Digital Commons Network^™

Dusty Plasma-Photocatalytic Reactor Technology For Gas Emission Treatment, N/A

Defensive Publications Series

This publication describes a dielectric barrier discharge plasma reactor configured to treat fluorinated greenhouse gas, sulfur hexafluoride, and nitrous oxide emissions. The technology integrates non-thermal plasma, localized heating assisted by nanocatalysts, and photocatalysis to decompose industrial emissions. Active species in the form of very small nanocatalysts are introduced into the plasma, allowing the nanocatalysts to become locally hot while the surrounding gas remains cool. A carrier gas generates light within the reactor to stimulate the catalytic materials. Various electrode configurations and sacrificial materials are employed to manage plasma density and capture reaction byproducts.

Keywords: dielectric barrier discharge, fluorinated greenhouse …

Plasma Jet Curtain, N/A

Defensive Publications Series

This publication describes a technology for the treatment of fluorinated greenhouse gas, sulfur hexafluoride, and nitrous oxide emissions. A plasma jet, which can be accompanied by a carrier gas, is directed into a low-pressure environment. The chamber includes a porous catalyst bed and a pressure differential is utilized behind the bed to control the dispersion of the plasma. Light is introduced near the plasma jet to illuminate the catalyst. This arrangement allows for plasma-based activation and species breakdown, while the degradation products are processed by the photocatalyst.

Keywords: plasma jet, fluorinated greenhouse gas, abatement, photocatalyst, pressure differential, bias, semiconductor …

Dynamic Resolution Of Security Configurations Via Sequential Rule Processing During Conference Instantiation And Label Transitions, N/A, N/A, N/A

Defensive Publications Series

Traditional data loss prevention configurations for video conferences are often statically applied during scheduling. Organizational policy updates occurring before a meeting start or sensitivity changes arising during a live session may not be accounted for by such approaches. In this disclosure, security rules are dynamically evaluated at conference instantiation and whenever classification labels are modified in-call. Rules associated with multiple labels are aggregated and processed sequentially, where the final rule in the evaluation chain determines the setting state. A resolved security configuration is applied to the conference instance in a single atomic update. Current organizational policies are consistently enforced regardless …

Conflict Resolution For Data Loss Prevention Rules In Video Communication Via Strict Linear Precedence, N/A, N/A, N/A

Defensive Publications Series

Conflicts in data protection policy enforcement for video conferencing are often addressed using subjective hierarchies of rule strictness or sensitivity. Such methods produce unpredictable results when multiple overlapping policies are applied to a single session. To address these limitations, disclosed is a method for evaluating data loss prevention rules using strict linear precedence. Rules are processed sequentially according to their order in a predefined administrative list. Every rule evaluated later in the sequence overrides any previous settings for the specific parameters it modifies, regardless of the perceived rigor of the rule. Evaluation is performed during session instantiation or when classification …

Llm Simulation Of Assistive Technology User Journeys For Temporal Accessibility Analysis, Nicolas Zimmermann, Isaac Leimgruber

Defensive Publications Series

Some automated accessibility testing tools analyze screens in isolation and can fail to detect complex, journey-level barriers that emerge over multiple interactions. A system can analyze the temporal experience of an assistive technology (AT) user by programmatically orchestrating a user journey and capturing a sequence of states, which may include the accessibility tree and focus state at each step. This sequential data can then be provided to a large language model. The model can be prompted to simulate a screen reader user's experience, evaluating the flow for issues related to focus expectation, context retention, and navigation consistency. This approach may …

Progress-Aware Framework For Governing Llm-Driven Evolutionary Search, Minghao Yan, Zhankui He, Chi Wang, Ed Chi, Ben Coleman, Ziqi Chen, Zhouhang Xie, Shuo Chen, Noveen Sachdeva, Beidou Wang, Isabella Ye, Derek Cheng, Weili Wang, Wang-Cheng Kang, Fernando Pereira, Bo Peng

Defensive Publications Series

Evolutionary search processes that use large language models may exhibit instability and inefficiency, which can lead to stagnation, context pollution, and suboptimal collaboration in parallel runs. A progress-aware framework can be used to govern these searches. The framework may integrate components such as hierarchical context management for organizing the search into a structured idea pool, momentum-based backtracking for adaptively escaping local minima by tracking a progress metric, and self-adaptive collaborative evolution for dynamically managing knowledge transfer between parallel agents. Functioning as a data-driven control system, the framework can make adaptive decisions regarding, for example, pruning, backtracking, and collaboration. This approach …

A Structurally Isomorphic Architecture For Real-Time Inference, N/A

Defensive Publications Series

This publication describes a structurally isomorphic architecture tailored for real-time inference in machine learning models. The architecture provides a dedicated pipelined structure that allows for the efficient execution of neural network computations by distributing tasks across discrete hardware units mapped to specific stages of a model. Features include synchronous operation, aggressive clock gating, and memory optimization based on capacity and throughput parameters. The design integrates compute capabilities directly with high-bandwidth memory stacks to reduce latency and power consumption.

