Competitive Network

A competitive network is a pre-configured subnetwork containing a competitive group with an input layer and training interface. The competitive group is where the learning algorithm and parameters are defined.

From the Simulations > Competitive > Competitive network (simple) simulation. This network has been trained on patterns P1-P5 and currently cannot distinguish P1-P3.

For details on the algorithm, parameters, and theory, see the competitive group page.

Structure

The competitive network consists of:

• An input layer (clamped)
• A competitive group
• All-to-all connections from input to competitive layer
• Training data management for unsupervised learning

Creation

When creating a competitive network, you specify:

• Number of inputs: Number of input neurons.
• Number of competitive neurons: Number of neurons in the competitive layer.

For competitive group parameters (update method, learning rate, etc.), see the competitive group documentation.

Right Click Menu

Common right-click items are described on the subnetwork page.

• Add Current Pattern to Training Data: Add the current input pattern to the training dataset.
• Train on current pattern…: Opens a dialog to train the network on the current input pattern for a specified number of iterations.
• Train once on current pattern: Train the network for one iteration on the current input pattern. Keyboard shortcut: T
• Edit / Train Competitive: Opens the training dialog to edit and train the competitive network.
• Randomize: Randomize synapses connected to the competitive group. Keyboard shortcut: R

Training

Training a competitive network involves specifying input data and then running the algorithm. The general process is covered in Unsupervised Learning. Double-click the interaction box to open the training dialog.

Training sets input patterns, and the competitive group learns to represent clusters in the input data. See the competitive group documentation for algorithm details.