Back to Glossary
Feedforward Network
Models & Architectures
Layers without feedback; data flow only forward.
Feedforward networks (multilayer perceptrons) transmit information from input to output without feedback loops.
- Components: Linear projections and nonlinear activations.
- Applications: Classification, regression, tabular data.
- Design aspects: Depth vs. width, regularization, initialization.