Keywords: Isomorphic Architecture, Real-Time Inference, Hardware Pipeline, Clock Gating, In-Memory Computing, Large Language Models.

Modular Rack-Level Closed-Loop Liquid Cooling System For Air-Cooled Data Center Environments, N/A

Defensive Publications Series

A rack-level closed-loop liquid cooling system is disclosed to enable high-density liquid-cooled information technology equipment to operate within standard air-cooled racks. The modular air-assisted liquid cooling distribution unit can be strategically sized for deployment in a sidecar configuration or as an individual unit within the same rack as the equipment it cools. The system utilizes a redundant and modular architecture, safety mechanisms to prevent simultaneous chassis extension, liquid leak prevention mechanisms, and control logic for system protection and efficiency.

Keywords: liquid cooling, coolant distribution unit, data center, air-assisted cooling, thermal management, redundant architecture.

System For Determining Subscription Efficiency By Correlating Application Usage Telemetry With Financial Notification Signals, Akash Verma, Vivek Yadav

Defensive Publications Series

Managing the costs of multiple streaming services, such as Video on Demand (VOD), Over-The-Top (OTT) media, or recurring digital memberships, can be difficult because billing cycles are often disconnected from how much a person actually uses the apps, leading to “zombie subscriptions,” or unused subscriptions that are hidden recurring payments for services that are rarely watched. The disclosed system acts as an automated financial dashboard that helps users save money and improve financial health by automatically comparing app usage (how many hours users spend watching) with financial data found in email receipts, app store records, or renewal notifications. By combining …

Carbon Aerogel Adsorbent For Hdd, Lihong Zhang

Defensive Publications Series

Carbon-based adsorbents used in current HDDs’ environmental-control-module (ECM) and recirculating filters have a close-end pore structure. Close-end pore structures usually have relatively low adsorption rate and adsorption capacity, which negatively impact the absorption effectiveness and efficiency. This invention proposes using Activated Carbon Aerogels to replace the conventional Activated Carbon for HDDs’ ECM and recirculating filters. Carbon Aerogels have open-end 3D-interconnected pore structures with a wider pore size distribution (ranging from micro-, meso- to macro pores) compared to Activated Carbons, enabling them to have faster adsorption rates, more effectiveness to a broader range of contaminants and much higher surface utilization (higher …

Privacy-Preserving Local Authorization For Regulated, Safety-Critical, And Financial Systems, Jonathan Stephen Baker

Defensive Publications Series

This disclosure describes systems and methods for controlling access to regulated, age-restricted, safety-critical, high-liability, and financial devices using privacy-preserving local authorization. In disclosed embodiments, a user device, terminal, embedded controller, or local authorization module verifies that a requesting user satisfies one or more access conditions before enabling operation of a protected physical or digital system. The authorization process may be performed locally or at the edge, without requiring persistent storage of raw biometric data, identity templates, or reusable personal credentials by a relying party. The system outputs a limited authorization result, proof, token, permission state, or access decision indicating that …

Gci Hydrocarbon Normalization & Mechanical Fuel Conditioning Platform (V1.5), Michael Victor Caldwell Mr.

Defensive Publications Series

A modular, low-temperature mechanical processing system designed to normalize variable hydrocarbon waste streams (used motor oil, vegetable oil, or pre-processed pyrolysis liquids). This platform does not perform traditional refining. Instead, it uses hydrodynamic shearing, controlled cavitation, centrifugal phase grading, contaminant removal, and adaptive recirculation to transform inconsistent waste fluids into stable, standardized, combustion-compatible fuel blends suitable for industrial engines, burners, and off-grid power systems.

This system emphasizes selective energy application, multi-pass conditioning, and real-time feedback control to maximize efficiency, reduce waste, and improve fuel usability without relying on high-temperature or catalytic refining processes.

