Publications
Publications and insights
A reading library of technical notes, mathematical explanations and engineering proposals. Every piece is written by Menlo Tech, reasoned from first principles, and checked numerically. None claims experimental results we have not obtained, and none is peer reviewed.
- 01Choosing a confidence threshold from the cost of being wrong
Research noteA short derivation of when an automated decision should go to a person. - 02The payment formula, derived
Mathematical explanationWhere PMT = P r(1+r)^n / ((1+r)^n − 1) comes from, step by step. - 03Cosine similarity: why vectors can retrieve meaning
Mathematical explanationThe geometry behind semantic retrieval, with a worked example. - 04From scores to decisions: priority and capacity
Technical articleA priority score orders the work. A knapsack decides how much of it fits. - 05Markov chains for journeys and collection states
Mathematical explanationAbsorption probabilities and expected steps, with a three-state example. - 06Calibration: when 90% should mean 90%
Engineering paper (proposal)How to measure calibration error and a proposed design for monitoring it. - 07Multi-Agent Coordination
Technical experimentCompare a simple allocation rule with the exact optimum. - 08Decision Optimization
Technical experimentChoose actions within a limited number of hours. - 09Agent Evaluation
Technical experimentReliability diagram, calibration error and the cost-optimal threshold. - 10Enterprise Knowledge Graph
Technical experimentShortest paths and downstream impact on a graph with provenance. - 11Predictive Modeling
Technical experimentExponential smoothing on a seasonal series and an absorbing Markov chain. - 12Mathematical Optimization
Technical experimentA two-variable linear program, solved exactly.
No industry research has been published yet. This category is reserved; nothing is invented to fill it.
No results for this filter.