diff --git a/src/mlpack/methods/ann/layer/CMakeLists.txt b/src/mlpack/methods/ann/layer/CMakeLists.txt
index e69e65e23d..3ed593a931 100644
--- a/src/mlpack/methods/ann/layer/CMakeLists.txt
+++ b/src/mlpack/methods/ann/layer/CMakeLists.txt
@@ -6,6 +6,8 @@ set(SOURCES
   alpha_dropout.hpp
   alpha_dropout_impl.hpp
   base_layer.hpp
+  concat.hpp
+  concat_impl.hpp
   concatenate.hpp
   concatenate_impl.hpp
   convolution.hpp
diff --git a/src/mlpack/methods/ann/layer/not_adapted/concat.hpp b/src/mlpack/methods/ann/layer/concat.hpp
similarity index 74%
rename from src/mlpack/methods/ann/layer/not_adapted/concat.hpp
rename to src/mlpack/methods/ann/layer/concat.hpp
index 1d99ff60c9..7d8aac0f3d 100644
--- a/src/mlpack/methods/ann/layer/not_adapted/concat.hpp
+++ b/src/mlpack/methods/ann/layer/concat.hpp
@@ -25,42 +25,47 @@ namespace ann /** Artificial Neural Network. */ {
  * feed-forward fully connected network container which plugs various layers
  * together.
  *
- * @tparam InputType Type of the input data (arma::colvec, arma::mat,
- *         arma::sp_mat or arma::cube).
- * @tparam OutputType Type of the output data (arma::colvec, arma::mat,
- *         arma::sp_mat or arma::cube).
+ * NOTE: this class is not intended to exist for long!  It will be replaced with
+ * a more flexible DAG network type.
+ *
+ * @tparam MatType Matrix representation to accept as input and use for
+ *    computation.
  */
-template <
-    typename InputType = arma::mat,
-    typename OutputType = arma::mat
->
-class ConcatType : public MultiLayer<InputType, OutputType>
+template <typename MatType = arma::mat>
+class ConcatType : public MultiLayer<MatType>
 {
  public:
   /**
-   * Create the Concat object using the specified parameters.
-   *
-   * @param run Call the Forward/Backward method before the output is merged.
+   * Create the Concat object.  The axis used for concatenation will be the last
+   * one.
    */
-  ConcatType(const bool run = true);
+  ConcatType();
 
   /**
    * Create the Concat object, specifying a particular axis on which the layer
    * outputs should be concatenated.
    *
    * @param axis Concat axis.
-   * @param run Call the Forward/Backward method before the output is merged.
    */
-  ConcatType(const size_t axis, const bool run = true);
+  ConcatType(const size_t axis);
 
   /**
    * Destroy the layers held by the model.
    */
-  ~ConcatType();
+  virtual ~ConcatType();
 
   //! Clone the ConcatType object. This handles polymorphism correctly.
   ConcatType* Clone() const { return new ConcatType(*this); }
 
+  //! Copy the given ConcatType layer.
+  ConcatType(const ConcatType& other);
+  //! Take ownership of the given ConcatType layer.
+  ConcatType(ConcatType&& other);
+  //! Copy the given ConcatType layer.
+  ConcatType& operator=(const ConcatType& other);
+  //! Take ownership of the given ConcatType layer.
+  ConcatType& operator=(ConcatType&& other);
+
   /**
    * Ordinary feed forward pass of a neural network, evaluating the function
    * f(x) by propagating the activity forward through f.
@@ -68,7 +73,7 @@ class ConcatType : public MultiLayer<InputType, OutputType>
    * @param input Input data used for evaluating the specified function.
    * @param output Resulting output activation.
    */
-  void Forward(const InputType& input, OutputType& output);
+  void Forward(const MatType& input, MatType& output);
 
   /**
    * Ordinary feed backward pass of a neural network, using 3rd-order tensors as
@@ -79,9 +84,9 @@ class ConcatType : public MultiLayer<InputType, OutputType>
    * @param gy The backpropagated error.
    * @param g The calculated gradient.
    */
-  void Backward(const InputType& /* input */,
-                const OutputType& gy,
-                OutputType& g);
+  void Backward(const MatType& /* input */,
+                const MatType& gy,
+                MatType& g);
 
