Patent Portfolio
This portfolio highlights selected intellectual property developed through applied systems thinking and invention. Each entry is presented in plain language to clarify intent, application, and current status.
How These Patents Are Presented
Patents are summarized to emphasize the problem addressed, the underlying insight, and potential applications. Legal scope and claims are discussed during licensing conversations rather than published here.
Regime Validation Engine for High‑Risk Systems
Complex models often produce confident outputs even when their underlying assumptions no longer hold. This creates hidden risk for organizations relying on simulations, AI, or predictive systems in critical decisions.
Our Regime Validation Engine is being developed to identify when analytical methods remain appropriate, when reliability begins to degrade, and when their use should be restricted or suspended. The approach is designed to help organizations avoid false insights, reduce decision risk, and apply advanced analysis only within regimes where underlying assumptions remain defensible.
Current development is expanding evaluation across multiple analytical models, strengthening the engine’s ability to distinguish valid operating regimes from conditions where model outputs require additional caution or constraint.
Applications
AI & Machine Learning R&D
Simulation & Digital Twins
High‑Risk Decision Platforms
Model Governance & AI Safety
Status
Patent Pending (US)
Availability
Licensing or acquisition discussion
Boundary-Safe Framework for Interpreting Complex Systems
Organizations can misclassify complex biological, cognitive, and behavioral phenomena by forcing them into explanatory models that do not fit the underlying conditions. In high-stakes research, safety, and evaluation contexts, these interpretation errors can lead to flawed decisions, misplaced interventions, and conclusions that extend beyond what the evidence supports.
Our Boundary-Safe Framework is being developed to keep distinct functional domains analytically separate while still allowing coordinated evaluation across them. The framework is designed to establish clearer boundaries, identify when an analytical model is being applied beyond its appropriate scope, and reduce diagnostic, causal, or interpretive overreach.
Current development includes a functional analysis workflow with model-scope detection, state tracking, reproducible workflow snapshots, and supporting test infrastructure. Evaluation is now expanding across varied analytical scenarios, with a focus on preserving domain boundaries and strengthening auditability as conditions evolve.
Applications
Consistent neurocognitive classification standards
AI and human‑systems teams
Policy risk analysis behavioral phenomena interpretability
Enterprise platforms enabling boundary-aware data analysis
Status
Patent Pending (US)
Availability
Licensing or acquisition discussion
Resilient Execution Control for Long‑Running Computational Systems
Complex software systems can become unstable when calculations produce undefined results, exceed expected operating limits, or encounter conditions outside the range anticipated by the original execution logic. In long-running or high-demand computational environments, these failures can interrupt continuity, propagate through dependent processes, and require costly recovery or restart.
Our Resilient Execution Control framework is being developed to manage unstable computational outcomes as execution conditions rather than treating them as valid system states. The approach is designed to redirect execution through protected control states, preserve continuity where appropriate, and reduce the risk of localized instability cascading into broader system failure.
Current development includes anomaly detection, configurable execution thresholds, lifecycle management, concurrent-state handling, and checkpoint-and-restore capabilities supported by an expanding automated test suite. Evaluation is now extending to longer-duration workloads, recovery edge cases, inter-domain dependencies, and performance under increased computational demand.
Applications
Mission-critical continuous extreme simulation platforms
Robust adaptive AI analytics pipelines
Long-horizon optimization planning systems
Fail-stop simulation environments
Status
Patent Pending (US)
Availability
Licensing or acquisition discussion
Dynamic Regulatory Framework for Non‑Diagnostic Neurocognitive State Classification
Neurocognitive assessment systems rely on static diagnostic labels that fail to capture dynamic regulation, variability, and recovery over time. These approaches leave gaps in population coverage and often pathologize differences rather than describing functional states.
This invention introduces a non‑diagnostic classification framework that organizes neurocognitive function into a complete set of dynamic regulatory states. It provides a living reference for balance and recalibration that supports longitudinal understanding without enforcing normative ideals.
The underlying system has completed comprehensive engineering validation:
Primary Results
- 92.7% subject-level LOSO accuracy (51/55 subjects) with nested out-of-sample calibration
- 100% accuracy at 87% coverage when abstention is permitted
- 94.4% accuracy at 98.2% coverage under standard confidence thresholds
- Multi-condition generalization confirmed across datasets
Calibration & Safety
- Per-fold temperature scaling with nested validation architecture — no data leakage between calibration and evaluation
- Post-calibration confidence closely matches empirical accuracy across all operating points
- Abstention gate refuses classification on uncertain subjects rather than producing errors
- Quality-aware data inclusion recovered 57% of additional subjects without accuracy degradation
Infrastructure
- Full 55-fold LOSO with nested calibration completes in under 10 minutes on standard cloud GPU
- Demographic confound analysis integrated (fails gracefully when metadata is unavailable)
- Classification confirmed to reflect genuine neurophysiological patterns, not circular label memorization
- Multi-format dataset ingestion with automated quality gating
Foundational EEG Research
This research established the conceptual foundation for Reinvent Develop’s EEG modeling work and the subsequent development of our multi-condition neurocognitive classification framework. The original publication documents the early research direction that evolved into the current system.
Applications
- Neurocognitive state monitoring
- Adaptive wellness platforms
- Human performance optimization
- Non‑diagnostic mental health analytics
- Clinical research support tools
- EEG device integration
Status
Patent Pending (US)
Availability
Licensing or acquisition discussion