Data attribution methods aim to quantify the influence of individual training samples on the prediction of artificial intelligence (AI) models.

We introduce a new approach, Generative mOdel finetun-ing with Reward-driven Sample selection (GORS), to boost the compositional text-to-image generation abilities of pretrained text-to-image models.

In this section, we provide a detailed description of the dataset following'datasheet for datasets' [

Work was done when interned at Zhipu AI.

Despite these strides, evaluating these models poses substantial challenges.

Explainable AI (XAI) is a rapidly growing domain with a myriad of proposed methods as well as metrics aiming to evaluate their efficacy.

This paper analyzes the impact of causal manner in the text encoder of text-to-image (T2I) diffusion models, which can lead to information bias and loss.