   /**
    * This is the overload of Backward() that runs only a specific layer with
@@ -92,9 +97,9 @@ class ConcatType : public MultiLayer<InputType, OutputType>
    * @param g The calculated gradient.
    * @param index The index of the layer to run.
    */
-  void Backward(const InputType& /* input */,
-                const OutputType& gy,
-                OutputType& g,
+  void Backward(const MatType& /* input */,
+                const MatType& gy,
+                MatType& g,
                 const size_t index);
 
   /**
@@ -104,9 +109,9 @@ class ConcatType : public MultiLayer<InputType, OutputType>
    * @param error The calculated error.
    * @param gradient The calculated gradient.
    */
-  void Gradient(const InputType& /* input */,
-                const OutputType& error,
-                OutputType& /* gradient */);
+  void Gradient(const MatType& /* input */,
+                const MatType& error,
+                MatType& /* gradient */);
 
   /**
    * This is the overload of Gradient() that runs a specific layer with the
@@ -117,29 +122,24 @@ class ConcatType : public MultiLayer<InputType, OutputType>
    * @param gradient The calculated gradient.
    * @param The index of the layer to run.
    */
-  void Gradient(const InputType& input,
-                const OutputType& error,
-                OutputType& gradient,
+  void Gradient(const MatType& input,
+                const MatType& error,
+                MatType& gradient,
                 const size_t index);
 
-  //! Get the value of run parameter.
-  bool Run() const { return run; }
-  //! Modify the value of run parameter.
-  bool& Run() { return run; }
-
   //! Get the axis of concatenation.
-  const size_t& ConcatAxis() const { return axis; }
+  const size_t& Axis() const { return axis; }
 
-  //! Get the size of the weight matrix.
-  size_t WeightSize() const { return 0; }
+  // We don't need to overload WeightSize(); MultiLayer already computes this
+  // correctly.  (It is the sum of weights of all child layers.)
 
   void ComputeOutputDimensions()
   {
     // The input is sent to every layer.
-    for (size_t i = 0; i < network.size(); ++i)
+    for (size_t i = 0; i < this->network.size(); ++i)
     {
-      network[i]->InputDimensions() = this->inputDimensions;
-      network[i]->ComputeOutputDimensions();
+      this->network[i]->InputDimensions() = this->inputDimensions;
+      this->network[i]->ComputeOutputDimensions();
     }
 
     // If the user did not specify an axis, we will use the last one.
@@ -147,7 +147,7 @@ class ConcatType : public MultiLayer<InputType, OutputType>
     // concatenating along is valid.
     if (!useAxis)
     {
-      axis  = this->inputDimensions.size() - 1;
+      axis = this->inputDimensions.size() - 1;
     }
     else if (axis >= this->inputDimensions.size())
     {
@@ -197,10 +197,10 @@ class ConcatType : public MultiLayer<InputType, OutputType>
   }
 
   /**
-   * Serialize the layer
+   * Serialize the layer.
    */
   template<typename Archive>
-  void serialize(Archive& ar,  const uint32_t /* version */);
+  void serialize(Archive& ar, const uint32_t /* version */);
 
  private:
   //! Parameter which indicates the axis of concatenation.
@@ -211,7 +211,7 @@ class ConcatType : public MultiLayer<InputType, OutputType>
 }; // class ConcatType.
 
