FUN

The FUN project proposed neural crawling to replace ranking via Page Rank logic – using language models to estimate the semantic quality of web pages and prioritise accordingly.

ProjectID

NGI-OpenWebSearchEU:FUN

Acronym

FUN

Additional Info

Full description: The FUN project proposed neural crawling to replace ranking via Page Rank logic – using language models to estimate the semantic quality of web pages and prioritise accordingly. The team developed four strategies and tested them on 87 million pages from ClueWeb22-B. On natural language queries, their best approach (DomQ) consistently outperformed PageRank in both crawling effectiveness and downstream retrieval quality. On keyword queries, it remained competitive.

Enduser Relevance

Giving endusers better quality results

Contact

Endorsements

Not available yet

Disclaimer

n/a yet

Country:  Italy

NGI Project:  OpenWebSearch.EU

Status: Early research demo

Category: Services and applications (for example email, instant messaging, video chat, collaboration, cloud storage)