[1] Mao W W[M]. Modern optical lens design: methods and examples(2013).

[2] Meiron J. Damped least-squares method for automatic lens design[J]. Journal of the Optical Society of America, 55, 1105-1109(1965).

[3] Sun J F, Li X T. Automatic design of machine vision lens based on improved genetic algorithm and damped least squares[J]. Proceedings of SPIE, 11895, 118950X(2021).

[4] Glatzel E, Wilson R. Adaptive automatic correction in optical design[J]. Applied Optics, 7, 265-276(1968).

[5] Forbes G W, Jones A E. Towards global optimization with adaptive simulated annealing[J]. Proceedings of SPIE, 1354, 144-153(1991).

[6] Höschel K, Lakshminarayanan V. Genetic algorithms for lens design: a review[J]. Journal of Optics, 48, 134-144(2019).

[7] Tang Y H, Wu Q Y, Chen X Y et al. Optimization design of the meridian line of progressive addition lenses based on genetic algorithm[J]. Acta Optica Sinica, 34, 0922005(2014).

[8] Isshiki M, Ono H, Hiraga K et al. Lens design: global optimization with escape function[J]. Optical Review, 2, 463-470(1995).

[9] Menke C. Application of particle swarm optimization to the automatic design of optical systems[J]. Proceedings of SPIE, 10690, 106901A(2018).

[10] Li Y H, Lü J, Jiang L et al. Reverse design of photonic devices based on a hybrid particle swarm algorithm[J]. Laser & Optoelectronics Progress, 59, 1113001(2022).

[11] Yang T, Jin G F, Zhu J. Automated design of freeform imaging systems[J]. Light: Science & Applications, 6, e17081(2017).

[12] Yang T, Cheng D W, Wang Y T. Direct generation of starting points for freeform off-axis three-mirror imaging system design using neural network based deep-learning[J]. Optics Express, 27, 17228-17238(2019).

[13] Gannon C, Liang R G. Using machine learning to create high-efficiency freeform illumination design tools[EB/OL]. https://arxiv.org/abs/1903.11166

[14] Côté G, Lalonde J F, Thibault S. Extrapolating from lens design databases using deep learning[J]. Optics Express, 27, 28279-28292(2019).

[15] Zhang H, Chen J W, Hu Y J et al. Freeform light distribution design based on deep learning and length-energy mapping[J]. Acta Optica Sinica, 42, 1022003(2022).

[16] Berthelot David. MixMatch: a holistic approach to semi-supervised learning[EB/OL]. https://arxiv.org/abs/1905.02249

[17] Sharma P, Singh A. Era of deep neural networks: a review[C](2017).

[18] Chen Y R, Xie Y, Song L H et al. A survey of accelerator architectures for deep neural networks[J]. Engineering, 6, 264-274(2020).

[19] Kiliçarslan S, Adem K, Çelİk M. An overview of the activation functions used in deep learning algorithms[J]. Journal of New Results in Science, 10, 75-88(2021).

[20] Klambauer G, Unterthiner T, Mayr A et al. Self-normalizing neural networks[C], 971-980(2017).

[21] Sun K L, Yu J M, Zhang L et al. A convolutional neural network model based on improved softplus activation function[M]. Abawajy J H, Choo K K R, Islan R, et al. International conference on applications and techniques in cyber intelligence ATCI 2019. Advances in intelligent systems and computing, 1017, 1326-1335(2019).

[22] Liang Q T[M]. Physical optics, 41-42(2012).