README
¶
Neurlang Binary Classifier (Hashtron)
Neurlang Classifier is a machine learning algorithm designed for training binary classifiers of integers. It is particularly efficient for tasks where the input data consists of uint32 integers. If your samples are larger than uint32, pre-hashing (e.g., using murmur hash) is recommended.
Key Features
- Integer-based Classification: The algorithm works directly with
uint32integers, avoiding the use of floating-point arithmetic. - Balanced Datasets: Performs best when the true and false sets are balanced. If they are not, you can use
Parity()to balance the data. - Hash-based Models: The resulting models are hash-based, making them extremely fast for inference.
- Scalable Networks: Supports larger networks, including those with majority pooling layers, reducing the need for large-scale matrix multiplication.
Implementing a Dataset
To implement a dataset for training the network, define a slice of samples where each sample has the following methods:
Feature(int) uint32: Returns the feature at the specified index. Each front layer hashtron reads its index starting from 0.Parity() uint16: Returns the parity of the sample. This is used to balance the dataset. If your classes contain equal number of samples, you can return 0.Output() uint16: Returns the output label. This is the prediction for the sample.
Implementing a Network
Here is an example of how to implement a network with majority pooling layers:
const fanout1 = 3
const fanout2 = 5
const fanout3 = 3
const fanout4 = 5
var net feedforward.FeedforwardNetwork
net.NewLayerP(fanout1*fanout2*fanout3*fanout4, 0, 1<<fanout4)
net.NewCombiner(majpool2d.MustNew(fanout1*fanout2*fanout4, 1, fanout3, 1, fanout4, 1, 1))
net.NewLayerP(fanout1*fanout2, 0, 1<<fanout2)
net.NewCombiner(majpool2d.MustNew(fanout2, 1, fanout1, 1, fanout2, 1, 1))
net.NewLayer(1, 0)
fanout1andfanout3define the majority pooling dimensions.fanout2andfanout4define the number of hashtrons.- The final layer contains one hashtron for predictions.
- The
0in the final layer can be replaced by the number of bits the network should predict (up to 16 supported). 0or1means 1 bit is predicted.
- The
Training and Inference
- Use
net.Tally4(sample, ...)to tally samples during training. - Use
net.Infer2(sample)to predict values from the network.
Compatibility
Neurlang Classifier is compatible with Go versions 1.13 and above. For CUDA-based learning, Go 1.16 with CUDA dependencies is required.
License
Neurlang Classifier is licensed under Apache 2.0 or Public Domain, at your option.
Directories
¶
| Path | Synopsis |
|---|---|
|
Package cmd contains various demo programs
|
Package cmd contains various demo programs |
|
Package datasets implements the Neurlang dataset type
|
Package datasets implements the Neurlang dataset type |
|
isalnum
Package isalnum implements the IsAlnum Dataset
|
Package isalnum implements the IsAlnum Dataset |
|
isvirus
Package isvirus contains TLSH hashes of viruses and clean files dataset for machine learning (without leading "T1" characters)
|
Package isvirus contains TLSH hashes of viruses and clean files dataset for machine learning (without leading "T1" characters) |
|
mnist
Package MNIST is the 60000 + 10000 handwritten digits dataset
|
Package MNIST is the 60000 + 10000 handwritten digits dataset |
|
Package Hash implements the fast modular hash used by the Neurlang classifier
|
Package Hash implements the fast modular hash used by the Neurlang classifier |
|
Package Hashtron implements a hashtron (classifier)
|
Package Hashtron implements a hashtron (classifier) |
|
Package layer defines a custom combiner and layer interface
|
Package layer defines a custom combiner and layer interface |
|
conv2d
Package conv2d implements a 2D bit-convolution layer and combiner
|
Package conv2d implements a 2D bit-convolution layer and combiner |
|
crossattention
Package crossattention implements a cross attetion connected layer and combiner
|
Package crossattention implements a cross attetion connected layer and combiner |
|
full
Package full implements a fully connected layer and combiner
|
Package full implements a fully connected layer and combiner |
|
majpool2d
Package majpool2d implements a 2D majority pooling layer and combiner
|
Package majpool2d implements a 2D majority pooling layer and combiner |
|
parity
Package parity implements a parity layer and combiner
|
Package parity implements a parity layer and combiner |
|
sochastic
Package sochastic implements a sochastic connected layer and combiner
|
Package sochastic implements a sochastic connected layer and combiner |
|
sum
Package sum implements a sum layer and combiner
|
Package sum implements a sum layer and combiner |
|
Package Learning implements the learning stage of the Neurlang classifier
|
Package Learning implements the learning stage of the Neurlang classifier |
|
avx
Package AVX implements the learning stage of the Neurlang classifier on AVX
|
Package AVX implements the learning stage of the Neurlang classifier on AVX |
|
cu
Package Cu implements the learning stage of the Neurlang classifier on CUDA
|
Package Cu implements the learning stage of the Neurlang classifier on CUDA |
|
Package net implements various hashtron network types
|
Package net implements various hashtron network types |
|
feedforward
Package feedforward implements a feedforward network type
|
Package feedforward implements a feedforward network type |
|
package parallel contains parallel LoopUntil() and parallel ForEach() plus other concurrency primitives.
|
package parallel contains parallel LoopUntil() and parallel ForEach() plus other concurrency primitives. |
Click to show internal directories.
Click to hide internal directories.