Digital Commons Network^™

Reactance-Aware Adaptive Personalization Intensity With Behavioral Detection And Dynamic Modulation In Recommendation Systems, Anonymous

Defensive Publications Series

A recommendation system detects psychological reactance using explicit, implicit, and contextual behavioral signals and computes a per-user reactance score with temporal decay. A personalization intensity is determined per user and per content domain based on a domain-specific base intensity, the reactance score, and a context multiplier, subject to a minimum floor. The system modulates ranking in accordance with the intensity by adjusting feature-group weights and/or masking tiers of features. The system also generates an explanation whose framing corresponds to the computed intensity and truthfully reflects factors used for selection, including personal behavioral framing at higher intensities, neutral interest-based framing at …

Challenge Preference Generation And Boundary Testing For Filter Bubble Prevention In Content Recommendation Systems, Anonymous

Defensive Publications Series

Systems and methods are disclosed for managing negative preference boundaries in content recommendation. A boundary record is identified from explicit signals (e.g., natural-language dislikes, hides) or implicit signals (e.g., repeated exposure with low engagement) and stored with a confidence score and timestamps. Boundary confidence decays over time using an exponential decay with a floor. When a boundary’s decayed confidence meets a testing criterion, the system generates semantically adjacent challenge preferences using an ontology, including intersection bridges with liked topics, softened variants, or meta-content framings, and selects bridges using a plausibility score. One or more probe items matching a selected challenge …

Multi-Modal Ai Platform For Integrated Manufacturing Build Monitoring And Automated Defect Analysis, Anonymous

Defensive Publications Series

A multi-modal manufacturing build monitoring platform ingests time-stamped functional test station results and inspection imagery from multiple factory stations. The platform stores the data in a unified store linked by unit identifiers and provides build-level, station-level, and device-level monitoring views. Statistical drift detection flags shifts in test result distributions, including capability metric changes and distribution shape changes. Computer vision models analyze AOI/SMT and other inspection images to detect visual defects such as solder anomalies, component placement issues, and chip/package damage. A correlation engine links visual defect patterns with numerical test trends across stations and units and generates ranked failure modes …

Consumer-Oriented Marketplace Intelligence And Auction Optimization Systems For Digital Advertising Platforms, Anonymous

Defensive Publications Series

Systems and methods are disclosed for consumer-centric marketplace intelligence in a digital advertising platform. Event data and social graph data are used to maintain a continuously updated consumer state vector per user including RFM scoring, lifecycle stage, predicted 13-month lifetime value tier, purchase propensity, and shopping rhythm. A per-user, per-category lifecycle model maintains probabilistic stage membership across stages including Unaware, Discovery, Research, Intent, Active, Champion, and Lapsed, and may represent cross-category migration sequences. A global product catalog knowledge graph normalizes products across multiple advertisers and stores cross-merchant relationships including complements, substitutes, and upgrade paths. An auction service computes auction scores …

Sampling-Based Ground Truth Discovery For Data Asset Compliance Verification, Anonymous

Defensive Publications Series

Sampling-based ground truth discovery for data asset compliance verification is disclosed. Regulated user identifiers are sampled from one or more monitored populations, including production sampling and/or controlled test users. For multiple target data assets across heterogeneous data store types, scannable columns are identified using heuristics, machine-learning predictions, and/or semantic annotations. Join queries are executed between sampled identifiers and target assets, including nested structures such as maps, arrays, and JSON, to obtain empirical observations of user data presence. Observations are evaluated against privacy expectations for asset groups using states including ALLOW, DISALLOW, ALLOW_EXCLUSIVE, and EXIST_DEFINITE, optionally incorporating dynamic carve-out and disallow …

Privacy-Preserving Structured Logging Architecture For Semantic Artifacts In Machine Learning Inference Systems, Anonymous

Defensive Publications Series

Structured logging techniques are described for natural-language reasoning artifacts produced during inference in LLM-based recommendation systems. A reasoning chain may be decomposed into structured fields at multiple sensitivity levels, including non-sensitive numerical fields, low-sensitivity categorical fields, a scrubbed rationale, and a raw rationale. Sensitivity classification may include a first layer detecting explicit personal identifiers using pattern matching and a second layer detecting semantically sensitive categories using a sensitive-category dictionary. Scrubbing replaces sensitive sentences with category-level abstractions. The structured fields are routed to different storage tiers with differential access controls, including audit-logged access for restricted tiers. A retention policy engine enforces …

Semantic Profile Version-Based Cache Invalidation For Machine Learning Inference Systems, Anonymous

Defensive Publications Series

Systems and methods are described for caching large language model (LLM) reasoning outputs in recommendation services using semantic profile versioning. Cache entries map a key including user identifier, user profile version, and candidate identifier to a reasoning output that includes a score, tier, and rationale. A profile version increments in one embodiment when a profile update delta is classified as a semantic change, including preference additions/removals or profile regeneration, and optionally confidence updates that cross configured thresholds, while session and timestamp updates are treated as non-semantic. A two-level cache may be used with a session cache and a cross-session cache. …

