Pith. sign in

REVIEW 1 cited by

Training Priors Predict Text-To-Image Model Performance

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2306.01755 v2 pith:HKBZ5B5D submitted 2023-05-23 cs.CV cs.AIcs.CL

classification cs.CVcs.AIcs.CL
keywords trainingastronautgeneratehorsealigneddataimagemodel
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Text-to-image models can often generate some relations, i.e., "astronaut riding horse", but fail to generate other relations composed of the same basic parts, i.e., "horse riding astronaut". These failures are often taken as evidence that models rely on training priors rather than constructing novel images compositionally. This paper tests this intuition on the stablediffusion 2.1 text-to-image model. By looking at the subject-verb-object (SVO) triads that underlie these prompts (e.g., "astronaut", "ride", "horse"), we find that the more often an SVO triad appears in the training data, the better the model can generate an image aligned with that triad. Here, by aligned we mean that each of the terms appears in the generated image in the proper relation to each other. Surprisingly, this increased frequency also diminishes how well the model can generate an image aligned with the flipped triad. For example, if "astronaut riding horse" appears frequently in the training data, the image for "horse riding astronaut" will tend to be poorly aligned. Our results thus show that current models are biased to generate images with relations seen in training, and provide new data to the ongoing debate on whether these text-to-image models employ abstract compositional structure in a traditional sense, or rather, interpolate between relations explicitly seen in the training data.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Relations, Negations, and Numbers: Looking for Logic in Generative Text-to-Image Models

    cs.CV 2024-11 conditional novelty 4.0 of 10

    DALL-E 3 fails to reliably produce images matching prompts about physical relations, negations, and exact numbers beyond three, and a grounded diffusion pipeline does worse on relations.

Pith tools