BLINK with Elasticsearch for Efficient Entity Linking in Business Conversations

Abstract

An Entity Linking system aligns the textual mentions of entities in a text to their corresponding entries in a knowledge base. However, deploying a neural entity linking system for efficient real-time inference in production environments is a challenging task. In this work, we present a neural entity linking system that connects the product and organization type entities in business conversations to their corresponding Wikipedia and Wikidata entries. The proposed system leverages Elasticsearch to ensure inference efficiency when deployed in a resource limited cloud machine, and obtains significant improvements in terms of inference speed and memory consumption while retaining high accuracy.

ChartSumm: A Comprehensive Benchmark for Automatic Chart Summarization of Long and Short Summaries

“…In the future, we would like to extend ChartSumm to a multilingual dataset to address the scarcity of well-formatted datasets in other low-resource languages. We will also study how to incorporate query relevance [33][34][35], question-answering [36][37][38][39], and entity recognition [40][41][42] capabilities in this task.…”

Multilingual bi‐encoder models for biomedical entity linking

“…Afterward, each candidate is analyzed in more detail with a cross‐encoder that combines the word and entity text (Wu et al 2020). Laskar, Chen, Martsinovich, et al (2022), Laskar, Chen, Johnston, et al (2022), Bhargav et al (2022) have also employed this method in other studies.…”

“…. Laskar, Chen, Martsinovich, et al (2022),,Bhargav et al (2022) have also employed this method in other studies.3.3.2 | Surface-based CGIf an alias table does not exist or is of limited scope, CG can be performed by exploiting the surface similarity between mentions and entities.…”