Online match making started as far back as in the 1950s…using a questionnaire and an IBM 650 machine to match 49 men with 49 women!

In over 50 years of its journey most match making apps still show profiles/images based on a superficial matching of individual’s interests, preferences based on profession, financial status, food habits etc..to the more interesting ones like seeking instant gratification and based on sexual orientation. All potential matches being clicked and filtered personally by the seeker.

With a heap of dating apps or match making sites currently in the market, a perfect differentiation between them would be the recommendation system. They have realised that physical characteristics play a vital role in a user’s preference as self-descriptors are hardly available and usually unreliable in this scenario, machine learning works its magic in predicting a user’s preference to find a potential soulmate. These online match makers like Blinq have started using their huge set of labelled data with like/dislike for a profile in training the deep learning models like CNN (Convolutional Neural Network) to find attractiveness of a user. Convolutional Neural Networks (CNNs) are a category of neural networks which have risen in prominence in recent years because of its wide application in image and video recognition, recommender systems etc. A host of companies like Facebook, Google, Amazon and others have been using them because of their unprecedented performance in the field of image classification. In simple/layman terms, CNN are just manifestation of deep learning and have layers and layers of filters. Every network layer acts as a detection filter for the presence of specific features or patterns present in the original image. For instance, see the image below;