Visualization View

An interactive visual representation of memory relationships:

Memory Statistics: Same statistics as table view:

• Total: Total number of memories in the system
• Filtered: Number of memories matching current filters
• Knowledge: Count of knowledge-based entries
• Conversation: Count of conversation-based entries

Visualization Settings:

Algorithm Selection: Choose the dimensionality reduction algorithm:

• t-SNE (t-Distributed Stochastic Neighbor Embedding):
  • Best for: Revealing clusters and local structure
  • Use when: You want to see how memories group together
  • Characteristics: Preserves local relationships, good for pattern discovery
• UMAP (Uniform Manifold Approximation and Projection):
  • Best for: Preserving global structure while showing local patterns
  • Use when: You want to understand overall memory organization
  • Characteristics: Faster than t-SNE, maintains both local and global structure

Fig 4.3

Points Setting: Control how many memory points to visualize:

• Range: Typically 100 to 2000 points
• Default: 500 points
• Purpose: Balance between detail and performance