Autoregressive Generation of Static and Growing Trees
This work addresses tree generation for computer graphics and simulation applications, but appears incremental as it builds on existing transformer and multi-resolution approaches.
The authors tackled the problem of generating complex trees with transformers by proposing a multi-resolution hourglass architecture with skip connections, achieving faster processing speed and lower memory consumption while enabling conditional generation and growth simulation.
We propose a transformer architecture and training strategy for tree generation. The architecture processes data at multiple resolutions and has an hourglass shape, with middle layers processing fewer tokens than outer layers. Similar to convolutional networks, we introduce longer range skip connections to completent this multi-resolution approach. The key advantage of this architecture is the faster processing speed and lower memory consumption. We are therefore able to process more complex trees than would be possible with a vanilla transformer architecture. Furthermore, we extend this approach to perform image-to-tree and point-cloud-to-tree conditional generation and to simulate the tree growth processes, generating 4D trees. Empirical results validate our approach in terms of speed, memory consumption, and generation quality.