Papers

    Papers & Publications

    A collection of my research work, publications, preprints, essays, and patents.

    2026

    Open attached file
    📝

    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.

    #preprint#PDE#inverse problems
    Year
    2026
    📝

    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.

    #preprint#PDE
    Year
    2026

    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.

    #patent#AI
    Year
    2025
    ⚡

    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.

    #patent#evaluation
    Year
    2025
    ⚡

    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.

    #patent#human-in-the-loop
    Year
    2025
    ⚡

    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.

    #patent#safety
    Year
    2025