Citrus

Citrus Search is a similarity-based search engine for scientific literature. It helps you find closely related publications by selecting a seed paper and exploring results driven by selected similarity metrics. It leverages graph and text-based machine-learning techniques to compare papers and is powered by Semantic Scholar’s Open Research Corpus.

---

What it does

• Lets you choose a seed paper and discover related work quickly without sifting through irrelevant results.
• Uses multiple similarity measures to define what "related" means (Citation Network, Content similarity, and others).
• Provides an overview of closely related work, helping researchers grasp the landscape around a topic at a glance.
• Runs on a live indexing source drawn from Semantic Scholar’s Open Research Corpus, spanning over 200 million publications and around 2 billion citations.

---

How Citrus determines relatedness

• Citation Network: Finds papers with similar connections in the citation network to your seed.
• Content: Finds papers with similar concepts, ideas, or research questions in their abstracts and titles.
• Behind the scenes, Citrus computes paper similarity using graph and text-based machine-learning techniques.

---

How to use Citrus Search

1. Select a seed paper that represents your topic of interest.
2. Start the search by choosing the similarity measure (e.g., Citation Network, Content) or add additional seed papers first.
3. View an overview of closely related work, typically presented on a timeline or in a summarized view.

---

Data and credits

• Citrus indexes data provided by Semantic Scholar's Open Research Corpus.
• The corpus includes a large-scale dataset of publications and citations to support comprehensive relatedness analysis.

---

Safety and scope

• Designed for academic exploration and literature discovery.
• Users should verify results and consider multiple similarity signals for robust literature reviews.

---

Core Features

• Seed-based literature discovery: start from a chosen paper and explore related work
• Multiple similarity measures (Citation Network, Content) for flexible exploration
• Timeline/overview view of related publications
• Large-scale data source: Semantic Scholar Open Research Corpus integration
• Quick navigation for efficient literature reviews
• Feedback-enabled: report issues or suggestions for improvement