Knowledge Tokens For Multi-Dimensional Epistemic Representation In Ai Systems, Johny Manuel-Devadoss, Deepthi Prasad

Defensive Publications Series

We propose a framework that replaces scalar confidence estimates in AI outputs with structured, multi-dimensional representations termed knowledge tokens. Each token encodes orthogonal epistemic attributes including provenance weight, empirical grounding, consensus density, and temporal freshness. Unlike conventional probabilistic confidence scores, knowledge tokens provide composable, inspectable, and tradeable representations of informational quality. This enables downstream systems to reason not only about likelihood but about the structure and reliability of knowledge itself. We formalize the representation, integration into model architectures, training objectives, and evaluation methodologies, and outline how such tokens enable new forms of reasoning, aggregation, and economic exchange over knowledge.

Market-Based Adversarial Verification For Ai Outputs, Johny Manuel-Devadoss, Deepthi Prasad

Defensive Publications Series

We introduce a market-based verification framework in which independent agents compete to validate or invalidate AI-generated claims through incentivized mechanisms. Unlike static evaluation pipelines or passive verification models, this system constructs a continuous adversarial marketplace where validators bid for the right to assess claims, stake capital on their judgments, and are rewarded or penalized based on correctness. The framework aligns incentives toward truth detection by making verification a profit-driven activity under uncertainty. We formalize the architecture, design, incentive compatibility, training strategies, and evaluation methodologies. The proposed system reframes truth validation as a decentralized, adversarial process rather than a static model …

An Adversarial Self-Invalidation Framework For Robust Reasoning, Johny Manuel-Devadoss, Deepthi Prasad

Defensive Publications Series

We propose a novel inference-time and training-time paradigm for improving the epistemic reliability of large language models (LLMs) and tool-augmented AI systems, termed Forced Epistemic Reset (FER). Unlike conventional self-consistency or critique-based methods, FER introduces a structured adversarial self-invalidation cycle in which a model is periodically required to challenge, falsify, and independently re-derive its own outputs using heterogeneous reasoning paths. The system explicitly treats intermediate conclusions as provisional hypotheses subject to internal falsification pressure. We formalize the architecture, training objectives, reset triggering mechanisms, and multi-path recomputation strategy. Empirical evaluation protocols are also outlined, focusing on contradiction recovery, robustness under reasoning …

Natural-Language Multi-Tenant Product Configurator Using Llm Tool-Calling Against A Curated Relational Database, Without A Constraint Solver, Christian Hinterdorfer

Defensive Publications Series

This disclosure describes a multi-tenant Configure-Price-Quote (CPQ) software architecture in which a Large Language Model (LLM) configures complex technical products through natural-language conversation while guaranteeing that every selected option is valid for the relevant product line. Validity is enforced not by translating the conversation into a formal constraint language and running a constraint solver (SAT/CSP/MIP/MiniZinc), but by restricting the LLM's action space: the LLM can only invoke a small, fixed set of tools that operate against a curated relational database, and it can only pass option identifiers that the database itself has previously returned. Hallucinated or invalid identifiers are structurally …

Generative Data Refinement With Semantic Preservation, Minqi Jiang, João Madeira Araújo, William Ellsworth, Sian Heather Gooding, Edward Grefenstette

Defensive Publications Series

A computing system can obtain input data containing semantic information. The computing system can access conditioning data configured to cause a generative model to generate modifications to the input data which alter the input data to satisfy a data refinement constraint while preserving the semantic information. The computing system can process the input data with the generative model conditioned on the conditioning data to generate, as an output of the generative model, refined data. According to an aspect of the present disclosure, the refined data can both satisfy the data refinement constraint and also contain the semantic information.

System And Method For Relational Matching Using Non-Presentational Behavioral Signals With Pre-Exposure Bilateral Model Correction, Joseph Jm Walker

Defensive Publications Series

This disclosure establishes prior art for a computer-implemented system and method for privacy-first, values-aligned relational matching and community formation. The system is defined by a fixed operational sequence: private reflection, behavioral observation, simulated interaction modeling, bilateral correction, and conditional identity revelation. Behavioral modeling is performed on non-presentational behavioral signals—user-generated data produced in the absence of a perceived audience—thereby reducing incentive for strategic self-representation and mitigating data distortion common in existing social and dating platforms.

A Behavioral Pattern Recognition Engine (BPRE) generates compatibility distributions grounded in established relational research frameworks. A Simulation and Correction Interface (SCI) presents interpretive models of potential …

Gci Aurora-Grid Ev Structural Battery Pack (G-Sef), Michael Victor Caldwell Mr.

