With the aim to offer a more nuanced content to the user, Pinterest have unveiled a machine learning tool that provides the user with the most personalized and relevant Pins.
The ML tool, titled Pinnability, runs on smart feed, “and estimates the relevance score of how likely a Pinner will interact with a Pin. With accurate predictions, we prioritize those Pins with high relevance scores and show them at the top of home feed,” explains Pinterest’s software engineer Yunsong Guo.
Earlier all home feed content from each source (e.g., following and Picked For You) was put together chronologically; a newer Pin from the same source would show up before an older Pin, irrespective of the degree of interest either of the pins hold to the user. It stunted the discovery of newer more interesting pins as a low-relevance Pin could very well appear before a high-relevance one.
“We experimented with multiple machine-learning models, including LR [logistical regression], GBDT [gradient-boosted decision trees], SVM [support vector machines], and CNN, and we use AUC [area under the curve] score in 10-fold cross-validation and 90/10 train-test split settings with proper model parameters for evaluation,” he wrote in a blog post last week.
“We observed that given a fixed feature set, the winning model always tends to be either LR or GBDT for Pinnability.”
More details about the latest addition can be found here.
Image credit: Pinterest
