Learn by building.
Follow the standard from a recorded memory to feeling, context and a visible relationship.
1. Build your first Memory Atom
Start with a memory record. Add one primary affect state with VAD. Add a sourced person, place or situation. Keep context separate from the interpretation, then link it to the relevant feeling.
- Read the core definition.
- Use the logical schema to identify the three parts.
- Open the context proximity lab and try assessed versus unknown influence.
- Download the lab's JSON. Notice that influence is separate from VAD and confidence.
2. Explore contextual similarity
Download the two-memory fixture. The records share a person and place, but have different feelings.
- Compare context entity IDs rather than names alone.
- Observe the VAD differences. Matching context does not make the emotions identical.
- List the shared context as an explanation. Do not invent a similarity percentage: weighting and calibration remain unestablished.
Expected result: two matching contextual entities, two independent memories and different affect coordinates.
Sample project: context proximity lab
A small, dependency-free browser project. Open its HTML file locally, move the influence slider and export the example. It demonstrates a relationship rule rather than a production emotion estimator.
Reference downloads
The schema is illustrative, not a complete runtime validator. Samples use synthetic content and demonstration values. Read the methods that still require calibration before assigning real scores.