Interface IContentClassifier<T>

Namespace

AiDotNet.AdversarialRobustness.Safety

Assembly

AiDotNet.dll

Defines the interface for ML-based content classification.

public interface IContentClassifier<T>

Type Parameters

T

The numeric data type used for calculations.

Remarks

Content classifiers provide machine learning-based analysis of content to detect harmful, toxic, or inappropriate material. Unlike regex-based pattern matching, ML classifiers can understand semantic meaning and context.

For Beginners: Think of this as an AI-powered content moderator that can understand the meaning of text, not just look for specific keywords. It can detect subtle forms of harmful content that simple pattern matching would miss.

Methods

Classify(Vector<T>)

Classifies content and returns the classification result.

ContentClassificationResult<T> Classify(Vector<T> content)

Parameters

content Vector<T>

The content to classify as a vector representation.

Returns

ContentClassificationResult<T>

The classification result with category predictions and confidence scores.

ClassifyBatch(Matrix<T>)

Classifies a batch of content items.

ContentClassificationResult<T>[] ClassifyBatch(Matrix<T> contents)

Parameters

contents Matrix<T>

Matrix where each row is a content item to classify.

Returns

ContentClassificationResult<T>[]

Array of classification results, one per input row.

ClassifyText(string)

Classifies content provided as text.

ContentClassificationResult<T> ClassifyText(string text)

Parameters

text string

The text content to classify.

Returns

ContentClassificationResult<T>

The classification result with category predictions and confidence scores.

GetSupportedCategories()

Gets the list of content categories this classifier can detect.

string[] GetSupportedCategories()

Returns

string[]

Array of category names supported by this classifier.

IsReady()

Checks if the classifier is ready to make predictions.

bool IsReady()

Returns

bool

True if the model is loaded and ready, false otherwise.