AI Club at CJS
Welcome to the Columbia Journalism School AI Club! We meet every other Monday in the Brown Institute (1st floor Pulitzer Hall).
This week, we’ll explore text embeddings (aka semantic maps). Imagine a government agency suddenly releases a large trove of pdf documents. There’s no way you can reasonably read millions of documents all at once so you’re left guessing at what might be in them. You search for keywords you think might surface documents with news value. But what if there was a way to leverage AI to show you all the documents at once, grouping similar ones together and allowing you to quickly get a sense of the whole corpus? Researchers at the Tow Center will show you a few examples of semantic maps in action (including of the Epstein files) and share some code that can help you make your own. As usual, there will be pizza.