FactoidL is meant to be an open source, free to use codebase to triangulate third-party content with a select group of databases (e.g., Wikipedia/DBPedia) in order to test the accuracy of the content. Essentially, this is an open source fact-checking algorithm in various formats (browser extension, mobile apps, web apps, etc.). The mission is to make fact-checking open and accessible to all so they can quickly and accurately gauge how truthful the content they are reading is compared to public human knowledge.
You can contact us at the official email: factoidlproject@gmail.com
We'd love to have those interested/experienced in Natural Language Processing (NLP) areas, like semantic search. Currently, we are focused on building a Google Chrome extension, a mobile app (PWA), and an Electron app for desktop.
You can also spread the love on Twitter!
The easiest way to use FactoidL is to add the Chrome Extension.
Normally, extensions are packed to obfuscate code and so it is condensed into a neat package for downloading from say, the Chrome Web Store. However, if you must add this extension "unpacked" instead, instructions are as follows:
- Download all of the files in this repo to a local directory on your computer.
- Enable Developer Mode in the latest version of Chrome. It is found under Settings → Extensions → Check Developer Mode. Do not navigate away from this page.
- On the same Extensions page, click "Load unpacked extension..." and select the local directory with all of the FactoidL files from step 1.
- Now, the FactoidL icon should show next to your omnibox (in the top right, where extensions are usually found).
- Navigate to a page (example_fact_content.html is provided as an example article that is periodically updated to simulate fact-checking a webpage) and click on the FactoidL extension icon.
- If an error is shown on the popup, try refreshing the page you are currently on. Otherwise, statistics about the facts checked should show up on the extension popup.
Alexander Kidd (Lead Dev)
-
Compromise.js (Awesome all-purpose NLP for js)
-
Wikipedia (Crowd-sourced information!)
-
JQuery (Who said you didn't need it?)