A Web Scale Entity Extraction System
This work addresses the problem of scalable semantic understanding for web practitioners, but it is incremental as it builds on existing Transformer methods for entity extraction.
The paper tackles the challenge of building a large-scale entity extraction system for web content by leveraging multi-modal Transformers, demonstrating effectiveness through multi-lingual, multi-task, and cross-document type learning, with results showing improved data quality via noise-minimizing label collection schemes.
Understanding the semantic meaning of content on the web through the lens of entities and concepts has many practical advantages. However, when building large-scale entity extraction systems, practitioners are facing unique challenges involving finding the best ways to leverage the scale and variety of data available on internet platforms. We present learnings from our efforts in building an entity extraction system for multiple document types at large scale using multi-modal Transformers. We empirically demonstrate the effectiveness of multi-lingual, multi-task and cross-document type learning. We also discuss the label collection schemes that help to minimize the amount of noise in the collected data.