Pinterest is bringing artificial intelligence (AI) to its platform to offer up better pin recommendations to its users.
Related Pins, which appear beneath pins, will now have deep learning applied to them, making them more relevant by using the data, such as saved pins, from a person’s most recent visit. That will be combined with the person’s interactions to label pins to surface better suggestions.
“Ultimately, we developed a scalable system that evolves with our product and people’s interests, so we can surface the most relevant recommendations through Related Pins,” Pinterest software engineer Kevin Ma said in a blog post. “We use deep learning to generate recommendation candidates, which, in testing, has increased engagement with Related Pins by five percent.”
Until now, Pinterest used a method called board co-occurrence to surface relevant pins, but that alone, Ma said, was not enough to guarantee a user didn’t get sucked into looking at a pile of pins he or she was not interested in.
Now, using deep learning, Pinterest is also applying a method called Pin2Vec to embed pins in the context of pinners’ activity.
“Pin2Vec has become an important source for generating candidates, but it doesn’t replace board co-occurrence,” he said. “In our tests, we’ve found the board co-occurrence is better performing for long tail Pins that are sparse in engagement data. Both board co-occurrence and Pin2Vec are used to generate candidates, while a separate rerank system sorts the candidates based on richer features.”
Ma also said Pinterest is “making our models faster and analyzing more signals to better personalize recommendations for Pinners around the world.”
To read Ma’s blog post, which explains the technical aspects of Pin2Vec, click here.
Jennifer Cowan is the Managing Editor for SiteProNews.
The post Pinterest Turns to Artificial Intelligence for Pin Recommendations appeared first on SiteProNews.