Transfer learning

Definition

The definition of transfer learning is given in terms of domains and tasks. A domain ${\displaystyle \mathcal{D}}$ consists of: a feature space ${\displaystyle \mathcal{X}}$ and a marginal probability distribution ${\displaystyle P(X)}$, where ${\displaystyle X=\{x_1,...,x_n\}\in \mathcal{X}}$. Given a specific domain, ${\displaystyle \mathcal{D}=\{\mathcal{X},P(X)\}}$, a task consists of two components: a label space ${\displaystyle \mathcal{Y}}$ and an objective predictive function ${\displaystyle f:\mathcal{X}\to \mathcal{Y}}$. The function ${\displaystyle f}$ is used to predict the corresponding label ${\displaystyle f(x)}$ of a new instance ${\displaystyle x}$. This task, denoted by ${\displaystyle \mathcal{T}=\{\mathcal{Y},f(x)\}}$, is learned from the training data consisting of pairs ${\displaystyle \{x_i,y_i\}}$, where ${\displaystyle x_i\in \mathcal{X}}$ and ${\displaystyle y_i\in \mathcal{Y}}$.

Given a source domain ${\displaystyle {\mathcal{D}}_S}$ and learning task ${\displaystyle {\mathcal{T}}_S}$, a target domain ${\displaystyle {\mathcal{D}}_T}$ and learning task ${\displaystyle {\mathcal{T}}_T}$, where ${\displaystyle {\mathcal{D}}_S\ne {\mathcal{D}}_T}$, or ${\displaystyle {\mathcal{T}}_S\ne {\mathcal{T}}_T}$, transfer learning aims to help improve the learning of the target predictive function ${\displaystyle f_T(\cdot )}$ in ${\displaystyle {\mathcal{D}}_T}$ using the knowledge in ${\displaystyle {\mathcal{D}}_S}$ and ${\displaystyle {\mathcal{T}}_S}$.

Applications

Algorithms for transfer learning are available in Markov logic networks and Bayesian networks. Transfer learning has been applied to cancer subtype discovery, building utilization, general game playing, text classification, digit recognition, medical imaging and spam filtering.

In 2020, it was discovered that, due to their similar physical natures, transfer learning is possible between electromyographic (EMG) signals from the muscles and classifying the behaviors of electroencephalographic (EEG) brainwaves, from the gesture recognition domain to the mental state recognition domain. It was noted that this relationship worked in both directions, showing that EEG brainwaves can likewise be used to classify EMG signals. The experiments noted that the accuracy of neural networks and convolutional neural networks were improved through transfer learning both prior to any learning (compared to standard random weight distribution) and at the end of the learning process (asymptote). That is, results are improved by exposure to another domain. Moreover, the end-user of a pre-trained model can change the structure of fully-connected layers to improve performance.