Table of Contents

Class MetaLearningTaskBase<T, TInput, TOutput>

Namespace
AiDotNet.Data.Structures
Assembly
AiDotNet.dll

Abstract base class for meta-learning tasks, providing common functionality and validation.

public abstract class MetaLearningTaskBase<T, TInput, TOutput> : IMetaLearningTask<T, TInput, TOutput>

Type Parameters

T

The numeric type used for calculations (e.g., float, double).

TInput

The input data type (e.g., Matrix<T>, Tensor<T>).

TOutput

The output data type (e.g., Vector<T>, Tensor<T>).

Inheritance
MetaLearningTaskBase<T, TInput, TOutput>
Implements
IMetaLearningTask<T, TInput, TOutput>
Inherited Members

Remarks

This base class implements the IMetaLearningTask<T, TInput, TOutput> interface and provides common functionality for all meta-learning task implementations. It includes validation, default values, and helper methods to reduce code duplication across concrete implementations.

Concrete implementations should inherit from this class and provide any additional functionality specific to their meta-learning scenario (e.g., episode tracking, hierarchical relationships, continual learning).

Constructors

MetaLearningTaskBase(int, int, int, string?, Dictionary<string, object>?)

Initializes a new instance of the MetaLearningTaskBase class.

protected MetaLearningTaskBase(int numWays = 5, int numShots = 1, int numQueryPerClass = 15, string? name = null, Dictionary<string, object>? metadata = null)

Parameters

numWays int

Number of classes (ways) in this task.

numShots int

Number of examples per class in the support set.

numQueryPerClass int

Number of query examples per class.

name string

Optional name for the task.

metadata Dictionary<string, object>

Optional metadata dictionary.

Properties

Metadata

Gets the additional metadata about the task.

public virtual Dictionary<string, object>? Metadata { get; }

Property Value

Dictionary<string, object>

Dictionary containing task-specific information.

Name

Gets or sets an optional name or identifier for the task.

public virtual string? Name { get; set; }

Property Value

string

Human-readable task name or null if not specified.

NumQueryPerClass

Gets the number of query examples per class.

public int NumQueryPerClass { get; protected set; }

Property Value

int

Remarks

The number of examples in the query set for each class.

NumShots

Gets the number of shots (examples per class) in the support set.

public int NumShots { get; protected set; }

Property Value

int

Remarks

In N-way K-shot learning, this represents the K (number of examples per class).

NumWays

Gets the number of ways (classes) in this task.

public int NumWays { get; protected set; }

Property Value

int

Remarks

In N-way K-shot learning, this represents the N (number of classes per task).

QueryInput

Gets or sets the input features for the query set.

public virtual TInput QueryInput { get; set; }

Property Value

TInput

Input data for evaluating adaptation performance.

Exceptions

ArgumentNullException

Thrown when trying to set null value.

QueryOutput

Gets or sets the target labels for the query set.

public virtual TOutput QueryOutput { get; set; }

Property Value

TOutput

True labels for evaluating query set performance.

Exceptions

ArgumentNullException

Thrown when trying to set null value.

QuerySetX

Gets the input features for the query set (alias for QueryInput).

public TInput QuerySetX { get; }

Property Value

TInput

Remarks

This property provides compatibility with legacy code that uses QuerySetX naming convention.

QuerySetY

Gets the target labels for the query set (alias for QueryOutput).

public TOutput QuerySetY { get; }

Property Value

TOutput

Remarks

This property provides compatibility with legacy code that uses QuerySetY naming convention.

SupportInput

Gets or sets the input features for the support set.

public virtual TInput SupportInput { get; set; }

Property Value

TInput

Input data containing examples for task adaptation.

Exceptions

ArgumentNullException

Thrown when trying to set null value.

SupportOutput

Gets or sets the target labels for the support set.

public virtual TOutput SupportOutput { get; set; }

Property Value

TOutput

Output data containing labels corresponding to SupportInput.

Exceptions

ArgumentNullException

Thrown when trying to set null value.

SupportSetX

Gets the input features for the support set (alias for SupportInput).

public TInput SupportSetX { get; }

Property Value

TInput

Remarks

This property provides compatibility with legacy code that uses SupportSetX naming convention.

SupportSetY

Gets the target labels for the support set (alias for SupportOutput).

public TOutput SupportSetY { get; }

Property Value

TOutput

Remarks

This property provides compatibility with legacy code that uses SupportSetY naming convention.

TaskId

Gets or sets an optional identifier for the task.

public virtual int? TaskId { get; set; }

Property Value

int?

Integer identifier for the task or null if not specified.

Methods

SetMetadata(string, object)

Adds or updates a metadata entry.

public void SetMetadata(string key, object value)

Parameters

key string

The metadata key.

value object

The metadata value.

Exceptions

ArgumentNullException

Thrown when key is null.

ToString()

Creates a string representation of the task.

public override string ToString()

Returns

string

String containing task name and type information.

TryGetMetadata<TValue>(string, out TValue?)

Gets a metadata value if it exists.

public bool TryGetMetadata<TValue>(string key, out TValue? value)

Parameters

key string

The metadata key.

value TValue

Output for the value if found.

Returns

bool

True if the value was found and successfully cast to TValue.

Type Parameters

TValue

The type of value to retrieve.

Validate()

Validates that the task has all required data populated.

public virtual void Validate()

Exceptions

InvalidOperationException

Thrown when required data is missing.

Edit this page