The redesign of the Readout Layer, introduction of the Cluster Chain as the standard computation unit for the Readout Unit and One Takes All group. Enhanced data bundle folderization. Enhanced and revised code comments.

Author: okozelsk

Demo/DemoConsoleApp/Examples/ExampleBase.cs

@@ -1,17 +1,17 @@
-using System;
-using Demo.DemoConsoleApp.Log;
+using Demo.DemoConsoleApp.Log;
 using RCNet.CsvTools;
 using RCNet.Neural.Data;
 using RCNet.Neural.Network.NonRecurrent;
 using RCNet.Neural.Network.SM;
 using RCNet.Neural.Network.SM.Preprocessing;
 using RCNet.Neural.Network.SM.Preprocessing.Input;
+using RCNet.Neural.Network.SM.Readout;
 
 
 namespace Demo.DemoConsoleApp.Examples
 {
     /// <summary>
-    /// Base class of the implemented examples
+    /// Implements the base class of the examples.
     /// </summary>
     public class ExampleBase
     {
@@ -28,10 +28,10 @@ protected ExampleBase()
         //Methods
         //Event handlers
         /// <summary>
-        /// Displays information about the verification progress.
+        /// Displays an information about the verification progress.
         /// </summary>
-        /// <param name="totalNumOfInputs">Total number of inputs to be processed</param>
-        /// <param name="numOfProcessedInputs">Number of processed inputs</param>
+        /// <param name="totalNumOfInputs">The total number of inputs to be processed.</param>
+        /// <param name="numOfProcessedInputs">The number of already processed inputs.</param>
         protected void OnVerificationProgressChanged(int totalNumOfInputs, int numOfProcessedInputs)
         {
             //Display progress
@@ -43,15 +43,15 @@ protected void OnVerificationProgressChanged(int totalNumOfInputs, int numOfProc
         }
 
         /// <summary>
-        /// Displays information about the preprocessing progress and at the end displays important NeuralPreprocessor's statistics.
+        /// Displays an information about the preprocessing progress and at the end displays important NeuralPreprocessor's statistics.
         /// </summary>
-        /// <param name="totalNumOfInputs">Total number of inputs to be processed</param>
-        /// <param name="numOfProcessedInputs">Number of processed inputs</param>
-        /// <param name="finalPreprocessingOverview">Final overview of the preprocessing phase</param>
+        /// <param name="totalNumOfInputs">The total number of inputs to be processed.</param>
+        /// <param name="numOfProcessedInputs">The number of already processed inputs.</param>
+        /// <param name="finalPreprocessingOverview">The final overview of the preprocessing.</param>
         protected void OnPreprocessingProgressChanged(int totalNumOfInputs,
-                                                    int numOfProcessedInputs,
-                                                    NeuralPreprocessor.PreprocessingOverview finalPreprocessingOverview
-                                                    )
+                                                      int numOfProcessedInputs,
+                                                      NeuralPreprocessor.PreprocessingOverview finalPreprocessingOverview
+                                                      )
         {
             if (finalPreprocessingOverview == null)
             {
@@ -72,39 +72,38 @@ NeuralPreprocessor.PreprocessingOverview finalPreprocessingOverview
         }
 
         /// <summary>
-        /// Displays information about the readout unit regression progress.
+        /// Displays information about the build process progress.
         /// </summary>
-        /// <param name="buildingState">Current state of the regression process</param>
-        /// <param name="foundBetter">Indicates that the best readout unit was changed as a result of the performed epoch</param>
-        protected void OnRegressionEpochDone(TrainedNetworkBuilder.BuildingState buildingState, bool foundBetter)
+        /// <param name="buildProgress">The current state of the build process.</param>
+        /// <param name="foundBetter">Indicates that the best network so far was found during the last performed epoch.</param>
+        protected void OnEpochDone(TNRNetBuilder.BuildProgress buildProgress, bool foundBetter)
         {
             int reportEpochsInterval = 5;
             //Progress info
             if (foundBetter ||
-                (buildingState.Epoch % reportEpochsInterval) == 0 ||
-                buildingState.Epoch == buildingState.MaxEpochs ||
-                (buildingState.Epoch == 1 && buildingState.RegrAttemptNumber == 1)
+                (buildProgress.Epoch % reportEpochsInterval) == 0 ||
+                buildProgress.Epoch == buildProgress.MaxEpochs ||
+                (buildProgress.Epoch == 1 && buildProgress.AttemptNumber == 1)
                 )
             {
                 //Build progress report message
-                string progressText = buildingState.GetProgressInfo(4);
+                string progressText = buildProgress.GetInfo(4);
                 //Report the progress
-                _log.Write(progressText, !(buildingState.Epoch == 1 && buildingState.RegrAttemptNumber == 1));
+                _log.Write(progressText, !(buildProgress.Epoch == 1 && buildProgress.AttemptNumber == 1));
             }
             return;
         }
 
