… In designing the 116 structural optimization tasks, our goal was to create a distribution of
diverse, well-studied problems with real-world significance. We started with a selection of …

… We use reparameterized local optimization to design metagratings that deflect normally …
by a generative neural network and reparameterized to constrained physical devices. Network …

… As the design and control of these buildings represent a significant change in how buildings
… rately replicate the observed behavior of the physical system. Physics-informed dynamical …

… Several numerical examples for design optimization of truss structures under deflection
constraints are investigated to evaluate the efficiency and reliability of the proposed framework. …

… design. By not relying on training data, our approach provides a robust framework for tackling
complex optimization … are derived for physical Hessians, the reparametrization approach to …

… and graph embeddings generated using graph neural networks. In [6], … correlations among
QoR metrics to be optimized, which may lead to … The reparameterization trick is exploited to …

… for topology optimization lies in the introduction of physical … Through this reparameterization,
we achieve a generalized … : neural reparameterization, the introduction of design constraints…

… the latent and physical space device … our reparameterization method minimally impact
device performance. To compare the two design strategies, we perform fully freeform optimization …

… the geometric variance of designs but also induces the variability in the designs’ physics. As
… efficient design exploration and the construction of improved surrogate models for designs’ …

… 4a), the neural reparameterized optimization fits the low-frequency … the scheduled frequency
of neural re-parameterization helps the … 4.1, we design the baseline pixel optimization for …