| 초록 |
Recently, as an emerging paradigm of material science, deep learning has shown its potential in various research fields; there have been attempts to apply deep learning in designing molecules’ structures, analyzing spectral data, and even sampling more accurate free energy landscapes. Among this unlimited potential of deep learning technology, image processing using deep learning has been particularly outstanding. Here, we propose deep learning algorithms to identify defects and quantify the ordering qualities in lamella- and cylinder-forming block copolymer films, replacing conventional defect inspection tools involved with either manual defect detections or inefficient image processing steps. We further use the developed deep learning neural networks for an inverse design of system parameters to achieve a targeting self-assembled structure. |
