Researchers have developed Co-DSL, a framework enabling humans and Large Language Models (LLMs) to collaboratively design reusable spatial concepts for layout generation. The system allows users to create personalized operations by providing descriptions that LLMs translate into initial implementations, which users can then refine through feedback. In a subsequent stage, LLMs use these libraries to author layout programs that satisfy user requests by composing concepts into a concept-graph, which a solver then uses to find a suitable layout. This approach has demonstrated success across five spatial layout domains, enabling LLMs to capture complex layout patterns and providing users with high-level parametric controls for exploring layout variations. AI
IMPACT Enables more intuitive and flexible creation of complex visual layouts by combining human expertise with LLM capabilities.
RANK_REASON The cluster contains an academic paper detailing a new framework for AI-assisted design. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
