POViT: Vision Transformer for Multi-objective Design and Characterization of Nanophotonic Devices

• URL: http://arxiv.org/abs/2205.09045v1
• Date: Tue, 17 May 2022 01:58:34 GMT
• Title: POViT: Vision Transformer for Multi-objective Design and Characterization of Nanophotonic Devices
• Authors: Xinyu Chen, Renjie Li, Yueyao Yu, Yuanwen Shen, Wenye Li, Zhaoyu Zhang, Yin Zhang
• Abstract summary: We propose the first-ever Transformer model (POViT) to efficiently design and simulate semiconductor photonic devices. Our work has the potential to drive the expansion of EDA to fully automated photonic design.
• Score: 8.969089378686299
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: We solve a fundamental challenge in semiconductor IC design: the fast and accurate characterization of nanoscale photonic devices. Much like the fusion between AI and EDA, many efforts have been made to apply DNNs such as convolutional neural networks (CNN) to prototype and characterize next-gen optoelectronic devices commonly found in photonic integrated circuits (PIC) and LiDAR. These prior works generally strive to predict the quality factor (Q) and modal volume (V) of for instance, photonic crystals, with ultra-high accuracy and speed. However, state-of-the-art models are still far from being directly applicable in the real-world: e.g. the correlation coefficient of V ($V_{coeff}$ ) is only about 80%, which is much lower than what it takes to generate reliable and reproducible nanophotonic designs. Recently, attention-based transformer models have attracted extensive interests and been widely used in CV and NLP. In this work, we propose the first-ever Transformer model (POViT) to efficiently design and simulate semiconductor photonic devices with multiple objectives. Unlike the standard Vision Transformer (ViT), we supplied photonic crystals as data input and changed the activation layer from GELU to an absolute-value function (ABS). Our experiments show that POViT exceeds results reported by previous models significantly. The correlation coefficient $V_{coeff}$ increases by over 12% (i.e., to 92.0%) and the prediction errors of Q is reduced by an order of magnitude, among several other key metric improvements. Our work has the potential to drive the expansion of EDA to fully automated photonic design. The complete dataset and code will be released to aid researchers endeavoring in the interdisciplinary field of physics and computer science.

Related papers

• PACE: Pacing Operator Learning to Accurate Optical Field Simulation for Complicated Photonic Devices [14.671301859745453]
  Existing SOTA approaches, NeurOLight, struggle with predicting high-fidelity fields for real-world complicated photonic devices. We propose a novel cross-axis factorized PACE operator with a strong long-distance modeling capacity. Inspired by human learning, we conquer the simulation task for extremely hard cases into two progressively easy tasks.
  arXiv  Detail & Related papers  (2024-11-05T22:03:14Z)
• All-optical modulation with single-photons using electron avalanche [69.65384453064829]
  We demonstrate all-optical modulation using a beam with single-photon intensity. Our approach opens up the possibility of terahertz-speed optical switching at the single-photon level.
  arXiv  Detail & Related papers  (2023-12-18T20:14:15Z)
• Design and simulation of a transmon qubit chip for Axion detection [103.69390312201169]
  Device based on superconducting qubits has been successfully applied in detecting few-GHz single photons via Quantum Non-Demolition measurement (QND) In this study, we present Qub-IT's status towards the realization of its first superconducting qubit device.
  arXiv  Detail & Related papers  (2023-10-08T17:11:42Z)
• High-throughput identification of spin-photon interfaces in silicon [5.412164094894868]
  We use first principles computational screening to identify spin-photon interfaces among more than 1000 substitutional and interstitial charged defects in silicon. We evaluate the most promising defects by considering their optical properties, spin multiplicity, and formation energies. We identify three new promising spin-photon interface as potential bright emitters in the telecom band: $rm Ti_i+$, $rm Fe_i0$, and $rm Ru_i0$.
  arXiv  Detail & Related papers  (2023-03-02T21:35:17Z)
• NeurOLight: A Physics-Agnostic Neural Operator Enabling Parametric Photonic Device Simulation [17.295318670037886]
  A physics-agnostic neural framework, dubbed NeurOLight, is proposed to learn a family of frequency-domain Maxwell PDEs for ultra-fast parametric photonic device simulation. We show that NeurOLight generalizes to a large space of unseen simulation settings, demonstrates 2-orders-of-magnitude faster simulation speed than numerical solvers, and outperforms prior neural network models by 54% lower prediction error with 44% fewer parameters.
  arXiv  Detail & Related papers  (2022-09-19T21:25:26Z)
• ViTAEv2: Vision Transformer Advanced by Exploring Inductive Bias for Image Recognition and Beyond [76.35955924137986]
  We propose a Vision Transformer Advanced by Exploring intrinsic IB from convolutions, i.e., ViTAE. ViTAE has several spatial pyramid reduction modules to downsample and embed the input image into tokens with rich multi-scale context. We obtain the state-of-the-art classification performance, i.e., 88.5% Top-1 classification accuracy on ImageNet validation set and the best 91.2% Top-1 accuracy on ImageNet real validation set.
  arXiv  Detail & Related papers  (2022-02-21T10:40:05Z)
• An Empirical Study of Training End-to-End Vision-and-Language Transformers [50.23532518166621]
  We present METER(textbfMultimodal textbfEnd-to-end textbfTransformtextbfER), through which we investigate how to design and pre-train a fully transformer-based VL model. Specifically, we dissect the model designs along multiple dimensions: vision encoders (e.g., CLIP-ViT, Swin transformer), text encoders (e.g., RoBERTa, DeBERTa), multimodal fusion (e.g., merged attention vs. co-
  arXiv  Detail & Related papers  (2021-11-03T17:55:36Z)
• Numerical optimization of a nanophotonic cavity by machine learning for near-unity photon indistinguishability at room temperature [0.0]
  High indistinguishability (I) at room-temperature is difficult to obtain due to the intrinsic dephasing of most deterministic single-photon sources (SPS) Here we present a numerical demonstration of the design and optimization of a hybrid slot-Bragg nanophotonic cavity that achieves theoretical near-unity I and high coupling efficiency (beta) at RT for a variety of singlephoton emitters. Although the proposal is not a scalable technology, it can be suitable for experimental demonstration of single photon operation.
  arXiv  Detail & Related papers  (2021-10-28T10:18:57Z)
• A new concept for design of photonic integrated circuits with the ultimate density and low loss [62.997667081978825]
  We propose a new concept for design of PICs with the ultimate downscaling capability, the absence of geometric loss and a high-fidelity throughput. This is achieved by a periodic continuous-time quantum walk of photons through waveguide arrays. We demonstrate the potential of the new concept by reconsidering the design of basic building blocks of the information and sensing systems.
  arXiv  Detail & Related papers  (2021-08-02T14:23:18Z)
• Improved Transformer for High-Resolution GANs [69.42469272015481]
  We introduce two key ingredients to Transformer to address this challenge. We show in the experiments that the proposed HiT achieves state-of-the-art FID scores of 31.87 and 2.95 on unconditional ImageNet $128 times 128$ and FFHQ $256 times 256$, respectively.
  arXiv  Detail & Related papers  (2021-06-14T17:39:49Z)
• Classical and quantum regression analysis for the optoelectronic performance of NTCDA/p-Si UV photodiode [0.0]
  The performance of a fabricated Au/NTCDA/p-Si/Al photodiode was explained in details and showed an excellent responsivity. The fabricated photodiodes exhibited a linear current-irradiance relationship under illumination up to 65 $mW/cm2$.
  arXiv  Detail & Related papers  (2020-01-02T17:40:20Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.