20230153870. UNSUPERVISED EMBEDDINGS DISENTANGLEMENT USING A GAN FOR MERCHANT RECOMMENDATIONS simplified abstract (Visa International Service Association)
UNSUPERVISED EMBEDDINGS DISENTANGLEMENT USING A GAN FOR MERCHANT RECOMMENDATIONS
Organization Name
Visa International Service Association
Inventor(s)
Yan Zheng of Sunnyvale CA (US)
Yuwei Wang of San Francisco CA (US)
Michael Yeh of Palo Alto CA (US)
Liang Wang of San Jose CA (US)
UNSUPERVISED EMBEDDINGS DISENTANGLEMENT USING A GAN FOR MERCHANT RECOMMENDATIONS - A simplified explanation of the abstract
This abstract first appeared for US patent application 20230153870 titled 'UNSUPERVISED EMBEDDINGS DISENTANGLEMENT USING A GAN FOR MERCHANT RECOMMENDATIONS
Simplified Explanation
The patent application describes a method for training a recommendation system to provide merchant recommendations based on payment transaction records.
- The system receives raw merchant embeddings and raw user embeddings generated from payment transaction records.
- A generative adversarial network (GAN) is trained to generate modified merchant embeddings by removing a location feature.
- The trained GAN and a preference model are used to generate a list of merchant rankings based on modified merchant embeddings, past user preferences, and the target location.
- The system can recommend merchants in the target location based on the generated rankings.
Potential Applications
- E-commerce platforms can use this technology to provide personalized merchant recommendations to users based on their preferences and location.
- Mobile payment apps can utilize this technology to suggest nearby merchants to users based on their transaction history and current location.
Problems Solved
- This technology solves the problem of providing accurate and relevant merchant recommendations to users based on their preferences and location.
- It addresses the challenge of removing location bias from merchant embeddings to ensure fair and unbiased recommendations.
Benefits
- Users can discover new merchants that align with their preferences and are conveniently located.
- Merchants can benefit from increased visibility and customer engagement through personalized recommendations.
- The system provides fair and unbiased recommendations by removing location bias from the merchant embeddings.
Original Abstract Submitted
embodiments for training a recommendation system to provide merchant recommendations comprise receiving, by a processor, raw merchant embeddings and raw user embeddings generated from payment transaction records, wherein the raw merchant embeddings include a plurality of embedded features. a generative adversarial network (gan) is trained to generate modified merchant embeddings from the raw merchant embeddings, where the modified embeddings remove a location feature. subsequent to training and responsive to receiving a request for merchant recommendations in the target location for the target user, the gan and a trained preference model are used to generate a list of merchant rankings based on a new set of modified merchant embeddings, past preferences of a target user, and the target location to recommend merchants in the target location.
