Sign in

Joe Barrow

@jbarrow.bsky.social
126 followers 191 following 23 posts

NLP @ Pattern Data Prev: Adobe Research, PhD UMD

PostsRepliesMedia
Joe Barrow @jbarrow.bsky.social · 24/09/2025
Paper thread of some work I’m *incredibly* proud of, my first single author paper! Converting a PDF to a fillable form is a hard problem, and a lot of solutions don’t work very well! In CommonForms, I show that you can train models that outperform Adobe Acrobat for <$500! 🧵
131
Joe Barrow @jbarrow.bsky.social · 23/04/2025
In which I argue that LLM-generated bounding boxes are impressive, but not that useful (yet): notes.penpusher.app/Misc/Horsesh...
notes.penpusher.app
Horseshoes (and Hand Grenades) - LLM Localization is not Close, but not Close Enough - Joe Barrow
TL;DRLarge Multimodal Models (LMMs) can now output bounding boxes when given images as inputs. The results are impressive, but for documents they aren't good enough for real world use, yet. The Probl…
020
Joe Barrow @jbarrow.bsky.social · 07/01/2025
Ah, yes, that ol' familiar unit of measure "AI TOPS"
140
Joe Barrow @jbarrow.bsky.social · 20/12/2024
I put together a little guide on getting started with Google Gemini -- how to make multimodal calls, get structured outputs, and image bounding boxes to build an object detector. notes.penpusher.app/Misc/Google+...
publish.obsidian.md
Google Gemini 101 - Object Detection with Vision and Structured Outputs - Joe Barrow - Obsidian Publish
This is a missing manual for how to get a simple working prototype up and running with Gemini's vision mode and structured outputs. I'm confident that manual exists elsewhere, but I haven't been able…
230
Reposted by Joe Barrow
Joe Barrow @jbarrow.bsky.social · 18/12/2024
Drake meme template.
No to: clear, concise prose
Yes to: negative vspace
031
Joe Barrow @jbarrow.bsky.social · 18/12/2024
Gemini 2.0 Flash is pretty good at localization in images. for an LMM (much better than GPT-4o in my experiments).
A picture of a teapot to the right of a teacup, both in a flat-bottomed basket. The teacup has eta and flowers in it. There are 2 blue bounding boxes on the image, one labeled "Teapot" and one labeled "Teacup" that are over the teapot and teacup respectively.
010
Joe Barrow @jbarrow.bsky.social · 09/12/2024
ML history question: is there an earlier reference to pixel-only in-context (i.e. no fine-tuning) DocVQA performance than the GPT-4 announcement from OpenAI?
000