KairosOptions<T>

Options & Configuration · AiDotNet.Models.Options

Configuration options for Kairos (Adaptive and Generalizable Time Series Foundation Model).

Inheritance: ModelOptions → RegressionOptions<T> → TimeSeriesRegressionOptions<T> → KairosOptions<T>

For Beginners

Kairos adapts its tokenization to your data:

Adaptive Tokenization: Unlike fixed-size patching, Kairos uses multiple patch sizes simultaneously and a learned router decides which granularity is best for each segment. Dense/volatile regions get fine-grained tokens; smooth regions get coarse tokens.

Mixture-of-Size Encoder: Multiple encoder branches process patches at different sizes, then a gating mechanism combines the results based on local information density.

How It Works

Kairos uses a Mixture-of-Size Encoder with adaptive tokenization that adjusts patch granularity based on local information density. This parameter-efficient approach handles diverse time series characteristics without fixed tokenization.

Reference: “Kairos: Towards Adaptive and Generalizable Time Series Foundation Models”, 2025. https://arxiv.org/abs/2509.25826

Constructors

Constructor Summary
KairosOptions() Initializes a new instance with default values.
KairosOptions(KairosOptions<T> other) Initializes a new instance by copying from another instance.

Properties

Property Summary
FoundationModelSize ModelSize { get; set; } Gets or sets the model size variant.
double DropoutRate { get; set; } Gets or sets the dropout rate.
double LearningRate { get; set; } Gets or sets the peak Adam learning rate used when no optimizer is supplied.
int ContextLength { get; set; } Gets or sets the context length.
int ForecastHorizon { get; set; } Gets or sets the forecast horizon.
int HiddenDimension { get; set; } Gets or sets the hidden dimension.
int IntermediateSize { get; set; } Gets or sets the intermediate size for the feed-forward network.
int NumHeads { get; set; } Gets or sets the number of attention heads.
int NumLayers { get; set; } Gets or sets the number of transformer layers.
int WarmupSteps { get; set; } Gets or sets how many optimizer steps the learning rate ramps linearly up to LearningRate.
int[] PatchSizes { get; set; } Gets or sets the multiple patch sizes for adaptive tokenization.