异第六感》（The Sixth Sense）；

  
  

• Icarus: A rise followed by a fall (eg On the Waterfront, Mary Poppins, A Very Long Engagement)

• “伊卡洛斯”型：情绪先扬后抑，如《码头风云》（On the Waterfront），《欢乐满人间》（Mary Poppins），《漫长的婚约》（A Very Long Engagement）；

• 【译注：伊卡洛斯是希腊神话中代达罗斯的儿子，与代达罗斯使用蜡和羽毛造的翼逃离克里特岛时，他因飞得太高，双翼上的蜡遭太阳融化跌落水中丧生，被埋葬在一个海岛上。】

  
  

• Cinderella: "rise-fall-rise" (egRushmore, Babe, Spider-Man 2)

• “灰姑娘”型：情绪呈升-降-升变化，如《青春年少》（Rushmore），《小猪宝贝》（Babe），《蜘蛛侠2》（Spider-Man 2）；

  
  

• Oedipus: "fall-rise-fall" (All About My Mother, As Good as It Gets, The Little Mermaid)

• “俄狄浦斯”型：情绪呈降-升-降变化，如《关于我母亲的一切》（All About My Mother），《尽善尽美》（As Good as It Gets），《小美人鱼》（The Little Mermaid）。

Following the money

票房为王

If man in a hole films are the most profitable, should the entertainment industry just make those?

如果说“绝处逢生”型电影最有赚头，那是不是说娱乐界就只需要生产制作这类电影呢？

The answer is probably not, for several reasons.

答案或许是否定的。并且有好几个方面的原因。

While those films do make the most money, we discovered that the right combination of genre and budget can produce a financially successful movie of any emotional shape.

的确，这样的电影利润最高。但我们发现，如果电影的类型与制作成本相互协调，巧妙结合，不管情感历程如何，都能赢得高票房收入。

For example, the "Icarus" arc - a rise followed by a fall - is good for low budget films, such as 2005's family drama Junebug.

例如，“伊卡洛斯”型情绪历程——即情绪先扬后抑——很适合应用在低成本制作的电影中，如2005年的家庭片《六月虫》Junebug。

However, if you want to shoot a financially successful tragedy, then it helps if it is epic, with a large budget of over $100m (£78m).

然而， 如果想拍摄一部票房喜人的悲剧片，最好能有超过1亿美元（即7800万英镑）的成本，拍出悲壮宏大的效果。

Sad superheroes

“超英”之伤

Superhero films are a perfect example of this.

超级英雄电影就是一个绝佳的例子。

In this genre, man in a hole films become only the second most profitable after epic tragedies, i.e. riches to rags films with plenty of funding.

这种电影中，“绝处逢生”型的利润仅次于史诗般的悲剧，也就是制作资金丰厚的“凤凰变麻雀”型。

However, superhero tragedies seem to attract bigger than average budgets anyway - $186m (£147m) - and bring in an average revenue of $310.5m (£245m).

然而，超级英雄的悲剧片似乎超出了平均预算成本——1亿8600万美元（即1亿4700万英镑），也带来了巨额的平均收益——3亿1050万美元（即2亿4500万英镑）。

Looking at all the mainstream Batman films, you can see the top performers were the epic tragedies - the Dark Knight films.

纵观主流的蝙蝠侠系列电影，你会发现，最为成功的作品都是史诗般的悲剧片——黑暗骑士系列电影。

Sci-fi, mysteries, and thrillers with happy endings  - the rags to riches shape - do not tend to do well at the box office.

一些结局幸福、“麻雀变凤凰”型的科幻片、悬疑片和恐怖片往往票房不佳。

This includes the Wachowskis' sci-fi epic Cloud Atlas, which flopped at the box office in 2012, making $27m (£21.3m) in the US against a budget of about $128.5m (£101.5m).

Equally, it is not a good idea to shoot a comedy with a sad ending.

其中包括由沃卓斯基姐弟导演的科幻史诗巨制《云图》（Cloud Atlas），该影片于2012年上映，制作预算高达12.85亿美元，却仅仅赢利2700万美元，惨遭滑铁卢。

同样，结局悲伤的喜剧片也收效较差。

Writing by robot

机器编剧

The entertainment industry tends to make ad hoc, top-down decisions about the content offered to the viewer.

娱乐业界常常自上而下、临时决定提供给观众的内容。

Determining which films get the green light is based primarily on the intuition, expertise and experience of a relatively small group of producers and studio executives, with limited input from focus groups.

一部影片是否可以审核通过，主要取决于相对小部分制片人和制片公司高管的直觉、专业知识和经验，而焦点小组的影响有限。

【译注：focus group：焦点小组，一种定性研究方式，常用于市场研究和政策分析。在调查过程中，通过开放讨论方式，组员表达对调查对象，如：某产品、服务、概念、广告、想法或包装的看法、态度，以预测调查对象的推出效果。（ Wikipedia ）】

Could machines do better?

机器人编剧会更好吗？

Recently, a bot was "employed" to produce a sci-fi film, with nonsensical and sometimes hilarious results, highlighting how far automated scriptwriting still has to go.

最近，一个机器人“受聘”制作一部科幻电影，结果，影片毫无意义，甚至还有点可笑。这突显出自动化编剧仍有一段很长的路要走。

That's because robots still aren't very good at replicating the nuances of human emotion, especially humour, and struggle to write scripts to which people can relate.

这是因为机器人仍然不能很好地复制人类情感的细微差别，尤其是幽默感，而且机器人很难写出让人类产生共鸣的电影脚本。

But we can use data science to enhance the efforts of human scriptwriters and production companies.

但是我们可以应用数据科学来帮助人类编剧和电影制作公司。

【译注：data science，交叉领域学科，通过系统科学分析手段从大量多样数据中提取有效信息。（ Wikipedia ）】

Time will tell whether these new techniques will actually shift decision-making about film content from producers to the viewers.

时间会证明这些新技术是否真的会将电影内容的决策权从制片人手中转移到观众手中。

演职人员表

编译：杨二一、郑可意、邱思怡、李昕瑞

排版：杨二一

指导老师：刘佳

文章来源：BBC

本译文由脑洞清奇的Quriosity呈献

扫码关注我们吧~