Multilingual Conceptual Coverage in Text-to-Image Models
This addresses the issue of evaluating how well generative models generalize across languages for researchers and developers, though it is incremental as it builds on existing benchmarking methods.
The paper tackles the problem of assessing multilingual parity in text-to-image models by proposing CoCo-CroLa, a technique to benchmark conceptual coverage across languages, and demonstrates its use for identifying weaknesses and biases without prior assumptions.
We propose "Conceptual Coverage Across Languages" (CoCo-CroLa), a technique for benchmarking the degree to which any generative text-to-image system provides multilingual parity to its training language in terms of tangible nouns. For each model we can assess "conceptual coverage" of a given target language relative to a source language by comparing the population of images generated for a series of tangible nouns in the source language to the population of images generated for each noun under translation in the target language. This technique allows us to estimate how well-suited a model is to a target language as well as identify model-specific weaknesses, spurious correlations, and biases without a-priori assumptions. We demonstrate how it can be used to benchmark T2I models in terms of multilinguality, and how despite its simplicity it is a good proxy for impressive generalization.