Summary
ResearchRabbit uses network-based algorithms to support search and discovery.
Recommendations are based on citation and authorship relationships within the scholarly literature.
Our approach prioritises transparent methods grounded in the structure of academic research.
The product is designed to be efficient and avoids unnecessary compute-heavy features.
How ResearchRabbit supports discovery
ResearchRabbit supports search and discovery by analysing relationships between papers and authors within the scholarly literature.
Seed paper search
When you search using keywords, a DOI, or a paper title, ResearchRabbit retrieves relevant papers from its database of over 280 million papers using industry-standard search infrastructure. This relies on precomputed lookup indexes and graph-based methods.
ResearchRabbit does not use large language models to power its core recommendation algorithms.
Connected papers and authors
Once you have seed papers, ResearchRabbit helps you explore related literature by looking at how papers are connected through:
Citations and references
Shared citations and co-citations
Authorship and co-authorship
ResearchRabbit uses network-based algorithms to understand how papers relate to one another across the research network, based on first-degree and second-degree connections. This approach is grounded in the structure of the literature rather than in content generation.
In practice, this helps relevant and influential work surface naturally as you explore. 🐇
Our approach to responsible AI use
We regularly think about whether and how to use new technologies, including AI. When we do, we consider things like:
Privacy and data protection
Explainability and transparency
Accuracy and reliability
Environmental impact
Whether it genuinely improves the research experience
We aim to adopt new technologies deliberately, and we care much more about usefulness than hype. This includes using AI. As ResearchRabbit evolves, we’ll keep this page updated so you can always see how things work behind the scenes. 🙂
You can read more here