Defensive Publications Series

The Aurora-Grid EV Battery Pack (G-SEF) is a next-generation electric vehicle battery pack architecture that replaces conventional EV pack structures—housing, cooling plates, and busbar assemblies—with a unified gyroid-based structural-energy frame.

The system is designed specifically for electric vehicle applications, where pack-level mass, thermal bottlenecks, and interconnect complexity limit charging speed, range efficiency, and long-term durability.

By embedding mechanical load-bearing, thermal distribution, and electrical routing into a single structural lattice, the design reduces subsystem redundancy while maintaining compatibility with standard lithium-ion EV cells (LFP/NMC).

The result is a structurally integrated EV battery pack intended to improve:

• driving range efficiency (mass reduction) …

  ---

  Go to article

Quantum‑Enhanced Wearable Diagnostic Device, Bala Surya Kiran Rumpavalasa, Sumedha Ghuliani

Defensive Publications Series

Current medical diagnostics for systemic pathologies, which may rely on periodic and invasive clinical procedures, create a gap where sub-cellular shifts may progress undetected until physical symptoms manifest. Existing wearable biometric devices evaluate macro-level physiological lagging indicators and may lack the sensitivity to detect underlying atomic or chemical drivers of disease. Further, internal electromagnetic interference within consumer electronics may disrupt high-sensitivity quantum probes. This disclosure describes a computing device (the “device”) that utilizes a system-in-package containing synthetic diamond enriched with nitrogen-vacancy centers. A nested mu-metal shield encases the hardware to create a quiet zone, which allows the device to perform …

Optical And Inertial Tracking For Extended Reality, Sumukh A. Shevde, Sudeep Chauhan

Defensive Publications Series

This document describes techniques for a system that pairs a wearable device, such as a smartwatch or fitness tracker, with an Extended Reality (XR) device such as an XR headset, Virtual Reality (VR) headset, Augmented Reality (AR) glasses, or Head-Mounted Display (HMD), to improve how the XR device tracks hand movements. Typically, XR devices may use cameras and computer vision for tracking, but this optical tracking may decrease in accuracy when hands are out of view, such as behind a user’s back or under a desk. In some examples, the XR device includes a calibration loop that uses optical data …

Predictive Physical Hardware Actuation Via Llm-Based Semantic Code Analysis, Tamar Yakar, Ilia Labzovsky

Defensive Publications Series

Reactive autoscaling systems may cause performance degradation and thermal issues when deploying new code, as such systems often scale resources based on lagging indicators. A system can address this by integrating a large language model (LLM) into a continuous integration/continuous deployment pipeline. Before deployment, the system analyzes the semantic and structural impact of codebase modifications by processing code diffs and abstract syntax tree deltas, which can be augmented with historical telemetry data. The LLM then generates a resource delta prediction vector to quantify anticipated physical demands, such as central processing unit cycles and thermal output. This prediction may enable a …

Non-Reciprocal Osmotic Rectifier Operating At The Landauer Limit For Water Desalination, Donevin Frownfelter

Defensive Publications Series

A non-reciprocal osmotic rectifier for water desalination operating at the Landauer Limit (W ≥ k_B T ln 2) is disclosed. The system replaces conventional pressure-driven reverse osmosis with Lorentz-force ion separation using 256 REBCO superconducting gates arranged in a breathing double-torus topology around a fused silica microfluidic spine.

https://doi.org/10.5281/zenodo.20012002

https://orcid.org/my-orcid?orcid=0009-0008-7546-6952

Water channels are machined with aperiodic widths scaling by the golden ratio f(n) = 1.618mm · φⁿ to eliminate resonant turbulence. A nanoporous graphene-vortex membrane (σ_g = 178) with matching aperiodic porosity provides mass separation without Euclidean distortion. Unidirectional flow is enforced by orthogonal fluid gates that reflect any reverse …

Mobile Municipal Organic Waste Refinery & Pelletization System, Michael Victor Caldwell Mr.

Defensive Publications Series

This system is a modular municipal organic waste processing platform that converts bagged wet food waste into stabilized fertilizer pellets using a staged mechanical and thermal process:

1. Mechanical shredding

2. Screw-press dewatering

3. Exhaust-driven thermal stabilization (mobile)

4. Controlled belt drying (fixed plant)

5. Nutrient reintroduction loop

6. Ring-die pelletization

The system is designed to minimize landfill diversion load, reduce biological decomposition during transport, and produce standardized fertilizer pellets using lignin-based binding without chemical additives.

