Provides deep learning models optimized for high-dimensional omics data using mlr3torch and torch. Includes TabTransformer and autoencoder-based representation learning.
Details
These learners require the 'mlr3torch' and 'torch' packages to be installed. They are optional components that provide state-of-the-art deep learning capabilities for biomarker discovery.
Available architectures: - **MLP**: Multi-Layer Perceptron with dropout regularization - **TabTransformer**: Attention-based model for tabular data - **Autoencoder**: Unsupervised feature compression (PipeOp)