Research project
4 chapters inside one connected body of work
Skill Memory Mapping System.
A research project spanning the memory model, graph structure, human edge cases, and the simulation evidence behind the system.
A connected research project on how memory systems should be modeled, structured, stress-tested, and validated before they become product behavior.
Date
April 2026
Status
Ongoing
Chapters
4
Access
Password protected
Project overview
One project that moves from the model to the schema, the human failure modes, and the validation evidence.
This is one research project, not a set of unrelated papers. The work starts with the core memory model, moves through the graph schema that supports it, then pressure-tests the design against real human behavior and simulation evidence.
The chapters are meant to be read together. Each one answers a different layer of the same question: how do we build a memory system that behaves well in the real world, not just in a neat mathematical abstraction?
Use the chapter navigation below to move through the project in order, or jump straight to the layer you care about most: the model, the structure, the failure modes, or the evidence.
Chapter navigation
The chapters stay inside this project page. Unlock the project to jump directly to each chapter section.
Skill Memory Mapping System
A research model for tracking difficulty, stability, and retrievability across a growing skill graph.
Graph Schema: Design, Gaps, and Recommendations
A structural review of the current graph model, the gaps caused by its flat schema, and the interventions needed for scoped planning and validation.
When Humans Break the Algorithm
A stress test of the algorithm against real human behavior, from noisy ratings to long absences and repeated lapses.
When Humans Break the Algorithm — Simulation Evidence
Simulation-backed evidence showing how the live algorithm behaves under eight predictable human error patterns.