MechaFormer: Sequence Learning for Kinematic Mechanism Design Automation
Diana Bolanos, Mohammadmehdi Ataei, Pradeep Kumar Jayaraman · 2025
Read MechaFormer: Sequence Learning for Kinematic Mechanism Design Automation by Diana Bolanos, Mohammadmehdi Ataei, Pradeep Kumar Jayaraman (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-09005.txt.
Description
Bolanos, Ataei, and Jayaraman introduce MechaFormer, a transformer model that designs mechanical linkages to trace a wanted path. Picking a mechanism means searching a huge space of layouts and sizes, a classic hard design problem. Their model treats design as sequence generation. Feed it a target curve, and it writes out a description of a mechanism in a purpose-built language. It fixes both the layout and the dimensions in one pass.
The model beat existing baselines on path-matching accuracy and produced a wide variety of new, workable designs. Different sampling settings let designers trade quality against diversity. Its outputs also proved excellent starting points for traditional optimizers, which then polished them quickly. The authors present the hybrid route as the most practical way to use the tool.
Catalog record
| CREATOR | Diana Bolanos, Mohammadmehdi Ataei, Pradeep Kumar Jayaraman |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 43 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-09005