Semantic Scholar

Semantic Scholar is a free academic search engine developed by the Allen Institute for AI. It focuses mainly on STEM subjects, especially computer science, biomedical sciences, and psychology.

It includes journal articles, conference papers, and preprints from sources like arXiv and bioRxiv — and uses AI tools to summarise, filter, and connect papers based on meaning, not just keywords.

Note: Semantic Scholar does not include books or connect to Abertay’s subscriptions. Use LibrarySearch to find full-text access and broader content coverage — including books, peer-reviewed articles, and specialist STEM databases.

Use Semantic Scholar alongside LibrarySearch — especially when:

• You’re doing a dissertation or literature review in a science or technology subject

• You want to find recent research or preprints not yet indexed elsewhere

• You’re exploring a complex topic and want to trace connections between papers

• You’re interested in citation context, related research, or author influence

• You’ve hit a wall with LibrarySearch and want to try a different approach

• Its semantic search helps you find papers based on meaning, even when different terminology is used across the literature. This can be especially useful when you're not sure what terminology to use or want to see how others describe your topic.

Semantic Scholar is not a replacement for LibrarySearch — but it can complement it when you're looking to expand your search strategy.

Need help getting started? See: Search Skills: Finding Information Effectively

Semantic search tries to match the meaning or intent behind your query — rather than just looking for exact keyword matches.

This can help when:

• You’re unfamiliar with the terminology in a new field

• You want to avoid missing results that use different wording

• You’re searching for concepts that aren’t captured by one fixed phrase

Example: A search for how AI detects disease in scans might return papers on machine learning in radiology or deep neural networks for image classification, even if those exact words weren’t used.

Semantic Scholar uses this approach to show relevant connections between papers — even across different disciplines or writing styles.

Semantic Scholar offers a range of tools to help you find, understand, and explore STEM research more effectively:

• Semantic (AI-enhanced) search: Matches results based on meaning and context — helping you find relevant papers even if they don’t use your exact search terms
  
  
• Citation insights: See who cited a paper, how, and why (e.g. support or contrast)
  
  
• Related paper suggestions: Explore clusters of similar or connected research
  
  
• Author profiles: Track a researcher’s output, collaboration network, and impact
  
  
• Preprint coverage: Includes content from arXiv, bioRxiv, medRxiv, and others
  
  
• Filters: Narrow by topic, publication year, venue, or paper type

The platform also includes optional GenAI features to help you read and compare research more efficiently:

• TL;DR – One-sentence summaries of research papers

• Ask This Paper – Type natural-language questions to explore a paper’s content

• Semantic Reader – An interactive reading mode available on some papers

These tools are useful for scanning and getting the gist of a paper — but they have limitations. Always verify by reading the full article before using or citing it.

Be aware of these limitations when using Semantic Scholar:

• No full-text access via Abertay – Check LibrarySearch if you hit a paywall

• No books or chapters – For monographs or textbooks, use LibrarySearch

• STEM-focused only – Limited coverage in business, law, humanities, or social sciences

• AI summaries may oversimplify – Good for scanning, but don’t replace close reading

To make the most of your searches:

• Use keywords and short phrases (not full questions)

• Use filters on the left (e.g. year, publication type, fields of study)

• Check the summary and citation context before deciding to read

• Use “Cited by” and “References” links to trace research forward and backward — a strategy known as citation chaining. This can help you find more relevant sources and understand how a topic has evolved.

Note: Semantic Scholar doesn’t support Boolean operators (AND, OR, NOT), truncation (*), or wildcards — because it uses semantic matching rather than keyword logic. 

For full-text access, try pasting the paper title into LibrarySearch or use LibKey Nomad.

Semantic Scholar makes it easy to follow the trail of research using citation links.

• Click References to see earlier work the author cited (backward chaining)

• Click Cited by to find newer papers that have cited the work (forward chaining)

Example:
You find a 2021 paper on machine learning and breast cancer.

• Backward chaining: Check the paper’s reference list to identify foundational studies

• Forward chaining: Use the “Cited by” link to explore newer papers that build on it

Want to learn more? See: Citation Chaining Techniques

Not all articles in Semantic Scholar are peer-reviewed — especially preprints, which haven’t gone through formal checks. That doesn’t mean they’re low quality — but it does mean you need to be critical.

Ask yourself:

• Has the paper been peer-reviewed?

• Is the author affiliated with a credible institution?

• Is it cited by other recent or relevant papers?

• Does the summary match what the full paper says?

Use these guides to help:

• How to Evaluate Information Effectively
• Academic vs Non-Academic Sources
• Critically evaluating journal articles

Semantic Scholar often links to Open Access versions — but not always.

If full text isn’t available:

• Copy the article title and try it in LibrarySearch

• Check for Open Access links on the right-hand side of the result

• Try using LibKey Nomad to simplify access (especially if browsing publisher sites)

Still stuck? See: Accessing full-text journal articles: a guide

The official help pages include detailed guides to each feature:
Semantic Scholar Product Overview

Covers:

• TL;DR summaries

• Ask This Paper

• Semantic Reader

• Author profiles and citation graphs

• Coverage and indexing policies