开源火种_火种艾完美的牵线搭桥-CSDN博客

开源火种

人工智能，意见，技术(Artificial Intelligence, Opinion, Technology)

“Bored in the house, in the house I’m bored”, Official Tinder theme song!

Tinder官方主题歌“在房子里闷死了，在我无聊的房子里！”

Content Table:-

内容表：-

What is Tinder?
什么是火种？
How it affected the Dating world?
它如何影响约会世界？
What users are expecting from Tinder?
用户对Tinder的期望是什么？
Possible Tinder recommendation system
可能的火种推荐系统
How other dating apps are calculating the “ELO” score?
其他约会应用程序如何计算“ ELO”分数？
Developing Tinder’s ML Model
开发Tinder的ML模型
Is Tinder an excellent matchmaker?
天德是一位出色的媒人吗？

什么是火种？ (What is Tinder?)

Tinder is a mobile dating app that can help you find singles in the local area. “Swipe right if you like her, Swipe left if you don’t” is a linchpin to the company’s success, and the format has been duplicated by numerous contemporaries.

Tinder是一款移动约会应用程序，可以帮助您在当地查找单身人士。 “如果喜欢她，请向右滑动，如果不喜欢，请向左滑动”是公司成功的关键，而且这种格式已经被许多同时代人复制。

Tinder was first launched as a location-based dating app in 2012 within incubator Hatch Labs and join a venture between IAC and Xtreme Labs and now it’s one of the most popular dating apps in the US with about 1.7 Billion swipes per day. Tinder has employed the freemium business model to earn revenue.

Tinder于2012年在孵化器Hatch Labs中作为基于位置的约会应用程序首次发布，并加入了IAC和Xtreme Labs之间的合资企业，现在它已成为美国最受欢迎的约会应用程序之一，每天刷卡约17亿次。 Tinder已采用免费增值业务模式来赚钱。

它如何影响约会世界？ (How it affected the Dating world?)

It went from a “location-based” dating app to a global dating app that is present in 190+ countries in less than 8 years. Comparing to adversaries, the motive of Tinder is not to entertain but to help you get your love life.

它从“基于位置的”约会应用程序转变为在不到8年的时间里已在190多个国家/地区推出的全球约会应用程序。与对手相比，Tinder的动机不是娱乐，而是帮助您获得爱情生活。

Image for post — SimpleTexting SimpleTexting

Tinder is used by 57 Million users and that doesn’t make it one of the most used dating apps. And yet there is something in particular about Tinder that causes it to feel like the characterizing application of the online dating era. In the western world at least — the Tinder seems to be consistent by all accounts humming endlessly out of sight, wherever you go. Its gamified style, flawlessness for easy access, and its legitimate straightforwardness maybe go a portion of the best approach to clarifying its runaway achievement.

Tinder有5700万用户使用，但这并不使其成为最常用的约会应用程序之一。然而，关于Tinder的某些事情特别使它感觉像是在线约会时代的典型应用。至少在西方世界，无论您走到哪里，《火种》似乎都是一致的，无休止地嗡嗡作响。它的游戏风格，易于访问的完美无瑕以及其合法的直率可能是澄清其失控成就的最佳方法的一部分。

Tinder users go on one million dates per week and over 20 Billion matches have been made since Tinder was launched.
Tinder用户每周要进行一百万次约会，自Tinder推出以来，已进行了200亿次匹配。
Active Tinder users log in on average four times a day.
活跃的Tinder用户平均每天登录四次。
95% of Tinder users meet their matches within a week, who thought it would be so easy to secure a date?
95％的Tinder用户在一周内满足他们的比赛要求，谁认为约会很容易？
The year 2017 experienced more couples that met online rather than offline.
2017年，有更多的情侣在网上相识而不是离线。

Talking about numbers, Tinder is estimated to be worth $10 billion, its revenue stood at $1.15 billion in 2019; 56% of total Match Group’s(Parent company)revenue of $2.05 billion. Tinder revenue grew enormously at a CAGR of 123% between 2015 and 2019.

