Love seeking is no longer an easy task. So, what is new? This generation is considered to be extremely lucky, as it has been at the forefront of the massive changes brought by the technological revolution. Our childhood was filled with tape recorders and video cassettes, wired phones, big-boxy computers and large libraries filled with books. Then, almost inexorably, we made that transition to iPads, YouTube, smart phones, iPads, iPads, and eBooks.

New channels of interaction:

The nature of our personal relationships has changed. There are many new channels and modes of interaction. “Being friends” has a new meaning. A web of invisible online networks is increasingly defining society. This got me thinking about another aspect. Has the idea and act of “falling in love” changed?

Do you recall how in simpler times love was often found? Friends and family set up blind dates for you. To find out if you had similar interests and beliefs, you interacted with others face-to-face. Time spent together was a key indicator of compatibility. Let’s cut to the present: Artificial Intelligence and data science have taken over the role as your friend, family member, acquaintance, and even your own judgement in finding your perfect partner.

Two axes have transformed the rules of engagement

1. How to connect with potential love interests

The channels for communication and introduction. Physical meetings are much more common now. You can discover potential romantic associations through your phone, tablet, or computer. With a swipe of your fingers, you can choose. Talk through an emoji screen to express your emotions. You hope that love blossoms.

2. Number of people to choose from

The number of people you have the opportunity to choose from. This is one of the most frightening and obvious changes. It’s no longer you, or anyone you know, who suggests a possible match. This role has been taken over by machines.

The current paradigm allows people, both humans and animals, to be parametrised and reduced down to a “feature-set”. AI-driven algorithms then can find the best ‘optimal matches”, i.e. Who you are most compatible with? Let’s take a look at how the ‘business of Love’ actually works behind-the scenes.

Today, recommendation engines are used in all areas of our lives. Recommender engines are at work when you shop on Amazon. You get recommendations for products you might like to purchase. Netflix also suggests movies that you might like.

• Historical Trends

Based on person X’s past behaviour, this engine would attempt to predict future choices X might make.

• People Like You

This approach uses a machine learning-based clustering algorithm to identify groups of people who are similar. The behaviour of the person could define similarity (You can either define similarity by their behaviour (what they buy, how they pay, frequency of purchase, and many other things) or by their characteristics (e.g. Their age, gender, and location. Or both. The recommendation engine would then predict X’s choices based on not only what X has done previously, but also on what others ‘like X have been doing.

• Content Based

This approach is more focused on the target items and the features of those items than the person. This would be used to identify the characteristics of Netflix movies that X enjoys (genre, length and language, actors, etc.). To determine his likely interest in a particular topic.

Most recommendation engines may use a combination of these factors. The engine will be able to predict behaviour more accurately the more feedback it receives. Feedback is basically the acceptance or rejection of a prediction (also known as a recommendation).

Let’s now get back to the business of love. It’s easy to see the parallels.

• Instead of the movie or item whose acceptance probability was predicted, it would be a person with whom X is most likely to get along.
• Historical trends refer to data about X’s past relationships.
• X’s behaviour and other characteristics will determine if X is “people like X”. age, ethnicity, education, location, employment status, food preferences, hobbies, eye colour, size, height etc.
• These same features in target items would be referred to as the characteristics of the people on the potential target list.

The better the algorithms will work, the longer the list of characteristics that are available. It doesn’t matter if X provides the data explicitly — most of the information in our world today is available online about everyone in any way shape or form — it’s about just accessing it and mining it.

These data points can be used to help AI algorithms assign probability scores or likelihood to each prospect in order to determine which prospects are most likely to like and be chosen by X. As in all the cases above, the system is self-learning and will continue to improve with every feedback. Every time X accepts or refuses a suggestion match.

Conclusion:

This is the problem: this approach assumes that love is logic-based or rule-based. In conclusion the matchmaking industry will not be able reap the benefits from AI in the same way as other applications. Because, in reality, and so I like it to believe, finding love doesn’t always seem logical. Despite the increasing use of AI and machines in all areas of our lives, I believe that the magic and mystery of this particular area will continue to exist.