Fig. 5: Resource and design time comparison between deep-learning enabled inverse design against electromagnetic simulation-based heuristic (genetic algorithm) optimization. | Nature Communications

Fig. 5: Resource and design time comparison between deep-learning enabled inverse design against electromagnetic simulation-based heuristic (genetic algorithm) optimization.

From: Deep-learning enabled generalized inverse design of multi-port radio-frequency and sub-terahertz passives and integrated circuits

Fig. 5

By removing the constraints of iterative EM simulations, the proposed method can drastically reduce the time of synthesis. Moreover, once a model trained, it can be utilized for different design goals and synthesize EM structures within minutes, which is in stark contrast with EM-based optimization.

Back to article page