谈到数字，Tinder估计价值100亿美元，2019年的收入为11.5亿美元; Match Group(母公司)总收入20.5亿美元的56％ 。在2015年至2019年之间，火种营收以123％的复合年增长率大幅增长。

Tinder has edged out Netflix and became the highest-grossing global non-gaming app in 2019. According to Sensor Tower, it has maintained that feat as of February 2020. The figures are set at $77.4 million, 42% of which was from the US, 7% from the UK, and 5% from Germany.

Tinder在Netflix上脱颖而出，并成为2019年全球收入最高的非游戏应用程序。根据Sensor Tower的数据，截至2020年2月，它一直保持着这一壮举。这一数字定为7,740万美元，其中42％来自美国。，英国的7％和德国的5％。

Match Group market cap $18.6 billion as of late March 2020

截至2020年3月，火柴集团的市值为186亿美元

用户对Tinder的期望是什么？(What users are expecting from Tinder?)

According to the SimpleTexting survey, looking for a serious, long-term relationship was by far the most preferred choice for netizens. A small ratio of men and women conveyed they were looking for friends, while others use the app to boost their self-esteem.

根据SimpleTexting的调查，到目前为止，寻找一个认真的长期关系是网民最喜欢的选择。一小部分男人和女人表示他们在寻找朋友，而其他人则使用该应用程序来提高他们的自尊心。

可能的火种推荐系统 (Possible Tinder Recommendation System)

There is no confirmed workflow for Tinder’s algorithm, these algorithms powering such platforms are proprietary and the company is least interested in dishing out the private details of their execution but based on the data posted by the company, and trails found by nerds. I made up the accompanying determination.

Tinder的算法尚无确定的工作流程，这些为此类平台提供支持的算法是专有的，该公司最不愿意透露其执行的私人细节，但基于公司发布的数据以及书呆子发现的踪迹。我下定了决心。

Your Tinder matches rely heavily on your data, based on your profile an “ELO” score is calculated that more or less defines the quality and quantity of your matches, from suggesting a sugar daddies to a lame hombre, it decides your fate. You rose in the ranks based on the numbers of right swipes you get, and also right swiper. The more right swipes that person had, the more their right swipe on you meant for your score.

Tinder比赛很大程度上取决于您的数据，根据您的个人资料计算得出的“ ELO ”分数或多或少定义了比赛的质量和数量，从暗示一个甜心小子到a脚的混血儿，它决定了您的命运。您会根据获得的右滑动次数和右滑动次数在排名中上升。此人向右滑动的次数越多，对您的分数向右滑动的次数就越多。

Tinder would then recommend people with the same score, assuming that people with similar opinions would be in approximately the same tier compatibility.

然后，Tinder将推荐具有相同分数的人，前提是具有相似观点的人具有大致相同的层兼容性。

Your desirability a.k.a “ELO” score is heavily dependent on:

您的可取性(也就是“ ELO”分数)在很大程度上取决于：

资料质量 (Quality of Profile)

This is the most important to determine the ELO score. Sometimes it is insurmountable and can debilitate your match. But it depends on your bio, photos, and, settings you’ve chosen.

这是确定ELO分数最重要的。有时，它是无法克服的，可能会使您的比赛失败。但这取决于您的个人简历，照片和选择的设置。

When you post some of your photos, the images are passed through their Machine learning servers which can easily define your preferences and choices.

当您发布一些照片时，图像将通过其机器学习服务器传递，这些服务器可以轻松定义您的偏好和选择。

Using Several object detection techniques, it can observe your interests, for illustration, if you put display pic of you on a bike enjoying in nature, the algorithm will feed that you like bikes and nature, now, you’re profile would be surged to girls with whom you share something common.

使用多种对象检测技术，它可以观察您的兴趣，例如，如果您将自己的显示图片放在自然享受的自行车上，该算法将满足您喜欢自行车和自然的需求，现在，您的个人资料将激增至与您分享共同点的女孩。

Similarly, when you put on a bio, the expressions are conveyed to the NLP system that can detect sentiments of your first impressions as well can find your traits.

同样，当您穿上生物时，这些表情会传达到NLP系统，该系统可以检测您的第一印象的情绪，也可以找到您的特质。

Apart from this, your tweaked settings can also contribute to enhancing your profile. The more distance you opt for, the more in exploration state you’re, the less distance you opt for, the more serious and safe relation you want.

