Concept-Aware Research Discovery for HCH DU

A transparent discovery layer that combines profile topics with concepts learned from the researcher's own publication corpus, then ranks public works by topical overlap, collaboration, recency and citation strength.

Ranked discoveries

Catalogue candidates

Profile concepts

Network-connected

Recent discoveries

Discovery feed

What is most relevant to this research profile?

Every result shows the signals that moved it upward. A high score represents stronger similarity signals in the sampled ResearchHub catalogue; it is not a claim of scientific quality or definitive relevance.

Add public research interests, research areas, keywords or publications to generate concept-aware discoveries.

Concept layer

How this differs from keyword search

1. Start with the researcher's explicit interests, areas and keywords.
2. Extract recurring non-stopword concepts from the researcher's title and abstract corpus.
3. Search the public catalogue across the combined concept set.
4. Rank candidates using topical overlap, collaborator links, recency and citation strength.

Evidence policy

Transparent by design

This release is concept-aware and lexical rather than embedding-based semantic AI. It uses public ResearchHub catalogue data and does not invent papers, relationships or relevance claims. Candidate sampling is limited to the highest-ranked catalogue matches available to this page.