Summary of Guided Trajectory Generation with Diffusion Models For Offline Model-based Optimization, by Taeyoung Yun et al.
Guided Trajectory Generation with Diffusion Models for Offline Model-based Optimization
by Taeyoung Yun, Sujin Yun, Jaewoo Lee, Jinkyoo Park
First submitted to arxiv on: 29 Jun 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
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 conditional generative modeling approach to optimize complex and high-dimensional black-box functions. This offline model-based optimization (MBO) method aims to find designs that maximize target functions using only pre-existing offline datasets. The approach constructs synthetic trajectories toward high-scoring regions while injecting locality bias for consistent improvement directions. A conditional diffusion model is then trained to generate trajectories conditioned on their scores. The method samples multiple trajectories from the trained model with guidance, exploring high-scoring regions beyond the dataset and selecting high-fidelity designs among generated trajectories using a proxy function. Extensive experiments demonstrate that this approach outperforms competitive baselines on Design-Bench and its practical variants. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps scientists and engineers optimize complex functions without needing to test every possibility. The authors created a new way to use existing data to find the best solutions by generating many possible paths forward. They then chose the most promising paths based on how well they did. This approach was tested and found to be better than others at finding good solutions. |
Keywords
* Artificial intelligence * Diffusion model * Optimization
