AutoTransfer: Subject Transfer Learning with Censored Representations on Biosignals Data
Authors:
- Niklas Smedemark-MarguliesNSNiklas Smedemark-Margulies
- Ye WangYWYe Wang
- Toshiaki Koike-AkinoTKToshiaki Koike-Akino
- Deniz ErdogmusDEDeniz Erdogmus
Preprint
Abstract
We provide a regularization framework for subject transfer learning in which we seek to train an encoder and classifier to minimize classification loss, subject to a penalty measuring independence between the latent representation and the subject label. We introduce three notions of independence and corresponding penalty terms using mutual information or divergence as a proxy for independence. For each penalty term, we provide several concrete estimation algorithms, using analytic methods as well as neural critic functions. We provide a hands-off strategy for applying this diverse family of regularization algorithms to a new dataset, which we call "AutoTransfer". We evaluate the performance of these individual regularization strategies and our AutoTransfer method on EEG, EMG, and ECoG datasets, showing that these approaches can improve subject transfer learning for challenging real-world datasets.Keywords
All Related Versions
This publication has 4 references indexed in Scilit:
- Learning Invariant Representations From EEG via Adversarial InferenceIEEE Access, 2020
- A library of human electrocorticographic data and analysesNature Human Behaviour, 2019
- EEGNet: a compact convolutional neural network for EEG-based brain–computer interfacesJournal of Neural Engineering, 2018
- Improving EEG-Based Emotion Classification Using Conditional Transfer LearningFrontiers in Human Neuroscience, 2017