What is it about?
This work proposes UCO, an approach to fine-tune Transformer encoder models using a dual-loss based optimization, using implicit user-intent within centrality labels.
Featured Image
Photo by Oberon Copeland @veryinformed.com on Unsplash
Why is it important?
Alphanumeric queries are a special case within user queries on e-commerce platforms which may consists of product metadata such as model number, SKU, etc. Such queries with (near) exact product metadata most likely do note achieve desired product results. This approach addresses the problem by transfer learning user-intent via centrality annotations.
Read the Original
This page is a summary of: Product Retrieval and Ranking for Alphanumeric Queries, October 2024, ACM (Association for Computing Machinery),
DOI: 10.1145/3627673.3679080.
You can read the full text:
Contributors
The following have contributed to this page