 // Standard Concat layer.
-typedef ConcatType<arma::mat, arma::mat> Concat;
+typedef ConcatType<arma::mat> Concat;
 
 } // namespace ann
 } // namespace mlpack
diff --git a/src/mlpack/methods/ann/layer/not_adapted/concat_impl.hpp b/src/mlpack/methods/ann/layer/concat_impl.hpp
similarity index 53%
rename from src/mlpack/methods/ann/layer/not_adapted/concat_impl.hpp
rename to src/mlpack/methods/ann/layer/concat_impl.hpp
index 7e231b5f81..29c345492b 100644
--- a/src/mlpack/methods/ann/layer/not_adapted/concat_impl.hpp
+++ b/src/mlpack/methods/ann/layer/concat_impl.hpp
@@ -19,38 +19,81 @@
 namespace mlpack {
 namespace ann /** Artificial Neural Network. */ {
 
-template<typename InputType, typename OutputType>
-ConcatType<InputType, OutputType>::ConcatType(
-    const bool run) :
-    axis(0),
-    useAxis(false)
-{
-  // Nothing to do.
-}
-
-template<typename InputType, typename OutputType>
-ConcatType<InputType, OutputType>::ConcatType(
-    const size_t axis,
-    const bool run) :
+template<typename MatType>
+ConcatType<MatType>::ConcatType(
+    const size_t axis) :
+    MultiLayer<MatType>(),
     axis(axis),
     useAxis(true)
 {
   // Nothing to do.
 }
 
-template<typename InputType, typename OutputType>
-ConcatType<InputType, OutputType>::~ConcatType()
+template<typename MatType>
+ConcatType<MatType>::ConcatType() :
+    MultiLayer<MatType>(),
+    axis(0),
+    useAxis(false)
 {
-  // Clear memory.
-  for (size_t i = 0; i < this->network.size(); ++i)
-    delete this->network[i];
+  // Nothing to do.
 }
 
-template<typename InputType, typename OutputType>
-void ConcatType<InputType, OutputType>::Forward(
-    const InputType& input, OutputType& output)
+template<typename MatType>
+ConcatType<MatType>::~ConcatType()
 {
-  this->InitializeForwardPassMemory();
+  // Nothing to do: the child layer memory is already cleared by MultiLayer.
+}
+
+template<typename MatType>
+ConcatType<MatType>::ConcatType(const ConcatType& other) :
+    MultiLayer<MatType>(other),
+    axis(other.axis),
+    useAxis(other.useAxis)
+{
+  // Nothing else to do.
+}
+
+template<typename MatType>
+ConcatType<MatType>::ConcatType(ConcatType&& other) :
+    MultiLayer<MatType>(std::move(other)),
+    axis(std::move(other.axis)),
+    useAxis(std::move(other.useAxis))
+{
+  // Nothing else to do.
+}
+
+template<typename MatType>
+ConcatType<MatType>& ConcatType<MatType>::operator=(const ConcatType& other)
+{
+  if (this != &other)
+  {
+    MultiLayer<MatType>::operator=(other);
+    axis = other.axis;
+    useAxis = other.useAxis;
+  }
+
+  return *this;
+}
+
+template<typename MatType>
+ConcatType<MatType>& ConcatType<MatType>::operator=(ConcatType&& other)
+{
+  if (this != &other)
+  {
+    MultiLayer<MatType>::operator=(std::move(other));
+    axis = std::move(other.axis);
+    useAxis = std::move(other.useAxis);
+  }
+
+  return *this;
+}
+
+template<typename MatType>
+void ConcatType<MatType>::Forward(const MatType& input, MatType& output)
+{
+  // The implementation of MultiLayer is fine: this will allocate a matrix that
+  // is able to hold each child layer's output.
+  this->InitializeForwardPassMemory(input.n_cols);
 
   // Pass the input through all the layers in the network.
   for (size_t i = 0; i < this->network.size(); ++i)
@@ -76,32 +119,42 @@ void ConcatType<InputType, OutputType>::Forward(
       std::accumulate(this->outputDimensions.begin() + axis + 1,
           this->outputDimensions.end(), 0);
 
-  std::vector<arma::Cube<typename OutputType::elem_type>> layerOutputAliases;
+  std::vector<arma::Cube<typename MatType::elem_type>> layerOutputAliases(
+      this->layerOutputs.size());
   for (size_t i = 0; i < this->layerOutputs.size(); ++i)
   {
-    layerOutputAliases.emplace_back(arma::Cube<typename OutputType::elem_type>(
-        this->layerOutputs[i].memptr(), rows,
-        this->network[i]->OutputDimensions()[axis], slices, false, true);
+    MakeAlias(layerOutputAliases.back(),
+              (typename MatType::elem_type*) this->layerOutputs[i].memptr(),
+              rows,
+              this->network[i]->OutputDimensions()[axis],
+              slices);
   }
 
