AveragedGEM<T>
Models & Types · AiDotNet.ContinualLearning
Implements Averaged Gradient Episodic Memory (A-GEM) for continual learning.
Implements: IContinualLearningStrategy<T>
For Beginners
A-GEM is a more efficient version of GEM that uses a single random sample from the episodic memory instead of checking constraints against all stored examples. This makes it much faster while maintaining similar performance.
How It Works
How it works:
Key Difference from GEM:
Projection Formula:
If g · g_ref < 0: g_proj = g - (g · g_ref / g_ref · g_ref) × g_ref
Reference: Chaudhry, A. et al. “Efficient Lifelong Learning with A-GEM” (2019). ICLR.
Constructors
| Constructor | Summary |
|---|---|
AveragedGEM(int memorySize, int sampleSize, double lambda, int? seed) |
Initializes a new instance of the AveragedGEM class. |
Properties
| Property | Summary |
|---|---|
bool AccumulatesAcrossTasks { get; } |
|
double Lambda { get; set; } |
|
int TaskCount { get; } |
Gets the number of tasks stored in episodic memory. |
int TotalMemorySize { get; } |
Gets the total number of samples in episodic memory. |
Methods
| Method | Summary |
|---|---|
T ComputeLoss(INeuralNetwork<T> network) |
|
Vector<T> ModifyGradients(INeuralNetwork<T> network, Vector<T> gradients) |
|
void AfterTask(INeuralNetwork<T> network, ValueTuple<Tensor<T>, Tensor<T>> taskData, int taskId) |
|
void BeforeTask(INeuralNetwork<T> network, int taskId) |
|
void Reset() |