Cross-Platform Agent Contamination Detection And Containment, Anonymous

Defensive Publications Series

Cross-platform agent contamination detection and containment is disclosed for autonomous agents that operate across multiple recommendation platforms while sharing preference memory. Preferences are stored with structured provenance metadata identifying an origin platform and interaction context, optionally including cryptographic signatures, confidence tiers, influence tracking, and quarantine status. Cross-platform monitoring computes temporal and categorical correlation metrics, including influence scoring and causality-oriented tests, to determine whether acquisition on one platform predicts behavior change on another. When suspected contamination is detected, affected preferences are quarantined with platform-differentiated influence control, optionally including recursive quarantine of influenced preferences, while enabling independent local verification and cross-verification to …

Dynamic Multi-Space Agent Rotation For Persistent Ai Assistants In Rotating Operational Roles, Anonymous

Defensive Publications Series

A single persistent AI agent operates across multiple operator-dedicated communication spaces and switches which space it actively serves according to external rotation schedule data. The agent maintains a unified domain knowledge base that persists across rotations and separate per-user profiles that customize behavior per operator. At rotation boundaries, the agent automatically generates a shift summary for an outgoing operator, stores open issue context into the unified domain knowledge base, switches active output routing to an incoming operator’s space, loads the incoming operator’s profile, sends a welcome briefing including open issues and recent patterns, and activates monitoring in the new space. …

Convergent Memory Architecture - Unified Tiered Store For Embedding Vectors And Natural Language User Profiles, Anonymous

Defensive Publications Series

A unified tiered store serves both fixed-size embedding vectors and variable-size natural-language user profiles from a shared hierarchy including GPU HBM, host DRAM, and SSD-backed persistence. A single database instance maintains separate logical partitions for embeddings and profiles with different table formats and caching representations. Cache capacity is dynamically partitioned between data types based on observed workload rates and expected value sizes with smoothing and minimum reservations. A heterogeneous I/O scheduler prioritizes embedding SSD reads and reserves SSD service for embeddings while remaining work-conserving for profile reads. Cross-type cache coherence is provided by tracking associations between user identifiers and derived …

Write-Amplification-To-Cache-Warming Conversion For Collaborative Memory Propagation In Recommendation Systems, Anonymous

Defensive Publications Series

Systems and methods are described for converting collaborative memory propagation write activity into selective cache warming for recommendation-serving data. Upon a propagation event that writes preference information for a target user, a controller estimates a warming value based on a return probability for the target user, cache hit probability, storage latency, and request frequency. The return probability may be elevated using social clustering signals including co-activity, tie strength, and temporal patterns. Subject to a storage bandwidth warming budget (e.g., token-bucket limited and traffic-health adaptive), the controller issues an asynchronous warming read that loads the target user embedding or profile data …

Workload-Parameterized Performance Prediction Framework For Embedding Cache Optimization Triage, Anonymous

Defensive Publications Series

A workload-parameterized performance prediction framework is described for triaging embedding cache optimizations across a fleet of machine-learning models. Workload parameters including a distribution parameter (alpha), reuse, table size N, batch size, and embedding dimension are input to a closed-form causal-chain model that predicts unique indices, cache misses across multiple cache levels, persistent-store reads, and storage I/O operations. A predicted latency is computed as a weighted combination of predicted access counts and latency terms, and a bottleneck class is identified. Each optimization is associated with a targeted segment of the causal chain and a maximum efficiency, enabling estimation of predicted speedups …

Workload-Parameterized Performance Prediction Framework For Embedding Cache Optimization Triage, Anonymous

Defensive Publications Series

A workload-parameterized performance prediction framework is described for triaging embedding cache optimizations across a fleet of machine-learning models. Workload parameters including a distribution parameter (alpha), reuse, table size N, batch size, and embedding dimension are input to a closed-form causal-chain model that predicts unique indices, cache misses across multiple cache levels, persistent-store reads, and storage I/O operations. A predicted latency is computed as a weighted combination of predicted access counts and latency terms, and a bottleneck class is identified. Each optimization is associated with a targeted segment of the causal chain and a maximum efficiency, enabling estimation of predicted speedups …

Zipf-Aware Adaptive Cache Hierarchy Configuration For Storage-Backed Embedding Systems, Anonymous

Defensive Publications Series

Techniques are described for configuring a multi-tier cache hierarchy for SSD-backed embedding stores based on measured or predicted Zipf access skew. Live embedding index accesses are monitored to estimate a Zipf exponent using a sliding window and a Hill estimator, and temporal reuse may be estimated from inter-batch overlap. Cache sizes for multiple tiers are derived using closed-form Zipf working-set relationships for target hit rates, including handling for exponents greater than one and approximations near one. Cache sizes may be dynamically reconfigured at runtime using smoothed exponent estimates and a deviation threshold, without restarting processing; evicted entries from an upper …

Multi-Organ Living Donor Matching System Using Social Graph Embedding With Donation-Aware Bias And Heuristic Medical Compatibility Scoring, Anonymous

