ProductNet: a Collection of High-Quality Datasets for Product Representation Learning

TL;DR

ProductNet introduces high-quality product datasets and a multi-modal neural network for improved product understanding, enabling efficient annotation, accurate categorization, and versatile transfer learning applications.

Contribution

The paper presents a novel iterative approach combining dataset curation and product representation learning using a multi-modal neural network and active learning.

Findings

94.7% top-1 accuracy on 1240 classes

High-quality datasets support diverse product understanding tasks

Embedding accelerates annotation and transfer learning

Abstract

ProductNet is a collection of high-quality product datasets for better product understanding. Motivated by ImageNet, ProductNet aims at supporting product representation learning by curating product datasets of high quality with properly chosen taxonomy. In this paper, the two goals of building high-quality product datasets and learning product representation support each other in an iterative fashion: the product embedding is obtained via a multi-modal deep neural network (master model) designed to leverage product image and catalog information; and in return, the embedding is utilized via active learning (local model) to vastly accelerate the annotation process. For the labeled data, the proposed master model yields high categorization accuracy (94.7% top-1 accuracy for 1240 classes), which can be used as search indices, partition keys, and input features for machine learning models. The product embedding, as well as the fined-tuned master model for a specific business task, can also be used for various transfer learning tasks.