除此之外，您的调整设置还可以有助于增强个人资料。选择的距离越长，处于探索状态的距离就越多，选择的距离越短，想要的关系就越认真和安全。

Tinder also records your Left/Right swipe ratio, if you’re profile’s ratio is high your’re profile will be promoted to more pool of opposite gender and vice-versa.

Tinder还会记录您的左右滑动比例，如果个人资料的比率很高，您的个人资料将被提升为更多的异性，反之亦然。

应用使用情况(App Usage)

Tinder knows humans are going to obsolete soon, so they’re trying to capitalize ASAP.

Tinder知道人类将很快过时，因此他们正试图尽快利用资本。

Tinder surely wants to make a lot of money, but spending a lot of time on their app also contributes highly to their intentions.

Tinder肯定想赚很多钱，但是花很多时间在他们的应用程序上也极大地促进了他们的意图。

The algorithm promotes those profiles whose app activity is high, after all, the more the number of users, the more their capital growth.

该算法会提升应用活动活跃度的配置文件，毕竟，用户数量越多，他们的资本增长就越多。

Tinder loves its users and never wants to lose its fanbase, and as a result, it often surge active profiles, the more visible profile means more matches which makes the user less prone to try adversaries.

Tinder喜欢它的用户，从不希望失去自己的支持者，因此，它经常激增活跃的个人资料，更可见的个人资料意味着更多的匹配项，从而使用户不太容易尝试攻击对手。

When a user’s app using frequency is low, they will lower down his/her possibility of getting a match, because due to low frequency it is highly possible they won’t reply back to their match.

当用户的应用使用频率较低时，他们将降低其获得匹配的可能性，因为由于频率较低，他们很有可能不会回复他们的匹配。

There is too many males on the app as compared to the contrary. Tinder prioritize active woman and active man who are likely to serve them.

与之相反，该应用上的男性人数过多。 Tinder优先考虑可能为她们服务的活跃女人和活跃男人。

刷卡活动(Swiping Activity)

Another factor contributing to the “ELO” score is swiping activity, Tinder tracks how often you swipe left or right.

造成“ ELO”得分的另一个因素是刷卡活动，Tinder跟踪您向左或向右滑动的频率。

If you swipe right, you’re too lenient and maybe spamming, if that would be the case, tinder would again lower down your possibility to get a match or in technical terms, it will ShadowBan you. Because more swiping, means less messaging and that means less trust on the app and no CEO wants that.

如果您向右滑动，可能表示您太宽容，甚至发了垃圾邮件，如果真是那样，那么火种会再次降低您获得比赛的可能性或从技术角度来讲，它将使您成为ShadowBan 。因为刷卡次数更多，意味着消息传递更少，这意味着对应用程序的信任度降低，没有CEO会希望这样做。

But again, if you rarely swipe right, it means you’re too picky, due to the high men to women ratio, it won’t suit the algorithm either.

但同样，如果您很少向右滑动，则表示您太挑剔，由于男女比例较高，因此也不适合该算法。

You’re restricted to 100 right swipes for each day in Tinder, to ensure you’re really taking a gander at profiles and not simply spamming everybody to pile on arbitrary matches.

在Tinder中，您每天只能进行100次右滑，以确保您确实对个人资料感兴趣，而不仅仅是向所有人发送垃圾邮件以进行任意比赛。

To keep getting promoted you need to find a balance to maximize this part of the equation.

为了保持晋升，您需要找到一个平衡点，以最大化方程式的这一部分。

讯息活动 (Messaging Activity)

In the era of digitalization, privacy is merely a word.

在数字化时代，隐私只是一个词。

Tinder tracks your messaging activity too, it tracks to how many matches you messaged or initiated a conversation, it tracks the sentiments of that conversation, it tracks the duration you had a conversation, and even if you both shared your contacts number or not.

Tinder也跟踪您的消息传递活动，跟踪您发送消息或发起对话的匹配项数，跟踪该对话的情绪，跟踪对话的持续时间，甚至您是否共享了联系人号码。

If you got a high rate of interaction success, the algorithm will reward you by promoting your profile and gaining you more matches.

如果您获得很高的互动成功率，该算法将通过提升您的个人资料并获得更多匹配来奖励您。

But if you will keep ladies on hang, it will punish you by degrading the “ELO” score of your profile.

