Autoencoder and restricted Boltzmann machine for transfer learning in functional magnetic resonance imaging task classification

Discovering individual fingerprints in resting‐state functional connectivity using deep neural networks

“…This approach may involve the direct transfer of taskrelated knowledge (in our case, individual identity) through supervised pretraining-based transfer learning, or the acquisition and transfer of knowledge unrelated to the downstream task from the source data via self-supervised pretraining-based transfer learning. Studies have shown that implementing transfer learning in deep learning models using fMRI data has enhanced model performance on downstream tasks (Hwang et al, 2023;Li et al, 2018;Rahman et al, 2022). Future studies using a deep learning model more suitable for dealing with FC data are warranted.…”

Adaptive habitat biogeography-based optimizer for optimizing deep CNN hyperparameters in image classification

“…These deep networks are different in terms of structure and connection weights. Many efforts, such as transfer learning, self-supervised learning, fuzzing with invocation ordering [ 23 , 24 ], fusion models [ 25 ], hierarchical models [ 26 ], adapting feature selection algorithms [ 27 ], subspace random optimization [ 28 ], multi-modal [ 29 ], and multi-label [ 30 ] techniques have improved the performance of these models [ 31 ].…”