Finding Publications Method

Criteria Node: Captures and stores search criteria specific to the research focus, such as Dual Targeting. This Node can be easily updated or modified for different topics, allowing the Method to be rerun seamlessly for various sections or evolving research needs
Prompt Instructions Node: Provides the instructions for the AI to follow based on the Criteria Node. This can be as simple as following the instructions in the Criteria Node all the way to a Mega Prompt. In our dual targeting section, the Criteria Node contained numbered bullets for specific topics that the AI created several keyword searches for each.
Generate Canvas Output Node: Generates a Canvas with several Nodes each containing a Keyword search based on the Criteria Node and Prompt Node. In our example, for each numbered bullet in the Criteria Node, a horizontal line of 5 Nodes were created with Keyword searches as Node titles.
Loop Node: An instructional Node that directs the AI to loop through all Nodes in the previous Canvas, ensuring comprehensive coverage of each keyword search.
Paper Finder Node: The AI is instructed to connect with the Semantic Scholar API, executing the keyword searches from #3, Generate Canvas Output Node.
Paper Generation Node: For each keyword search, the AI outputs corresponding papers into a new Canvas, creating a Node for each paper and the same format of #3, Generate Canvas Output Node.
Loop Node: An instructional Node that directs the AI to loop through all Nodes in the previous Canvas, ensuring comprehensive coverage of each Research Paper from #6 Paper Generation Node
Inclusion/Exclusion Context Node: Relevant context for the AI to use to evaluate whether to inlude each paper into the Method or exclude it. In the Dual Targeting section, we put the entire Outline of the paper as context.
Prompt Instructions Node: Provides instructions for the AI to follow as it loops through each paper from the #6 Paper Generation Node and uses the context provided in the #7 Inclusion/Exclusion Context Node. In the Dual Tageting section, the promp instructs the AI to evaluate each paper’s abstract according to the outline to see if its relevant. It instructs the AI to set a high bar for relevance and output a “Yes/No” response along with a brief explanation on its reasoning.

Although not applied in this review, we plan to adopt the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 framework to enhance transparency and rigor in the inclusion and exclusion process. This will be especially useful in the 'Living Publications' section, where it will support consistent study selection and ensure well-documented updates over time

This Method can be run in its entirety or in parts, allowing users to manually adjust keyword Nodes or add papers for evaluation.

This Method will continue on to output a Canvas for only the included papers. This Canvas is shown in the next part of the Method for ‘Data Extractions Method’ where we loop through all of the papers that meet the inclusion criteria for analysis.

PreviousLiterature Review Method NextData Extraction Method

Last updated 7 months ago

Criteria Node: Captures and stores search criteria specific to the research focus, such as Dual Targeting. This Node can be easily updated or modified for different topics, allowing the Method to be rerun seamlessly for various sections or evolving research needs
Prompt Instructions Node: Provides the instructions for the AI to follow based on the Criteria Node. This can be as simple as following the instructions in the Criteria Node all the way to a Mega Prompt. In our dual targeting section, the Criteria Node contained numbered bullets for specific topics that the AI created several keyword searches for each.
Generate Canvas Output Node: Generates a Canvas with several Nodes each containing a Keyword search based on the Criteria Node and Prompt Node. In our example, for each numbered bullet in the Criteria Node, a horizontal line of 5 Nodes were created with Keyword searches as Node titles.
Loop Node: An instructional Node that directs the AI to loop through all Nodes in the previous Canvas, ensuring comprehensive coverage of each keyword search.
Paper Finder Node: The AI is instructed to connect with the Semantic Scholar API, executing the keyword searches from #3, Generate Canvas Output Node.
Paper Generation Node: For each keyword search, the AI outputs corresponding papers into a new Canvas, creating a Node for each paper and the same format of #3, Generate Canvas Output Node.
Loop Node: An instructional Node that directs the AI to loop through all Nodes in the previous Canvas, ensuring comprehensive coverage of each Research Paper from #6 Paper Generation Node
Inclusion/Exclusion Context Node: Relevant context for the AI to use to evaluate whether to inlude each paper into the Method or exclude it. In the Dual Targeting section, we put the entire Outline of the paper as context.
Prompt Instructions Node: Provides instructions for the AI to follow as it loops through each paper from the #6 Paper Generation Node and uses the context provided in the #7 Inclusion/Exclusion Context Node. In the Dual Tageting section, the promp instructs the AI to evaluate each paper’s abstract according to the outline to see if its relevant. It instructs the AI to set a high bar for relevance and output a “Yes/No” response along with a brief explanation on its reasoning.

This Method can be run in its entirety or in parts, allowing users to manually adjust keyword Nodes or add papers for evaluation.