Summary of Score Distillation Via Reparametrized Ddim, by Artem Lukoianov et al.
Score Distillation via Reparametrized DDIM
by Artem Lukoianov, Haitz Sáez de Ocáriz Borde, Kristjan Greenewald, Vitor Campagnolo Guizilini, Timur Bagautdinov, Vincent Sitzmann, Justin Solomon
First submitted to arxiv on: 24 May 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Graphics (cs.GR); Machine Learning (cs.LG)
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 This paper investigates the limitations of 3D shape generation methods like Score Distillation Sampling (SDS) built on 2D diffusion models. While these methods generate realistic 2D images, they produce cartoon-like, over-smoothed shapes in 3D. The authors show that SDS can be understood as a high-variance version of Denoising Diffusion Implicit Models (DDIM) with differently-sampled noise terms. This excessive variance leads to over-smoothing and unrealistic outputs. To address this issue, the authors propose modifying SDS by inverting DDIM in each update step, which removes over-smoothing and preserves higher-frequency detail. The modified method achieves better or similar 3D generation quality compared to other state-of-the-art Score Distillation methods without training additional neural networks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps us understand why some 2D image models don’t work well in 3D. It shows that a method called Score Distillation Sampling (SDS) can be improved by using another model called Denoising Diffusion Implicit Models (DDIM). This modification makes the SDS method better at generating realistic 3D shapes without needing extra training or special equipment. |
Keywords
» Artificial intelligence » Diffusion » Distillation
