temporal-graph-gen VS pal

It has been possible to generate impressive graphs from text since GPT-2. Though you need a few tricks to make it work.

Here's an example (my work): https://aclanthology.org/2021.naacl-main.67.pdf

TLDR of the input/output: https://madaan.github.io/res/tldr/graph_gen_tldr.jpg

On a related note, I did some work on generating temporal graphs from text: https://arxiv.org/pdf/2010.10077.pdf

Code/models: https://github.com/madaan/temporal-graph-gen

Using the terminology that I'm working with this is an example of a second-order analytic augmentation!

Here's another approach of second-order analytic augmentation, PAL: https://reasonwithpal.com

And third-order, Toolformer: https://arxiv.org/abs/2302.04761

The difference isn't in what is going on but rather with framing the approach within the analytic-synthetic distinction developed by Kant and the analytic philosophers who were influenced by his work. There's a dash of functional programming thrown in for good measure!

I have scribbled on a print-out of the article on my desk:

  Nth Order

GitHub: https://github.com/reasoning-machines/pal

tl;dr -- LLMs are bad at basic arithmetic and logic (as their opening examples with math word problems show), but they do much better if instead of asking them for the answer, you ask for code to compute the answer. Then evaluate or run the code to get the answer.