-  arma::Cube<typename OutputType::elem_type> output(output.memptr(), rows,
-      this->outputDimensions[axis], slices, false, true);
+  arma::Cube<typename MatType::elem_type> outputAlias;
+  MakeAlias(outputAlias,
+            (typename MatType::elem_type*) output.memptr(),
+            rows,
+            this->outputDimensions[axis],
+            slices);
 
   // Now get the columns from each output.
   size_t startCol = 0;
   for (size_t i = 0; i < layerOutputAliases.size(); ++i)
   {
     const size_t cols = layerOutputAliases[i].n_cols;
-    output.cols(startCol, startCol + cols - 1) = layerOutputAliases[i];
+    outputAlias.cols(startCol, startCol + cols - 1) = layerOutputAliases[i];
     startCol += cols;
   }
 }
 
-template<typename InputType, typename OutputType>
-void ConcatType<InputType, OutputType>::Backward(
-    const InputType& /* input */, const OutputType& gy, OutputType& g)
+template<typename MatType>
+void ConcatType<MatType>::Backward(
+    const MatType& /* input */, const MatType& gy, MatType& g)
 {
-  this->InitializeBackwardPassMemory();
+  // The implementation of MultiLayer is fine: this will allocate a matrix that
+  // is able to hold each child layer's delta (which has the same size as the
+  // input).
+  this->InitializeBackwardPassMemory(gy.n_cols);
 
   // Just like the forward pass, we can treat our inputs as a cube, but here we
   // have to distribute the correct parts of `gy` to the layers.
@@ -113,18 +166,20 @@ void ConcatType<InputType, OutputType>::Backward(
       std::accumulate(this->outputDimensions.begin() + axis + 1,
           this->outputDimensions.end(), 0);
 
-  arma::Cube<typename OutputType::elem_type> gyTmp(gy.memptr(), rows,
-      this->outputDimensions[axis], slices, false, true);
+  arma::Cube<typename MatType::elem_type> gyTmp;
+  MakeAlias(gyTmp,
+            (typename MatType::elem_type*) gy.memptr(),
+            rows,
+            this->outputDimensions[axis],
+            slices);
 
   size_t startCol = 0;
   for (size_t i = 0; i < this->network.size(); ++i)
   {
     const size_t cols = this->network[i]->OutputDimensions()[axis];
-    // TODO: is delta size correct?
-    // TODO: no copy!
-    OutputType delta = gyTmp.cols(startCol, startCol + cols - 1);
-    // TODO: consider batch size correctly
-    delta.reshape( ... );
+    MatType delta = gyTmp.cols(startCol, startCol + cols - 1);
+    // Reshape so that the batch size is the number of columns.
+    delta.reshape(delta.n_elem / gy.n_cols, gy.n_cols);
     this->network[i]->Backward(this->layerOutputs[i], delta,
         this->layerDeltas[i]);
 
@@ -138,11 +193,11 @@ void ConcatType<InputType, OutputType>::Backward(
   }
 }
 
-template<typename InputType, typename OutputType>
-void ConcatType<InputType, OutputType>::Backward(
-    const InputType& /* input */,
-    const OutputType& gy,
-    OutputType& g,
+template<typename MatType>
+void ConcatType<MatType>::Backward(
+    const MatType& /* input */,
+    const MatType& gy,
+    MatType& g,
     const size_t index)
 {
   // We only intend to perform a backward pass on one layer.
@@ -156,8 +211,12 @@ void ConcatType<InputType, OutputType>::Backward(
       std::accumulate(this->outputDimensions.begin() + axis + 1,
           this->outputDimensions.end(), 0);
 