Recipient-Specific Watermarking Of Query Results For Live And Api-Delivered Structured Data, Aaron Burton

Defensive Publications Series

Recipient-specific watermarking of query results applies a per-recipient signal to the rows returned from a query, an API endpoint, or a streaming feed, so that structured data delivered as live query responses, rather than as static files, carries an attribution signal that survives normal analytical use. The watermark is embedded in the response payload at the moment the query is answered, and is keyed to the recipient identity associated with the request. When a later-observed copy of the data appears to have originated from the producer, the producer can match residual signals in the copy against the recipient keys it …

Confidential Computing Hosted Watermarking And Verification Of Structured Datasets, Aaron Burton

Defensive Publications Series

Confidential computing hosted watermarking applies and later verifies recipient-bound watermarks on a structured dataset entirely inside a trusted execution environment or a multi-party data clean room, so that the producer's source data is not exposed to the verifier and the verifier's signals are not exposed to the producer. The enclave receives the source dataset, the recipient identity, and the watermark policy, produces a recipient-bound watermarked variant whose carriers are visible only to the enclave, and later, given a suspected derivative, evaluates whether the derivative carries signals consistent with any registered recipient, all without decrypting the source dataset outside the enclave.

Boundary-Time Recipient-Specific Watermarking For Cross-Account Structured Data Sharing, Aaron Burton

Defensive Publications Series

Boundary-time recipient-specific watermarking applies a per-recipient watermark to a structured dataset at the moment the dataset is exposed across a cross-account sharing boundary, where a producer makes a hosted table or view visible to one or more recipient accounts without copying the underlying data. A transformation layer interposed between the producer's source objects and the recipient-visible binding produces a distinct watermarked variant for each recipient, without modifying the source schema or the recipient's tooling. Each recipient reads what appears to be a normal binding, while the producer retains the ability to attribute any later-found copy of the data to a …

Zenbrain: A Neuroscience-Inspired 7-Layer Memory Architecture For Autonomous Ai Systems (V7), Alexander Bering

Defensive Publications Series

ZenBrain is an open-source memory architecture for autonomous AI systems implementing seven distinct memory layers inspired by cognitive neuroscience: working, short-term, episodic, long-term, procedural, core, and predictive memory. A unified MemoryCoordinator integrates fifteen validated neuroscience mechanisms — nine foundational algorithms (vmPFC-coupled FSRS spaced repetition with prediction-error gating, two-factor synaptic Hebbian dynamics for knowledge graphs, simulation-selection sleep consolidation modeling hippocampal replay, Bayesian confidence propagation, compositional context embeddings, spectral knowledge-graph health monitoring, context-adaptive information-bottleneck budgeting, dual-process chain-of-thought consolidation, and iMAD selective debate) and six Predictive Memory Architecture (PMA) components: NeuromodulatorEngine (dopamine/norepinephrine/serotonin/acetylcholine tonic levels with phasic prediction-error bursts and opponent-process dynamics), ReconsolidationEngine (memory …

The Composite Optical System That Contains A Polygonal Composite Light Ray Deviating System And A Composite Cuboidal Light Ray Projection System., Punarjeewa Abeysekera

Defensive Publications Series

This paper describes a composite optical system that contains a polygonal composite light ray deviating system and a composite cuboidal light ray projection system. The entire composite optical system consists of a polygonal composite light ray deviating system and a composite cuboidal light ray projection system. And in the entire composite optical system, the polygonal composite light ray deviating system will be positioned on top of the composite cuboidal light ray projection system. The polygonal composite light ray deviating system will consists of a collection of individual lens elements that are collectively configured in an adjacent manner. In addition to …

Gci #301 — Ariadne Heart-Shield (V5.3), Michael Victor Caldwell Mr.

Defensive Publications Series

This disclosure presents a passive vascular interface device designed to condition hemodynamic and biochemical boundary conditions at vascular anastomoses. The system utilizes controlled multi-helical flow guidance, compliance-matched materials, and anti-fouling nitric oxide-active surfaces to reduce disturbed flow, leukocyte adhesion probability, and thrombotic activation at the graft junction.

The device does not prevent immune recognition but instead modifies the local microenvironment to reduce the physical and biochemical conditions required for inflammatory amplification. A full preclinical validation pathway is defined, including computational modeling, benchtop testing, ex vivo validation, and large-animal studies.