Interface MemoryGraphConfig

Configuration for the memory graph subsystem.

The memory graph powers spreading activation (Collins & Quillian model), Hebbian co-activation learning ("neurons that fire together wire together"), conflict detection, clustering, and graph-boosted retrieval scoring.

Enabled by default when CognitiveMemoryManager is initialized. Set disabled: true to opt out entirely.

interface MemoryGraphConfig {
    disabled?: boolean;
    backend?: "graphology" | "knowledge-graph";
    maxDepth?: number;
    decayPerHop?: number;
    activationThreshold?: number;
    hebbianLearningRate?: number;
}

Properties

disabled? backend? maxDepth? decayPerHop? activationThreshold? hebbianLearningRate?

Properties

Optional disabled

disabled?: boolean

Set to true to disable the memory graph entirely. When disabled, spreading activation, Hebbian co-activation, and graph-based retrieval boosting are all skipped.

Default

false

Optional backend

backend?: "graphology" | "knowledge-graph"

Which graph backend to use.

Default

'knowledge-graph'

Optional maxDepth

maxDepth?: number

Max hops for spreading activation.

Default

3

Optional decayPerHop

decayPerHop?: number

Activation decay per hop (0-1).

Default

0.5

Optional activationThreshold

activationThreshold?: number

Minimum activation to continue spreading (0-1).

Default

0.1

Optional hebbianLearningRate

hebbianLearningRate?: number

Hebbian learning rate for co-activation edge strengthening (0-1).

Default

0.1

Settings

Member Visibility

Theme

On This Page

disabledbackendmaxDepthdecayPerHopactivationThresholdhebbianLearningRate