ECNLP 3 Program

ECNLP 3 will be held as an online workshop due to the COVID-19 situation.

Join the workshop online:

• 05:30-14:15 (PDT; UTC/GMT-7), Friday, July 10, 2020

• Instructions: join us live from https://virtual.acl2020.org/workshop_W16.html

  • This link is accessible to everyone who registered for the ACL 2020

  • Pre-recorded talks are available for viewing on demand

• Join our mailing list to get the latest news about ECNLP and e-Commerce research: https://bit.ly/3cu0uIk

All times are in Pacific Time (GMT-7).

The workshop will be held remotely, so you will need to convert this to your local time.

Personalized Natural Language Processing for e-Commerce

Speaker: Prof. Julian McAuley

In this talk I'll discuss how techniques from user modeling and recommender systems can be used to develop personalized models of natural language. Personalized language models are valuable when dealing with language that deals with preferences, is subjective, or depends on personal context, etc. We show that such applications are abundant in e-commerce settings, and give examples of personalized NLP systems ranging from question answering to recipe generation.

Entity Discovery and Linking: Moving from Academia to Industry

Speaker: Prof. Heng Ji

Entity discovery and linking (EDL) is a key technology to discover and fuse entity knowledge across domains and languages together, and thus it has many wide applications in both government-supported academic research (e.g., disaster monitoring) and industrial research (e.g., customized entity recommendation). In this talk I will give an overview of recent advances on EDL in the academic world, focusing on cross-genre cross-lingual transfer and fine-grained EDL. Then I will summarize some challenges on deploying these techniques to industrial applications and potential required extensions, including real-time online entity profiling, and joint entity extraction, linking and recommendation.

The workshop papers can be viewed here:

https://www.aclweb.org/anthology/events/acl-2020/#2020-ecnlp-1