Layer hierarchy potentially overly flat

[realmarcin]

Consider the context of an AI ontology, which of these DNN Layers could be grouped into categories?

GPT4o

Normalization Layers Batch Normalization Layers:

BatchNorm1D Layer BatchNorm2D Layer BatchNorm3D Layer SyncBatchNorm Layer LazyBatchNorm1D Layer LazyBatchNorm2D Layer LazyBatchNorm3D Layer BatchNormalization Layer Instance Normalization Layers:

InstanceNorm1D Layer InstanceNorm2D Layer InstanceNorm3D Layer LazyInstanceNorm1D Layer LazyInstanceNorm2D Layer LazyInstanceNorm3D Layer Other Normalization Layers:

GroupNorm Layer LayerNorm Layer LayerNormalization Layer LocalResponseNorm Layer UnitNormalization Layer Normalization Layer Convolutional Layers Standard Convolutional Layers:

Convolution1D Layer Convolution2D Layer Convolution3D Layer Transposed Convolutional Layers:

Convolution1DTranspose Layer Convolution2DTranspose Layer Convolution3DTranspose Layer Depthwise Convolutional Layers:

DepthwiseConv1D Layer DepthwiseConv2D Layer Separable Convolutional Layers:

SeparableConvolution1D Layer SeparableConvolution2D Layer Convolutional LSTM Layers:

ConvLSTM1D Layer ConvLSTM2D Layer ConvLSTM3D Layer Other Convolutional Layers:

LocallyConnected1D Layer LocallyConnected2D Layer Locally-connected Layer Pooling Layers Average Pooling Layers:

AvgPool1D Layer AvgPool2D Layer AvgPool3D Layer AveragePooling1D Layer AveragePooling2D Layer AveragePooling3D Layer GlobalAveragePooling1D Layer GlobalAveragePooling2D Layer GlobalAveragePooling3D Layer Max Pooling Layers:

MaxPooling1D Layer MaxPooling2D Layer MaxPooling3D Layer GlobalMaxPooling1D Layer GlobalMaxPooling2D Layer GlobalMaxPooling3D Layer Adaptive Pooling Layers:

AdaptiveAvgPool1D Layer AdaptiveMaxPool1D Layer AdaptiveAvgPool2D Layer AdaptiveMaxPool2D Layer AdaptiveAvgPool3D Layer AdaptiveMaxPool3D Layer Other Pooling Layers:

FractionalMaxPool2D Layer FractionalMaxPool3D Layer MaxUnpool1D Layer MaxUnpool2D Layer MaxUnpool3D Layer LPPool1D Layer LPPool2D Layer RNN Layers Standard RNN Layers:

SimpleRNN Layer GRU Layer LSTM Layer RNN Layer RNN Cell Layers:

GRUCell Layer SimpleRNNCell Layer LSTMCell Layer AbstractRNNCell StackedRNNCells Layer Bidirectional and Time-Distributed Layers:

Bidirectional Layer TimeDistributed Layer Attention Layers Attention Mechanisms: Attention Layer MultiHeadAttention Layer AdditiveAttention Layer Preprocessing Layers Categorical Features Preprocessing:

CategoryEncoding Layer IntegerLookup Layer StringLookup Layer Hashing Layer Numerical Features Preprocessing:

Discretization Layer Text Preprocessing:

TextPreprocessing Layer TextVectorization Layer Image Preprocessing and Augmentation:

ImagePreprocessing Layer ImageAugmentation Layer CenterCrop Layer RandomBrightness Layer RandomContrast Layer RandomCrop Layer RandomFlip Layer RandomRotation Layer RandomTranslation Layer RandomHeight Layer RandomWidth Layer RandomZoom Layer Rescaling Layer Resizing Layer Dense and Hidden Layers Dense Layers:

Dense Layer DenseFeatures Layer Hidden Layers:

Hidden Layer Probabilistic Hidden Layer Spiking Hidden Layer Activation Layers Activation Functions: Activation Layer PReLU Layer ELU Layer LeakyReLU Layer ReLU Layer Softmax Layer ThresholdedReLU Layer Regularization Layers Dropout and Noise Layers: Dropout Layer GaussianDropout Layer SpatialDropout1D Layer SpatialDropout2D Layer SpatialDropout3D Layer GaussianNoise Layer AlphaDropout Layer NoiseDense Layer ActivityRegularization Layer Input, Output, and Other Layers Input Layers:

Input Layer NoisyInput Layer BackfedInput Layer MatchedInputOutput Layer Output Layers:

Output Layer Wrapper Layers:

Wrapper Layer Merging Layers:

Add Layer Average Layer Maximum Layer Minimum Layer Concatenate Layer Multiply Layer Subtract Layer Dot Layer Reshaping and Permutation Layers:

Reshape Layer Permute Layer Flatten Layer RepeatVector Layer Cropping and Padding Layers:

Cropping1D Layer Cropping2D Layer Cropping3D Layer ZeroPadding1D Layer ZeroPadding2D Layer ZeroPadding3D Layer Upsampling Layers:

UpSampling1D Layer UpSampling2D Layer UpSampling3D Layer Other Layers:

Lambda Layer Policy Layer MemoryCell Layer Weighted Layer Layer Layer InputSpec Layer Masking Layer Kernel Layer

[cmungall]

Looks like this is done?