Entity Register

DNA focuses on entities that are discussed and shared throughout the conversation. This entity register is mapped in a 3-dimensional vector representation that looks a bit like a web of first, second, and third-degree connections.

The following breaks down how entities are scored by DNA:

• Relevance (range(0,1)): Checks how closely an entity is related to the current topic of the conversation.

• Depth (range(0,1)): Looks at how deeply an entity is connected to the details of the discussion.

• Relationship (range(0,1)): Measures the connections between the entity and other topics or entities talked about.

• Context (range(0,1)): Considers how well the entity fits into the overall subject of the conversation.

• Emotional Resonance (range(0,1)): This one is about how much the entity affects the feelings of the person in the

• No. Of mentions: This is the total number of time an entity has been mentioned in a conversation over different sessions.

• Score(E) = w1 * Relevance(E) + w2 * Depth(E) + w3 * Relationship(E) + w4 * Context(E) + w5 * Emotional Resonance(E) + w6 * (Number of mentions of E / Maximum number of mentions observed.