FunCAD-Harness checks whether LLM CAD designs actually function
Original title: Can It Actually Work? Synthesizing Functional CAD Mechanisms
FunCADBench spans 46 mechanism families; the harness turns abstract requirements into candidate-specific physical tests.
Researchers propose FunCAD-Harness for functional CAD design, where LLM agents interpret underspecified requests and revise mechanisms using physical simulation. The framework converts abstract requirements into candidate-specific tests because different designs can realize the same function with different parts and sequences. On FunCADBench, covering 46 mechanism families with interdependent multi-step operations, the authors report higher functional success than standard simulation-augmented baselines and failures that visual inspection misses.
Unconfirmed: Functional-success gains are author-reported and not independently verified here.
- FunCADBench covers 46 mechanism families with multi-step operations.
- Agents iteratively revise designs from simulation feedback.
- Authors say the harness beats standard simulation-augmented baselines.
- The paper does not name the frontier models used in the experiments.
From the original article
1 min read at arxiv.orgWhile LLM agents can generate visually plausible CAD assemblies, these designs often fail to fulfill their intended physical functions. We study functional CAD design, a setting where agents interpret underspecified requests to construct and iteratively revise mechanisms that satisfy functional requirements in physical simulation.
By Wenqian Zhang · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.