Papers & Publications
A collection of my research work, publications, preprints, essays, and patents.
2026
Reconstruction of the Initial Condition for a Non-Homogeneous Heat Equation from Finite Measurements
Authors not listed
Shows how to rigorously reconstruct initial temperature distribution of a 1-D system for the ill-posed non-homogeneous heat equation with Dirichlet boundary conditions.
Reconstruction of the Initial Condition for a Complex Non-Homogeneous Heat Equation from Finite Measurements
S. Frerichs, et al.
Demonstrates how to rigorously reconstruct the unknown initial state of a generalized, complex-valued system—such as the Schrödinger equation or a heat equation with periodic boundaries using a finite number of discrete measurements.
2025
Intent-Conscious Data Unit (ICDU) Systems and Methods
Provisional # 63/923,146
A new structured data format for AI training that teaches models why a user is asking something — not just the text. It encodes user intent, governing principles (like safety or tone), persona, and context so the model learns to reason, not memorize.
AI Judge Systems for Qualitative and Quantitative Model Evaluation
Provisional # 63/923,149
Uses novel testing models and evaluation pipelines to judge an AI's learning evolution via separate AI model.
Human-in-the-Loop Quality Control and Nuance Guiding Systems
Provisional # 63/923,151
A human-in-the-loop system for grading the subjective parts of AI (empathy, clarity, coaching quality). Humans score with a rubric schema based on unique testing models, an AI judge consolidates their feedback, and incorporates it into the training pipeline.
AI Stress-Test and Scenario Perturbation Systems
Provisional # 63/923,152
A “stress-test generator” for AI. It creates controlled variations (role, tone, channel, constraints, etc.) around a scenario to see if the model stays consistent. It measures stability, fairness, invariance, and failure modes, and acts as a safety gate.