Hourglass Software LLC provides RESTful APIs based on Intellectual Property (owned by Hourglass Software) for using Facial Recognition, Natural Language Processing, and other physical attributes for double-model Machine Learning predictions for online dating applications. Any mobile or web application can use the RESTful API over HTTP to integrate with the Smart Match Engine and Algorithm. Thus, by leveraging Machine Learning, there is a higher likelihood of the application user to “like” the profile being provided to them. It is based on a confidence value which can be configured per user. Each user has his/her own personalized Machine Learning Model/s built for him/her that makes prediction on who the user is more likely to LIKE. The purpose is to supplement your application, not to perform holistic dating application solution. The web API is built using C# ASP.NET MVC 4.6. Machine Learning updates every 50 records to retrain the model. It also adapts to the users most recent behavior by allowing “noise injection”.
• TITLE - High Confidence Predicted Profiles for Online Dating Matching Using Continuously Learning Machine Learning Models Trained for each User with Physical and Facial Attributes • EFS ID – 37625118 • Application Number – 16670883 • Inventor – Faisal A Qureshi • Assignment – Hourglass Software, LLC. • Status – FILED, PENDING • Summary • Key components of technology used have been filed for non-provisional patent • This includes: • Using facial recognition calculations, other physical attributes (ie: height, eye color, hair color), and NLP analyzed “about you” text, to be used as feature vector for Machine Learning model • Dual-model based Machine Learning • Foreground making predictions • Background for training • Then swap (function) • Noise Injection • Show any profile (< X% prediction) so machine learning gets smarter and self-adjusts Infrastructure and Customization
We own and use: • Microsoft Azure for hosting the web service API • Rackspace MySQL database • AWS S3 • AWS Rekognition • AWS Comprehend • Microsoft ML.NET We customize parts of the API for specific for your application. You can also make requests to further customize search filter fields for profile retrieval. These require code changes and database changes. The core logic, however, will not be changed. Supported face picture sizes are also customizable. Domains will be determined at time of contract.
See User Guide
for more information and to purchase this product with subscription, contact firstname.lastname@example.org (price varies depending on number of users)