Search for Papers (AKA Literature Reviews)

Elicit.ai logo If you have any questions or get stuck as you work through this in-class GenAI exercise, please ask the instructor for assistance. Have fun!

Elicit - Summaries of Top 4 Articles

Elicit AI is a research assistant tool that utilizes natural language processing to help users extract insights from academic papers and data. It makes a consolidated summary of the top 4 papers on the topic (in its judgment) and gives you links to 8 other related papers. Don’t depend on the summaries being 100% accurate!

Elicit Training Data sources: Web scraping (including Open Access data) & Indexing agreements with 50 publishers.

  1. Let’s do a search in Elicit on the same topic this time using more natural language.
    • Open Elicit, and if you want to use it you’ll need to create an account.
    • Type the following into the search bar and click the search button:
      How useful is informal credentialling for academic makerspace skills in student job searches?
    • Try searching for information about a topic that you are interested in to further explore the capabilities of Elicit. Be curious and have some fun!
  2. In Elicit, try doing a natural language search for a topic you know a lot about so that you can evaluate the quality of the results.
    • For example:
      Is informal credentialing helpful for academic makerspaces users in finding jobs?
    • Now try the same search in Google Scholar and compare the Elicit results with Google Scholar.
  3. Reflection time:
    • How useful do the articles look for your search?
    • Compared to the Google Scholar results how high is the quality of the articles Elicit found?
    • Does the combined 4 article summary look reasonable?
    • How can you verify the accuracy of the summary?
    • Does this look like a tool that could help you with your research?

ResearchRabbit - Make Connections from Groups of Articles

ResearchRabbit AI is a tool designed to help researchers and academics navigate and discover relevant literature efficiently. It uses AI algorithms to map out connections between research papers, enabling users to find related work, identify trends, and explore new research areas. It also creates article summaries, but don’t depend on the summaries being 100% accurate!

ResearchRabbit Training Data sources: Microsoft Academic Graph.

  1. Let’s do a search in ResearchRabbit on the same topic this time by adding one or two articles on the topic of your choice.
    • Open ResearchRabbit, and if you want to use it you’ll need to create an account.
    • Click the green Add Papers button, and then type the following into the search bar and click the search button (note that as of June 2024 ResearchRabbit does not support natural language queries, so we will do a keyword search. ResearchRabbit does use artificial intelligence for other aspects of its service):
      Informal credentialling academic makerspaces
    • Try using ResearchRabbit to explore a topic you’re interested in and hopefully find additional useful articles. Be curious and have some fun!
  2. Test ResearchRabbit on a topic you know a lot about:
    • In ResearchRabbit, try doing a keyword search for a topic you know a lot about so that you can evaluate the quality of the results.
    • Now try the same search in Google Scholar and compare the ResearhRabbit results with Google Scholar.
  3. Reflection time:
    • Do the suggested “Similar Work”, “Earlier Work”, and “Later Work” look relevant and useful?
    • Compared to the Google Scholar results how high is the quality of the articles ResearchRabbit found?
    • Does the network map of articles look like it might be helpful for you in your research?
    • How useful do the articles look informal credentialling for academic makerspace skills in student job searches?
    • Overall does this look like a tool that could help you with your research?

Semantic Scholar - Article Summaries

Semantic Scholar uses GenAI tools to create summaries of journal articles, but don’t depend on the summaries being 100% accurate! Semantic Scholar summaries are typically better than general GenAI tool summaries but can still “make stuff up” so you still need to make your own summaries for articles you’ll directly use in your research.

Semantic Scholar Training Data sources: Web scraping (including Open Access data) & Indexing agreements with 50 publishers.

  1. Let’s do a keyword search in Semantic Scholar on a topic of interest.
    • Open Semantic Scholar (no need to create an account unless you want to).
    • Type the following into the search bar and click the search button (note that as of June 2024 Semantic Scholar does not support natural language queries, so we will do a keyword search. Semantic Scholar does use Generative AI for other aspects of its service):
      Informal credentialling academic makerspaces
    • Try searching for information about a topic that you are interested in to further explore the capabilities of Perplexity. Be curious and have some fun!
  2. Test Semantic Scholar on a topic you know a lot about:
    • In Semantic Scholar, try doing a keyword search for a topic you know a lot about so that you can evaluate the quality of the results.
    • Now try the same search in Google Scholar and compare the Semantic Scholar results with Google Scholar.
  3. Reflection time:
    • How useful do the articles look informal credentialling for academic makerspace skills in student job searches?
    • Compared to the Google Scholar results how high is the quality of the articles Semantic Scholar found?
    • Do the article summaries look reasonable?
    • How can you verify the accuracy of the summary?
    • Does this look like a tool that could help you with your research?

NEXT STEP: Critically Review All GenAI Output