Semantic Scholar, an AI tool to complement your literature search

Note to students: before using AI tools in your coursework, check what is allowed in your course guide. By default, using AI generated content is considered fraud. You can find more information about using AI on this student support page.

What is Semantic Scholar?

Semantic Scholar is a tool developed by the Allen Institute for AI. It is a search engine that uses machine learning and natural language to retrieve relevant papers. Some features rely on generative AI as well.

Coverage

Semantic Scholar covers scientific literature, mostly in the fields of biology, medicine, geography, history, social sciences, and engineering. Note that you may not have access to all papers indexed by Semantic Scholar, although WUR Library subscribes to a several journals. You can always double check if you can access the full text by searching for a specific title in WUR Library Search.

Interesting features

• You can search for publications with natural language. You can also refine your search with a few filters

• TLDRs (Too Long; Didn’t Read): you will get a short AI-generated summary of the scientific publication you selected. Note that this feature is in beta and is not available for all publications. You will most likely be able to use it on papers related to medicine, biology and computer sciences.
• Ask this paper: you can ‘ask’ a publication questions such as “what methods were used?” or “what were the main results?”. You will receive an AI-generated answer based on the information in that specific publication. Note that this feature is also in beta and is not available for all publications.
• Skimming highlights (via the semantic reader): some key elements will be highlighted directly in the text of the article and summarized for you. This includes the paper's goal, the methods used and the main results.

Other interesting features are paper recommendations (you will need to create an account), influential citations and topic pages.

Limitations

Using AI when searching for literature has its limitations:

• As with any AI tool, the output may include factual errors that could be difficult to spot.
• Semantic Scholar indexes several publications, but publications behind a paywall are not included or are partially included (only the title, abstract and keywords). This means that you must use Semantic Scholar in combination with a search in a bibliographic database.
• Your results will be potentially biased, depending on how your search is formulated, what articles are available within Semantic Scholar, or how the search engine was trained to interpret your searches in natural language. This is an additional reason why you should also perform systematic searches in bibliographic databases in addition to using Semantic Scholar.

Want to know more?

Semantic Scholar has an extensive FAQ as well as some video tutorials that explain more about the tool and how you can use it to its full capacity.