We introduce a deep learning-based listing embedding method using a Siamese network for similar home recommendations at Zillow. Our embedding model utilizes both collaborative and content information, models complex interactions between attributes, and effectively handles cold-start items. We have demonstrated the strong performance of our embedding method in surfacing relevant homes to the users.
For more information, see http://nu32.org. This video is a supplement to the book "Embedded Computing and Mechatronics with the PIC32 Microcontroller," Lynch...