Royi Rassin Profile picture
Oct 20, 2022 8 tweets 4 min read Read on X
New work! :D

We show evidence that DALL-E 2, in stark contrast to humans, does not respect the constraint that each word has a single role in its visual interpretation.

Work with @ravfogel and @yoav.
BlackboxNLP @ #emnlp2022

Below, "a person is hearing a bat" Image
We detail three types of behaviors that are inconsistent with the single role per word constraint
(1) A noun with multiple senses in an ambiguous prompt may cause DALL-E 2 to generate the same noun twice, but with different senses. In “a bat is flying over a baseball stadium”, “bat” is visualized as both: a flying mammal and a wooden stick. Image
(2) A noun may be interpreted as a modifier of two different entities. For “a fish and a gold ingot”, DALL-E 2 will consistently include a goldfish, despite “gold” being a modifier of “ingot”, not “fish” Image
(3) A noun is interpreted simultaneously as an entity and as a modifier of a different entity. For “a seal is opening a letter”, DALL-E 2 tends to attribute the “sealed” property to the letter Image
Bonus finding: some nouns can be replaced with their description, and still lead to the first behavior. “crane” is not explicitly mentioned in “a tall, long-legged, long-necked bird and a construction site”, but two types of crane appear for its interpretation Image
If you’re interested to read more about this cool phenomenon, give our paper a read!

Paper: arxiv.org/pdf/2210.10606…
I tagged the wrong Yoav.
Shared work with @yoavgo and @ravfogel

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Royi Rassin

Royi Rassin Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @RoyiRassin

Nov 7
How diverse are the outputs of text-to-image models and how can we measure that? In our new work, we propose a measure based on LLMs and Visual-QA (VQA), and show NONE of the 12 models we experiment with are diverse. 🧵
1/11 Image
Project page (with interactive results to explore): royira.github.io/GRADE/
Code: github.com/RoyiRa/GRADE-Q…
HF: huggingface.co/papers/2410.22…
arXiv: arxiv.org/abs/2410.22592
In the world, cookies come in many shapes and kinds, but generated cookies are consistently round and chocolate chip. This happens for many other concepts as well. How can we estimate and quantify the distribution of such behaviors more systematically 🤔?
2/11
Read 12 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(