         /// <summary>
-        /// Loads given file and executes StateMachine training.
-        /// This version of function requires configured NeuralPreprocessor.
+        /// Loads the specified file and executes the StateMachine training.
         /// </summary>
-        /// <param name="stateMachine">Instance of StateMachine to be trained</param>
-        /// <param name="trainingDataFileName">Name of the csv file containing training data</param>
-        /// <param name="predictionInputVector">Returned vector to be used for next prediction (relevant only in case of continuous feeding of the input)</param>
+        /// <param name="stateMachine">An instance of StateMachine to be trained.</param>
+        /// <param name="trainingDataFileName">The name of the csv file containing the training data.</param>
+        /// <param name="predictionInputVector">The vector to be used for next prediction (relevant only in case of continuous feeding of the input).</param>
         protected void TrainStateMachine(StateMachine stateMachine, string trainingDataFileName, out double[] predictionInputVector)
         {
-            //Register to RegressionEpochDone event
-            stateMachine.RL.RegressionEpochDone += OnRegressionEpochDone;
+            //Register to EpochDone event
+            stateMachine.RL.EpochDone += OnEpochDone;
             //Load csv data
             CsvDataHolder trainingCsvData = new CsvDataHolder(trainingDataFileName);
             //Convert csv data to VectorBundle useable for StateMachine training
@@ -153,13 +152,12 @@ out predictionInputVector
         }
 
         /// <summary>
-        /// Loads given file and executes StateMachine verification.
-        /// This version of function requires configured NeuralPreprocessor.
+        /// Loads the specified file and executes the StateMachine verification.
         /// </summary>
-        /// <param name="stateMachine">Instance of StateMachine to be trained</param>
-        /// <param name="verificationDataFileName">Name of the csv file containing verification data</param>
-        /// <param name="omittedInputVector">Remaining input vector from training phase (relevant only in case of continuous feeding of the input)</param>
-        /// <param name="predictionInputVector">Returned vector to be used for next prediction (relevant only in case of continuous feeding of the input)</param>
+        /// <param name="stateMachine">An instance of StateMachine to be verified.</param>
+        /// <param name="verificationDataFileName">The name of the csv file containing the verification data.</param>
+        /// <param name="omittedInputVector">Remaining input vector from training phase (relevant only in case of continuous feeding of the input).</param>
+        /// <param name="predictionInputVector">The vector to be used for next prediction (relevant only in case of continuous feeding of the input).</param>
         protected void VerifyStateMachine(StateMachine stateMachine, string verificationDataFileName, double[] omittedInputVector, out double[] predictionInputVector)
         {
             //Load csv data
@@ -175,7 +173,7 @@ protected void VerifyStateMachine(StateMachine stateMachine, string verification
                     //Continuous input feeding
                     //Last known input values from training (predictionInputVector) must be pushed into the reservoirs to keep time series continuity
                     //(first input data in verification.csv is output of the last data in training.csv)
-                    double[] tmp = stateMachine.Compute(omittedInputVector);
+                    double[] tmp = stateMachine.Compute(omittedInputVector, out ReadoutLayer.ReadoutData readoutData);
                     //Load verification data and get new predictionInputVector for final prediction
                     verificationData = VectorBundle.Load(verificationCsvData,
                                                          stateMachine.Config.NeuralPreprocessorCfg.InputEncoderCfg.VaryingFieldsCfg.ExternalFieldsCfg.GetFieldNames(),