但是，如果您让女士保持绞刑状态，则会通过降低个人资料的“ ELO”得分来惩罚您。

For a personalized recommendation, the algorithm will keep an eye on the conversation and sentiments of conversation. Depending on the sentiments and traits you possess, if you and your match had a great messaging activity it will recommend you more profiles sharing some common traits with the former one.

对于个性化推荐，该算法将密切关注对话和对话情感。根据您的情绪和特质，如果您和您的比赛有出色的信息交流活动，它会建议您与其他人共享一些共同特征的个人资料。

The energy with which male message doesn’t appear to be coordinated with their articulacy, with the normal note tipping the scales at distinctly unromantic 12 characters. Messages from ladies will in general normal at a more artistic 122 characters.

男性信息所表现出的能量似乎并没有与他们的表达能力相协调，正常音符使音阶明显地变得不浪漫，只有12个字符。来自女士的消息通常会以更具艺术感的122个字符显示。

The algorithm will track the sentiment and will make sure your messages are positive but not too much alongside it also keeps an eye on the message sent per message received.

该算法将跟踪情绪，并确保您的消息是肯定的，但同时又不要过多，它还会关注收到的每个消息中发送的消息。

结合在一起 (Combining it together)

So it happens like this when you’ll install and signup for the app, it will ask for your data like ethnicity, race, education, height, company, etc.

因此，当您安装并注册该应用时，会发生这种情况，它将询问您的数据，例如种族，种族，教育程度，身高，公司等。

For the starters, the app doesn’t know much about you, except for the data that you’ve fed to it. The app will cooperate with you as a “beginner’s luck” because it still hasn’t classified you as a bad or good user, it will surge your profile to see your activity, it will track what kind of people you exactly thrive for?

对于初学者来说，该应用程序对您的了解不多，除了您提供给它的数据外。该应用程序将与您合作，成为“初学者的运气”，因为它仍然没有将您归类为不良用户或好用户，它会激增您的个人资料以查看您的活动，还会跟踪您真正适合哪些人？

If you mostly swiped for the Asian race with a Master’s level of education, it will try to show you profiles with the same eligibility.

如果您主要是通过硕士学历来参加亚洲竞赛，它将尝试向您显示具有相同资格的个人资料。

Gradually when you’ll keep on using the app, it will try to understand you better, now it not only trace your swipe activity but rather also track your affiliated Spotify, Instagram’s accounts activity, to give you a personalized experience of ads.

逐渐地，当您继续使用该应用程序时，它将尝试更好地了解您，现在它不仅可以跟踪您的滑动活动，而且还可以跟踪附属的Spotify(Instagram帐户活动)，从而为您提供个性化的广告体验。

Combining your past activity and social media interaction it will try to calculate your “ELO” score, it will check if the user is actively using the app and if not it will shadowban it, next it will trace for swipe frequency, if that too will be high, it will go onto next step computing user messaging frequency, if all criteria are met then “ELO” score would be high and the user’s profile will surface to same “ELO” scored profiles, if user messaging activity represents a threat to any specie, race or personal offenses, the user will be shadowbanned.

结合您过去的活动和社交媒体互动，它将尝试计算您的“ ELO ”分数，它将检查用户是否正在积极使用该应用，如果没有，它将对其进行遮挡，其次将跟踪滑动频率，如果也如此较高，则将继续执行下一步计算用户消息传递频率，如果满足所有条件，则“ ELO ”得分将很高，并且如果用户消息传递活动对任何用户构成威胁，则用户的个人资料将显示为相同的“ ELO ”得分个人资料物种，种族或人身伤害，将禁止用户使用。

其他约会应用程序如何计算“ ELO”分数？(How other dating apps are calculating the “ELO” score?)

Popular dating apps like OkCupid or eHarmony claim to use a special type of ML technique to predict your taste and present you with the most compatible match. These are expected to use the Gale-Shapley algorithm that was developed in 1962 by two economists who wanted to prove that any pool of people could be sifted into stable marriages.

