A mixture-density-based tandem optimization network for on-demand inverse design of thin-film high reflectors
Deep learning (DL) has emerged as a promising tool for photonic inverse design. Nevertheless, despite the initial success in retrieving spectra of modest complexity with nearly instantaneous readout, DL-assisted design methods often underperform in accuracy compared with advanced optimization techni...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
De Gruyter
2021-10-01
|
| Series: | Nanophotonics |
| Subjects: | |
| Online Access: | https://doi.org/10.1515/nanoph-2021-0392 |
| _version_ | 1811320123617705984 |
|---|---|
| author | Unni Rohit Yao Kan Han Xizewen Zhou Mingyuan Zheng Yuebing |
| author_facet | Unni Rohit Yao Kan Han Xizewen Zhou Mingyuan Zheng Yuebing |
| author_sort | Unni Rohit |
| collection | DOAJ |
| description | Deep learning (DL) has emerged as a promising tool for photonic inverse design. Nevertheless, despite the initial success in retrieving spectra of modest complexity with nearly instantaneous readout, DL-assisted design methods often underperform in accuracy compared with advanced optimization techniques and have not proven competitive in handling spectra of practical usefulness. Here, we introduce a tandem optimization model that combines a mixture density network (MDN) and a fully connected (FC) network to inversely design practical thin-film high reflectors. The multimodal nature of the MDN gives access to infinite candidate designs described by probability distributions, which are iteratively sampled and evaluated by the FC network to allow for rapid optimization. We show that the proposed model can retrieve the reflectance spectra of 20-layer thin-film structures. More interestingly, it reproduces with high precision the periodic structures of high reflectors derived from physical principles, even though no such information is included in the training data. Improved designs with extended high-reflectance zones are also demonstrated. Our approach combines the high-efficiency advantage of DL with the optimization-enabled performance improvement, enabling efficient and on-demand inverse design for practical applications. |
| first_indexed | 2024-04-13T12:53:45Z |
| format | Article |
| id | doaj.art-3a69ffa99f924ab8bbdc9fb92a288007 |
| institution | Directory Open Access Journal |
| issn | 2192-8614 |
| language | English |
| last_indexed | 2024-04-13T12:53:45Z |
| publishDate | 2021-10-01 |
| publisher | De Gruyter |
| record_format | Article |
| series | Nanophotonics |
| spelling | doaj.art-3a69ffa99f924ab8bbdc9fb92a2880072022-12-22T02:46:07ZengDe GruyterNanophotonics2192-86142021-10-0110164057406510.1515/nanoph-2021-0392A mixture-density-based tandem optimization network for on-demand inverse design of thin-film high reflectorsUnni Rohit0Yao Kan1Han Xizewen2Zhou Mingyuan3Zheng Yuebing4Walker Department of Mechanical Engineering, The University of Texas at Austin, Austin, TX78712, USAWalker Department of Mechanical Engineering, The University of Texas at Austin, Austin, TX78712, USADepartment of Statistics and Data Science, The University of Texas at Austin, Austin, TX78712, USADepartment of Statistics and Data Science, The University of Texas at Austin, Austin, TX78712, USAWalker Department of Mechanical Engineering, The University of Texas at Austin, Austin, TX78712, USADeep learning (DL) has emerged as a promising tool for photonic inverse design. Nevertheless, despite the initial success in retrieving spectra of modest complexity with nearly instantaneous readout, DL-assisted design methods often underperform in accuracy compared with advanced optimization techniques and have not proven competitive in handling spectra of practical usefulness. Here, we introduce a tandem optimization model that combines a mixture density network (MDN) and a fully connected (FC) network to inversely design practical thin-film high reflectors. The multimodal nature of the MDN gives access to infinite candidate designs described by probability distributions, which are iteratively sampled and evaluated by the FC network to allow for rapid optimization. We show that the proposed model can retrieve the reflectance spectra of 20-layer thin-film structures. More interestingly, it reproduces with high precision the periodic structures of high reflectors derived from physical principles, even though no such information is included in the training data. Improved designs with extended high-reflectance zones are also demonstrated. Our approach combines the high-efficiency advantage of DL with the optimization-enabled performance improvement, enabling efficient and on-demand inverse design for practical applications.https://doi.org/10.1515/nanoph-2021-0392artificial neural networksdeep learninginverse designnanophotonicsoptimizationthin-film optics |
| spellingShingle | Unni Rohit Yao Kan Han Xizewen Zhou Mingyuan Zheng Yuebing A mixture-density-based tandem optimization network for on-demand inverse design of thin-film high reflectors Nanophotonics artificial neural networks deep learning inverse design nanophotonics optimization thin-film optics |
| title | A mixture-density-based tandem optimization network for on-demand inverse design of thin-film high reflectors |
| title_full | A mixture-density-based tandem optimization network for on-demand inverse design of thin-film high reflectors |
| title_fullStr | A mixture-density-based tandem optimization network for on-demand inverse design of thin-film high reflectors |
| title_full_unstemmed | A mixture-density-based tandem optimization network for on-demand inverse design of thin-film high reflectors |
| title_short | A mixture-density-based tandem optimization network for on-demand inverse design of thin-film high reflectors |
| title_sort | mixture density based tandem optimization network for on demand inverse design of thin film high reflectors |
| topic | artificial neural networks deep learning inverse design nanophotonics optimization thin-film optics |
| url | https://doi.org/10.1515/nanoph-2021-0392 |
| work_keys_str_mv | AT unnirohit amixturedensitybasedtandemoptimizationnetworkforondemandinversedesignofthinfilmhighreflectors AT yaokan amixturedensitybasedtandemoptimizationnetworkforondemandinversedesignofthinfilmhighreflectors AT hanxizewen amixturedensitybasedtandemoptimizationnetworkforondemandinversedesignofthinfilmhighreflectors AT zhoumingyuan amixturedensitybasedtandemoptimizationnetworkforondemandinversedesignofthinfilmhighreflectors AT zhengyuebing amixturedensitybasedtandemoptimizationnetworkforondemandinversedesignofthinfilmhighreflectors AT unnirohit mixturedensitybasedtandemoptimizationnetworkforondemandinversedesignofthinfilmhighreflectors AT yaokan mixturedensitybasedtandemoptimizationnetworkforondemandinversedesignofthinfilmhighreflectors AT hanxizewen mixturedensitybasedtandemoptimizationnetworkforondemandinversedesignofthinfilmhighreflectors AT zhoumingyuan mixturedensitybasedtandemoptimizationnetworkforondemandinversedesignofthinfilmhighreflectors AT zhengyuebing mixturedensitybasedtandemoptimizationnetworkforondemandinversedesignofthinfilmhighreflectors |