Could Google passage indexing be leveraging BERT? [Dawn Anderson @SearchEngineLand]
Google’s language model BERT has (and still is) created some attention in the NLP and Search community when it was released in 2018, and incorporated in Google’s web search engine in 2019. Given that BERT is, among others, used to rank search results, it also has high relevance for the recommender system community, which also requires to rank items.
Dawn Anderson from SearchEngineLand wrote a nice summary about BERT including its history, what BERT is used for, and its competitors (GPT-2/3, etc.). The most interesting part, however, is her view on BERT for passage indexing, i.e. the little snippets that Google displays sometimes for the top search query:
While her article focuses on web search and search engine optimization, passage indexing is also relevant for the recommender system community (think of, for instance, explainable recommendations).
Read the full article here https://searchengineland.com/could-google-passage-indexing-be-leveraging-bert-342975.
And, if you want to learn even more on BERT, this Reddit thread is worth a read.
About The Author
Joeran Beel
I am the founder of Recommender-Systems.com and head of the Intelligent Systems Group (ISG) at the University of Siegen, Germany https://isg.beel.org. We conduct research in recommender-systems (RecSys), personalization and information retrieval (IR) as well as on automated machine learning (AutoML), meta-learning and algorithm selection. Domains we are particularly interested in include smart places, eHealth, manufacturing (industry 4.0), mobility, visual computing, and digital libraries. We founded or maintain, among others, LensKit-Auto, Darwin & Goliath, Mr. DLib, and Docear, each with thousand of users; we contributed to TensorFlow, JabRef and others; and we developed the first prototypes of automated recommender systems (AutoSurprise and Auto-CaseRec) and Federated Meta Learning (FMLearn Server and Client).