You can start a search by inserting a topic, article name or author in the search bar. Then, the AI model looks up and evaluates literature using a machine learning algorithm. While conventional search models use textual similarity to provide relevant articles, Semantic Scholar uses machine learning to take into account more parameters for each search. This way, the tool can provide more meaningful and relevant results for your search.

In the page header, several search filters such as date range, journal and author. Moreover, results can be sorted by relevance, recency, number of citations and the influence. The most influential papers are determined by a machine learning algorithm, combining the citation count, and the context in which the respective paper is cited. More information on how the influence of papers is determined can be found here.

After finding a relevant paper, Semantic Scholar offers multiple functionalities that help you gain quick understanding of a topic.

• TLDR: The Too Long, Didn't Read feature gives a short AI generated summary of a selected paper. This feature is still in beta and therefore does not yet summarize all papers.
• A selection of Related Papers
• Ask this paper (beta): Semantic Scholar also has an AI-tool that helps you with quickly searching for the answers to specific questions in the paper. The answers to your questions can easily be checked using the supporting statements option. As this tool is in beta, it cannot yet be used on all articles.
• Semantic reader: By highlighting key elements,  the Semantic Reader helps to get quick insight into the contents and structure of papers. Unfortunately, the Semantic reader is limited to papers published on ArXiv at this moment.
• Topic pages (beta): Topic pages help you to gain a quick understanding of a topic by giving an AI-generated explanation of the topic, accompanied by related topics and relevant papers on the topic.

In Semantic Scholar, you can easily archive your accessed literature in a library. You can access this library by clicking the arrow in the top right corner of the screen.

To organize your library, you can create folders. Another advantage of creating folders is the possibility to get recommendations on articles related to the articles saved in the folder.

The research dashboard gives an overview of publications that might be relevant for your research project. Based on the contents of the literature in your library, Semantic Scholar can present you new literature recommendations per library folder.

The research dashboard can be entered by clicking on the arrow in the top right corner and selecting Research Feeds. You can filter the feed per library folder, by selecting the desired folder on the right side of the screen. You can choose to save the recommended articles by pressing the save button beneath the respective article. These articles will be used as a positive signal for future recommendations. By pressing the Not Relevant, button, the model will also be trained to give more relevant recommendations in the future.

To get email notifications for your topics of interest, you can enable mail updates in the General Settings tab. Here you can select the research feeds you want to receive emails about, and set the mailing frequency.

You can easily download the Semantic Scholar extension in Firefox and Chrome. Then, by selecting text like an author name, a paper title or interesting keywords, you can start a search. In the example below, the name of our Rector Magnificus was highlighted on the WUR website. Then the blue and yellow Semantic Scholar icon was pressed in Chrome, opening the extension. By pressing Search, we are redirected to the Semantic Scholar site, with the search results for Prof. dr. Carolien Kroeze.