Summary of Conditional Vendi Score: An Information-theoretic Approach to Diversity Evaluation Of Prompt-based Generative Models, by Mohammad Jalali et al.
Conditional Vendi Score: An Information-Theoretic Approach to Diversity Evaluation of Prompt-based Generative Models
by Mohammad Jalali, Azim Ospanov, Amin Gohari, Farzan Farnia
First submitted to arxiv on: 5 Nov 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Information Theory (cs.IT)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper introduces a novel approach to quantify the diversity of samples generated by prompt-based text-conditioned generation models. The proposed method decomposes the entropy of the generated data into two components: conditional entropy, which measures the internal diversity of the model, and mutual information, which assesses the statistical relevance between the generated data and text prompts. Two new scores are introduced: Conditional-Vendi and Information-Vendi, which can be used to evaluate the prompt-induced and model-induced diversity in the generated samples. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper explores a way to measure the diversity of images and videos generated by AI models. It develops a new method to calculate this diversity based on how much the generated data varies when different text prompts are given. The approach is tested using various experiments, showing that it can effectively quantify the internal diversity of the model. |
Keywords
» Artificial intelligence » Prompt
