BarlowTwins<T>

Models & Types · AiDotNet.SelfSupervisedLearning

Barlow Twins: Self-Supervised Learning via Redundancy Reduction.

Inheritance: ModelBase<T, Tensor<T>, Tensor<T>> → SelfSupervisedLearningMethodBase<T> → BarlowTwins<T>

For Beginners

Barlow Twins learns representations by making the cross-correlation matrix between embeddings of two augmented views close to the identity matrix. This achieves both invariance (diagonal = 1) and reduces redundancy between features (off-diagonal = 0).

How It Works

Key innovations:

Loss components:

Cross-correlation matrix:

Reference: Zbontar et al., “Barlow Twins: Self-Supervised Learning via Redundancy Reduction” (ICML 2021)

Best for: Interpretable approach, avoiding collapse naturally.

Pros: Interpretable loss, naturally avoids collapse, no negative samples.

Cons: Requires careful scaling of loss terms.

Constructors

Constructor Summary
BarlowTwins(INeuralNetwork<T> encoder, IProjectorHead<T> projector, SelfSupervisedLearningConfig<T> config) Initializes a new instance of the BarlowTwins class.

Properties

Property Summary
SelfSupervisedLearningMethodCategory Category { get; }
bool RequiresMemoryBank { get; }
bool UsesMomentumEncoder { get; }
double Lambda { get; } Gets the lambda (redundancy reduction weight) used in the loss.
string Name { get; }

Methods

Method Summary
BarlowTwins<T> Create(INeuralNetwork<T> encoder, int encoderOutputDim, int projectionDim, int hiddenDim, double lambda) Creates a Barlow Twins instance with default configuration.