-  arma::Cube<typename OutputType::elem_type> gyTmp(gy.memptr(), rows,
-      this->outputDimensions[axis], slices, false, true);
+  arma::Cube<typename MatType::elem_type> gyTmp;
+  MakeAlias(gyTmp,
+            (typename MatType::elem_type*) gy.memptr(),
+            rows,
+            this->outputDimensions[axis],
+            slices);
 
   size_t startCol = 0;
   for (size_t i = 0; i < index; ++i)
@@ -165,22 +224,22 @@ void ConcatType<InputType, OutputType>::Backward(
     startCol += this->network[i]->OutputDimensions()[axis];
   }
 
-  // TODO: no copy!
   const size_t cols = this->network[index]->OutputDimensions()[axis];
-  OutputType delta = gyTmp.cols(startCol, startCol + cols - 1);
-  delta.reshape( ... );
+  MatType delta = gyTmp.cols(startCol, startCol + cols - 1);
+  // Reshape so that the batch size is the number of columns.
+  delta.reshape(delta.n_elem / gy.n_cols, gy.n_cols);
 
   this->network[index]->Backward(this->layerOutputs[index], delta, g);
 }
 
-template<typename InputType, typename OutputType>
-void ConcatType<InputType, OutputType>::Gradient(
-    const InputType& input,
-    const OutputType& error,
-    OutputType& gradient)
+template<typename MatType>
+void ConcatType<MatType>::Gradient(
+    const MatType& input,
+    const MatType& error,
+    MatType& gradient)
 {
   // Just like the forward pass, we can treat our inputs as a cube, but here we
-  // have to distribute the correct parts of `gy` to the layers.
+  // have to distribute the correct parts of `error` to the layers.
 
   size_t slices = (axis == 0) ? input.n_cols :
       std::accumulate(this->outputDimensions.begin(),
@@ -189,8 +248,12 @@ void ConcatType<InputType, OutputType>::Gradient(
       std::accumulate(this->outputDimensions.begin() + axis + 1,
           this->outputDimensions.end(), 0);
 
-  arma::Cube<typename OutputType::elem_type> errorTmp(error.memptr(), rows,
-      this->outputDimensions[axis], slices, false, true);
+  arma::Cube<typename MatType::elem_type> errorTmp;
+  MakeAlias(errorTmp,
+            (typename MatType::elem_type*) error.memptr(),
+            rows,
+            this->outputDimensions[axis],
+            slices);
 
   size_t startCol = 0;
   size_t startParam = 0;
@@ -199,11 +262,13 @@ void ConcatType<InputType, OutputType>::Gradient(
     const size_t cols = this->network[i]->OutputDimensions()[axis];
     const size_t params = this->network[i]->WeightSize();
 
-    OutputType err = errorTmp.cols(startCol, startCol + cols - 1);
-    err.reshape(input.n_cols, err.n_elem / input.n_cols);
-    // TODO: what about layerGradients?
-    OutputType gradientAlias(gradient.colptr(startParam, 1, params, false,
-        true);
+    MatType err = errorTmp.cols(startCol, startCol + cols - 1);
+    err.reshape(err.n_elem / input.n_cols, input.n_cols);
+    MatType gradientAlias;
+    MakeAlias(gradientAlias,
+              (typename MatType::elem_type*) gradient.colptr(startParam),
+              1,
+              params);
     this->network[i]->Gradient(input, err, gradientAlias);
 
     startCol += cols;
@@ -211,16 +276,15 @@ void ConcatType<InputType, OutputType>::Gradient(
   }
 }
 
-// TODO: adapt
-template<typename InputType, typename OutputType>
-void ConcatType<InputType, OutputType>::Gradient(
-    const InputType& input,
-    const OutputType& error,
-    OutputType& gradient,
+template<typename MatType>
+void ConcatType<MatType>::Gradient(
+    const MatType& input,
+    const MatType& error,
+    MatType& gradient,
     const size_t index)
 {
   // Just like the forward pass, we can treat our inputs as a cube, but here we
-  // have to distribute the correct parts of `gy` to the layers.
+  // have to distribute the correct parts of `error` to the layers.
 
