arXiv AI recent: eCNNTO: A Highly Generalizable ConvNet for Accelerating Topology Optimization
The authors propose eCNNTO, an element-based convolutional neural network designed to accelerate density-based topology optimization. eCNNTO replaces per‑element deep belief networks with...
The paper describes how topology optimization typically requires many finite‑element analysis iterations, creating a computational bottleneck. eCNNTO builds on prior work by Kallioras et al.