Visual Discovery at Pinterest

By Andrew Zhai, Dmitry Kislyuk, Yushi Jing, Michael Feng, Eric Tzeng, Jeff Donahue, Yue Li Du, Trevor Darrell. 2017

84 citations recorded by the source corpus.

Over the past three years Pinterest has experimented with several visual search and recommendation systems, from enhancing existing products such as Related Pins (2014), to powering new products such as Similar Looks (2015), Flashlight (2016), and Lens (2017). This paper presents an overview of our visual discovery engine powering these services, and shares the rationales behind our technical and product decisions such as the use of object detection and interactive user interfaces. We conclude that this visual discovery engine significantly improves engagement in both search and recommendation tasks.

Subfields: Computer Vision and Pattern Recognition, Information Retrieval & Search

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Information Retrieval & Search

Citations in the corpus

Speed/Accuracy Trade-Offs for Modern Convolutional Object Detectors (2016, background): “Since appearing in 2015, Faster R-CNN has been particularly influential, and has led to a number of follow-up works [2, 34, 33, 45, 13, 5, 19, 44, 23, 46] (including SSD and R-FCN).”
Classification is a Strong Baseline for Deep Metric Learning (2018, background): “Metric learning, also known as learning image embeddings, is a core problem of a variety of applications including face recognition [13, 16, 22, 23], ﬁne-grained retrieval [17, 18, 27], clustering [31], and visual search [7, 11, 32, 33].”
Related Pins at Pinterest: The Evolution of a Real-World Recommender System (2017, methodology): “We have two visual candidate sources, described further in [13] and [25].”
Recommending what video to watch next: a multitask ranking system (2019, methodology): “…used co-occurrences of items to generate candidates, [11] adopted a collaborative fltering based method, [14] and [19] applied a random walk on (co-occurrence) graph, [42] learned content representation to flter items to candidates, and [10] described a hybrid approach using mixture of features.”
Snorkel DryBell: A Case Study in Deploying Weak Supervision at Industrial Scale (2018, background): “We observe other organizations already developing broad-purpose models for domains such as language modeling [13, 14], object recognition [20], and object detection [39].”
Self-supervised Learning for Large-scale Item Recommendations (2020, background): “Deep learning has led to many successes in building industry-scale recommender systems such as video suggestion [11], news recommendation [38], and visual discovery in social networks [30, 45].”