   size_t slices = (axis == 0) ? input.n_cols :
       std::accumulate(this->outputDimensions.begin(),
@@ -229,8 +293,12 @@ void ConcatType<InputType, OutputType>::Gradient(
       std::accumulate(this->outputDimensions.begin() + axis + 1,
           this->outputDimensions.end(), 0);
 
-  arma::Cube<typename OutputType::elem_type> errorTmp(error.memptr(), rows,
-      this->outputDimensions[axis], slices, false, true);
+  arma::Cube<typename MatType::elem_type> errorTmp;
+  MakeAlias(errorTmp,
+            (typename MatType::elem_type*) error.memptr(),
+            rows,
+            this->outputDimensions[axis],
+            slices);
 
   size_t startCol = 0;
   size_t startParam = 0;
@@ -243,19 +311,22 @@ void ConcatType<InputType, OutputType>::Gradient(
   const size_t cols = this->network[index]->OutputDimensions()[axis];
   const size_t params = this->network[index]->WeightSize();
 
-  // TODO: no copy!
-  OutputType err = errorTmp.cols(startCol, startCol + cols - 1);
-  err.reshape(input.n_cols, err.n_elem / input.n_cols);
-  OutputType gradientAlias(gradient.memptr(), 1, params, false, true);
+  MatType err = errorTmp.cols(startCol, startCol + cols - 1);
+  err.reshape(err.n_elem / input.n_cols, input.n_cols);
+  MatType gradientAlias;
+  MakeAlias(gradientAlias,
+            (typename MatType::elem_type*) gradient.colptr(startParam),
+            1,
+            params);
   this->network[index]->Gradient(input, err, gradientAlias);
 }
 
-template<typename InputType, typename OutputType>
+template<typename MatType>
 template<typename Archive>
-void ConcatType<InputType, OutputType>::serialize(
+void ConcatType<MatType>::serialize(
     Archive& ar, const uint32_t /* version */)
 {
-  ar(cereal::base_class<MultiLayer<InputType, OutputType>>(this));
+  ar(cereal::base_class<MultiLayer<MatType>>(this));
 
   ar(CEREAL_NVP(axis));
   ar(CEREAL_NVP(useAxis));
diff --git a/src/mlpack/methods/ann/layer/layer_types.hpp b/src/mlpack/methods/ann/layer/layer_types.hpp
index fb179b6457..ec1848f486 100644
--- a/src/mlpack/methods/ann/layer/layer_types.hpp
+++ b/src/mlpack/methods/ann/layer/layer_types.hpp
@@ -20,6 +20,7 @@
 #include <mlpack/methods/ann/layer/add.hpp>
 #include <mlpack/methods/ann/layer/alpha_dropout.hpp>
 #include <mlpack/methods/ann/layer/base_layer.hpp>
+#include <mlpack/methods/ann/layer/concat.hpp>
 #include <mlpack/methods/ann/layer/concatenate.hpp>
 #include <mlpack/methods/ann/layer/convolution.hpp>
 #include <mlpack/methods/ann/layer/dropconnect.hpp>
diff --git a/src/mlpack/methods/ann/layer/serialization.hpp b/src/mlpack/methods/ann/layer/serialization.hpp
index 5e13d2a9b4..5b0306cfa3 100644
--- a/src/mlpack/methods/ann/layer/serialization.hpp
+++ b/src/mlpack/methods/ann/layer/serialization.hpp
@@ -29,6 +29,7 @@
     CEREAL_REGISTER_TYPE(mlpack::ann::ElishType<__VA_ARGS__>); \
     CEREAL_REGISTER_TYPE(mlpack::ann::GaussianType<__VA_ARGS__>); \
     /* (end of base_layer.hpp) */ \
+    CEREAL_REGISTER_TYPE(mlpack::ann::ConcatType<__VA_ARGS__>); \
     CEREAL_REGISTER_TYPE(mlpack::ann::ConcatenateType<__VA_ARGS__>); \
     CEREAL_REGISTER_TYPE(mlpack::ann::ConvolutionType< \
         mlpack::ann::NaiveConvolution<mlpack::ann::ValidConvolution>, \