Demo/DemoConsoleApp/Examples/FFNetBoolAlg.cs

@@ -1,20 +1,22 @@
-using System;
-using System.Globalization;
-using RCNet.Neural.Activation;
+using RCNet.Neural.Activation;
 using RCNet.Neural.Data;
 using RCNet.Neural.Network.NonRecurrent.FF;
+using System;
+using System.Globalization;
 
 namespace Demo.DemoConsoleApp.Examples
 {
     /// <summary>
-    /// This "Hello world" example shows how to use implemented FF network as the independent component.
-    /// It trains the multilayer Feed Forward network to solve AND, OR and XOR.
+    /// Trains a multilayer Feed Forward network to solve AND, OR and XOR.
     /// </summary>
+    /// <remarks>
+    /// This "Hello world" example shows how to use the feed forward network component independently.
+    /// </remarks>
     public class FFNetBoolAlg : ExampleBase
     {
         /// <summary>
-        /// Creates training data.
-        /// Input vector contains 0/1 combination and output vector contains appropriate results of the AND, OR and XOR operation
+        /// Creates the training data.
+        /// Input vector contains 0/1 combination and output vector contains appropriate results of the AND, OR and XOR operation.
         /// </summary>
         private VectorBundle CreateTrainingData()
         {
@@ -48,7 +50,7 @@ public void Run()
             //Training
             _log.Write("Training");
             _log.Write("--------");
-            //Create trainer instance
+            //Create the trainer instance
             RPropTrainer trainer = new RPropTrainer(ffNet,
                                                     trainingData.InputVectorCollection,
                                                     trainingData.OutputVectorCollection,
@@ -63,7 +65,7 @@ public void Run()
             _log.Write(string.Empty);
 
             //Training is done
-            //Display network computation results
+            //Display the network computation results
             _log.Write("Trained network computations:");
             _log.Write("-----------------------------");
             foreach (double[] input in trainingData.InputVectorCollection)

Demo/DemoConsoleApp/Examples/LibrasClassificationESNDesigner.cs

@@ -1,17 +1,17 @@
-using System;
-using RCNet.Neural.Activation;
+using RCNet.Neural.Activation;
 using RCNet.Neural.Data.Filter;
 using RCNet.Neural.Network.NonRecurrent;
 using RCNet.Neural.Network.SM;
 using RCNet.Neural.Network.SM.Preprocessing;
 using RCNet.Neural.Network.SM.Preprocessing.Input;
 using RCNet.Neural.Network.SM.Preprocessing.Neuron.Predictor;
 using RCNet.Neural.Network.SM.Readout;
+using System;
 
 namespace Demo.DemoConsoleApp.Examples
 {
     /// <summary>
-    /// Example code shows how to setup StateMachine as a pure ESN for multivariate timeseries classification using StateMachineDesigner.
+    /// Example code shows how to use StateMachineDesigner and setup StateMachine as a pure ESN for multivariate timeseries classification.
     /// Example uses LibrasMovement_train.csv and LibrasMovement_verify.csv from ./Data subfolder.
     /// The dataset is from "Anthony Bagnall, Jason Lines, William Vickers and Eamonn Keogh, The UEA & UCR Time Series Classification Repository, www.timeseriesclassification.com"
     /// https://timeseriesclassification.com/description.php?Dataset=Libras
@@ -44,9 +44,9 @@ public void Run()
                                                                                 );
             //Simplified readout layer configuration
             ReadoutLayerSettings readoutCfg = StateMachineDesigner.CreateClassificationReadoutCfg(new CrossvalidationSettings(0.0825d, CrossvalidationSettings.AutoFolds, 1),
-                                                                                                  StateMachineDesigner.CreateSingleLayerRegrNet(new AFAnalogIdentitySettings(), 5, 400),
+                                                                                                  StateMachineDesigner.CreateSingleLayerFFNetCfg(new AFAnalogIdentitySettings(), 5, 400),
+                                                                                                  1,
                                                                                                   "Hand movement",
-                                                                                                  new NetworkClusterSecondLevelCompSettings(new CrossvalidationSettings(0.25d, CrossvalidationSettings.AutoFolds, 2), StateMachineDesigner.CreateMultiLayerRegrNet(10, new AFAnalogLeakyReLUSettings(), 1, 5, 400)),
                                                                                                   "curved swing",
                                                                                                   "horizontal swing",
                                                                                                   "vertical swing",