OkCupid或eHarmony等流行的约会应用程序声称使用特殊类型的ML技术来预测您的口味并为您提供最兼容的匹配项。预期这些算法将使用由两名经济学家于1962年开发的Gale-Shapley算法，他们希望证明任何人都可以被筛分为稳定的婚姻。

In the first iteration, each unengaged man proposed to the woman he chose, and then the woman is expected to reply “maybe” to her match she prefers the most and “no” to others. She is then engaged to the suitor she most prefers so far, and that suitor is likewise provisionally engaged to her.
在第一次迭代中，每个未婚男士向他选择的女人求婚，然后期望该女人对她最喜欢的比赛回答“也许”，而对其他人则回答“不”。然后，她与到目前为止最喜欢的求婚者订婚，该求婚者也暂时与她订婚。
In the next round, each unengaged man proposed to the most-suited woman to whom he hasn’t proposed and then each woman replies “maybe” if she is currently not engaged or prefers him to her already an engaged partner.
在下一轮中，每个未婚男子向未曾推荐过的最合适的妇女求婚，然后每个妇女回答“也许”，如果她目前未订婚或相对于已订婚的伴侣更喜欢他。
This process is repeated until everyone is engaged.
重复此过程，直到每个人都参与。

This algorithm is guaranteed to produce a stable marriage for all participants in time.

该算法保证为所有参与者及时产生稳定的婚姻。

开发Tinder的ML模型 (Developing Tinder’s ML Model)

Step 1. Data labeling and cleaning: Go through 500 to 1000 profiles, each with 4–5 photos, and classify them into “like”, “dislike”, or “neutral”.

步骤1.数据标记和清理：遍历500至1000个配置文件，每个配置文件包含4-5张照片，并将它们分类为“喜欢”，“不喜欢”或“中性”。

Step 2. Neural Network: Using transfer learning train an existing neural network to swipe right or left based on already classified images.

第2步：神经网络：使用转移学习训练一个现有的神经网络，以基于已分类的图像向右或向左滑动。

Step 3. Algorithm: Write a function that scores a profile based on the sum of the scores on each image that the above network has come up with.

第3步：算法：编写一个函数，根据上述网络提出的每个图像上的得分总和对配置文件进行得分。

天德是一位出色的媒人吗？ (Is Tinder an excellent matchmaker?)

You feed in some information, Tinder collects some more information about you and you’re surfaced online popped with multiple matches.

您输入一些信息，Tinder会收集有关您的更多信息，并且您会在网上看到多个匹配项。

But what Tinder lack is the result, nobody ever discloses what happened after they met? They hooked up, they had babies, they fought, they’re leading a perfect loving life or they ghosted each other?

但是结果是Tinder缺乏什么，没有人透露他们见面后发生了什么？他们勾搭起来，生了孩子，战斗了，过着完美的爱情生活，还是彼此鬼混了？

Tinder never received the reward/punishment for its recommendation system so practically it can never improve by its experience making it more robust, it’s algorithm might change making it more robust but since it can never get real human feedback it can never improve practically with the same algorithm in-charge.

Tinder从未对其推荐系统获得过奖赏/惩罚，因此从实践上讲，它不可能变得更健壮，因此永远无法得到改进；算法可能会变得更健壮，但由于无法获得真正的人为反馈，因此在相同的条件下，它永远无法得到改进算法负责。

Secondly, I don’t believe Tinder can be trusted, it is data-hungry and it tracks you like most as it can, to understand you, and I am not in favor of AI ruling us in few years.

其次，我不相信Tinder是值得信赖的，它非常耗数据，并且会尽可能地追踪您，以了解您，并且我不赞成AI在几年内统治我们。

Valentines may come & go, but, your data will be put online forever.

情人可能来来往往，但是，您的数据将永远在线存储。

结论 (Conclusion)

Hopefully, this article has given you how Tinder is using AI and how its recommendation workflow to find you a loving partner.

希望本文能给您Tinder如何使用AI及其推荐流程如何找到您的爱人。

As always, thank you so much for reading, and please share this article if you found it useful!

与往常一样，非常感谢您的阅读，如果发现有用，请分享这篇文章！

Feel free to connect:

随时连接：

LinkedIn ~ https://www.linkedin.com/in/dakshtrehan/

领英〜https: //www.linkedin.com/in/dakshtrehan/

Instagram ~ https://www.instagram.com/_daksh_trehan_/

Instagram〜https: //www.instagram.com/_daksh_trehan_/

Github ~ https://github.com/dakshtrehan

Github〜https: //github.com/dakshtrehan