Abstract
A system and method for automatically generating a highly relevant list of candidate products from an advertiser’s product catalog to feature alongside a video advertisement are provided herein. The manual process currently used for selecting products is often time-consuming, unscalable, and results in sub-optimal product relevance. The solution provided herein uses an automated, AI-powered approach to identify products featured in a video and then utilizes a two-stage similarity ranking process involving text and image data to find the most contextually and visually similar items in a potentially massive product catalog. This multi-modal approach enhances relevance compared to simple text-based searches and enables the efficient creation of scalable shoppable video experiences. The automated system addresses the challenge of maximizing the e-commerce potential of video ads by reducing the time from product discovery to purchase.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Shreenivasa, Satish; Brenner, Morgan; Blossom, Leigh Purcell; Li, Xi; and Wallace, Briana, "Automated Multi-Modal System for Generating Product Recommendations for Video Advertisements", Technical Disclosure Commons, ()
https://www.tdcommons.org/dpubs_series/9084