Summary of Dstruct2design: Data and Benchmarks For Data Structure Driven Generative Floor Plan Design, by Zhi Hao Luo et al.
DStruct2Design: Data and Benchmarks for Data Structure Driven Generative Floor Plan Design
by Zhi Hao Luo, Luis Lara, Ge Ya Luo, Florian Golemo, Christopher Beckham, Christopher Pal
First submitted to arxiv on: 22 Jul 2024
Categories
- Main: Computation and Language (cs.CL)
- 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 This paper explores text-conditioned generative models for floorplan generation, a specific type of raster image generation task. While current approaches focus on aesthetic results, the authors highlight the importance of numerical properties in many use cases. They propose an intermediate data structure containing these properties to generate final floorplan images while respecting constraints like room sizes. The paper constructs a new dataset and provides tools for converting procedurally generated ProcTHOR floorplan data into this format. It also introduces metrics and benchmarks to evaluate constraint-respecting samples from models. Finally, the authors demonstrate the feasibility of using large language model (LLM) Llama3 for floorplan generation respecting numerical constraints. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about using computers to generate pictures of buildings that follow certain rules. Instead of just making it look nice, we want to make sure certain things are a specific size or in a specific place. Right now, the way we do this isn’t very good. The authors created new data and ways to measure how well these computer models do at following these rules. They also tested a special kind of language model that can help with this problem. |
Keywords
» Artificial intelligence » Image generation » Language model » Large language model