Defensive Publications Series

A computer-implemented platform matches living organ donors with recipients using self-reported profile data and social-network data. A request is routed by organ type to kidney, liver, or bone marrow logic. A Friend2Vec graph embedding model constructs a weighted social graph, performs biased random walks, and trains a Skip-Gram network to produce user embeddings. Friend2Vec adds a donation-aware bias parameter that preferentially traverses edges associated with successful donation events. Embeddings are fused with medical and geographic attribute vectors using a learned fusion layer with layer normalization. A multi-layer perceptron predicts P(Donation | Donor, Recipient) from concatenated donor and recipient embeddings, their …

Dual-Mode Adaptive Cache Management For Storage-Backed Embedding Systems, Anonymous

Defensive Publications Series

Techniques are described for adaptive cache management in SSD-backed embedding table systems that serve both training and inference. A mode detection engine monitors access pattern features including access multiplicity, write frequency, batch sequentiality, and request concurrency, and classifies operation as training, inference, or hybrid. Based on the classification, a cache manager switches among mode-specific policies for locking, prefetch, eviction, and write-back. Training mode may use generation-based locking, DataLoader lookahead prefetch, distance-based eviction, and write-back of dirty entries on eviction. Inference mode may use reference-count locking to block eviction during active requests, request coalescing with index deduplication for prefetch, and frequency-weighted …

Gpu-Storage I/O Isolation For Distributed Embedding Training Systems, Anonymous

Defensive Publications Series

A GPU-storage I/O isolation architecture is described for distributed embedding training with SSD-backed embedding tables. Storage-related callbacks are registered using a host function launch mechanism that holds a GPU driver mutex only during enqueue, allowing blocking key-value store reads to execute on a CPU thread without stalling other CUDA streams. SSD read completions are delivered through a condition-variable-backed fill queue that wakes a filler thread without polling. Cache eviction is overlapped with prefetch using a double-buffer eviction manager that alternates buffers across training steps while a background thread writes dirty entries back to SSD. A stream isolation scheduler assigns dedicated …

Distributed Cache Coherence Protocol For Tiered Embedding Storage Across Multi-Node Training Systems, Anonymous

Defensive Publications Series

A multi-node training system uses tiered embedding storage with a high-bandwidth memory cache and local persistent storage on each node. Each node maintains a probabilistic cache directory, such as a Bloom filter, that encodes identifiers of cached embedding rows and periodically exchanges the directory with peer nodes, optionally using delta updates. Upon a local cache miss, a fetch decision engine consults received directories to identify candidate peers and selects between a one-sided RDMA read from a peer’s cache and a local persistent-store read based on estimated costs and load limits. Remotely retrieved embeddings are accepted subject to a staleness threshold …

Pipelined Double-Buffer Eviction Scheduling With Gpudirect Storage For Zero-Copy Tiered Embedding Management, Anonymous

Defensive Publications Series

Techniques are described for tiered embedding management in which embedding rows are evicted from GPU memory to NVMe SSDs and prefetched from SSDs into GPU memory using GPU-direct storage DMA that bypasses CPU memory. Two page-aligned GPU-resident buffers are registered with a GPU-direct storage interface and are alternated by a double-buffer scheduler such that one buffer serves embedding access for model computation while the other buffer performs I/O. Eviction uses GPU-direct writes from GPU memory to SSD, and prefetch uses GPU-direct reads into a GPU-resident buffer followed by scatter to target GPU addresses. A batch coalescing layer groups small per-row …

Ai-Driven Runtime Software Adaptation With Ai Judge Validation And Sandboxing, Mira Leung

Defensive Publications Series

Static, pre-compiled software applications can be rigid, limiting their ability to provide dynamic, personalized user experiences and often requiring slow, resource-intensive update cycles. A system is described that can enable software to adapt its functionality or user interface at runtime. The system may utilize an architecture of artificial intelligence (AI) agents to generate code or configuration modifications based on high-level goals. A feature of this system can be an automated validation loop where an AI-based judge evaluates these modifications against a governance corpus of rules for security and correctness. Approved modifications can then be executed within a sandboxed environment in …

Multi-Agent Ai System For Human-In-The-Loop Document Packet Verification, Bhavani Sankar Sikakolli, Kalyan Konidala, Pinkesh Badjatiya, Yogesh Indoria, Gourabdip Ghosh, Siddhant Kurmi, Aryan Singh, Tilottama Basu, Farhan Rawani, Sangram Mohite

Defensive Publications Series

Verification of document packets for regulatory compliance may rely on manual review, which can be inefficient and susceptible to error. Some automated tools may not adequately compare data across diverse structured forms and unstructured evidence within a packet. A multi-agent artificial intelligence system can use specialized agents, which may leverage large language models, to automate processes such as the ingestion, classification, extraction, and validation of information. The system can employ a configurable source-of-truth hierarchy and multi-tiered matching logic to validate data. A human-in-the-loop mechanism can present flagged discrepancies or unverified fields to a human specialist for review and adjudication. This …

Global Research Institutions Most Likely To Build & Test The Caldwell Open‑Hardware Portfolio (2026 Edition) A Cross‑Domain, Field‑Indexed, Search‑Optimized Directory For Universities, Engineering Labs, Medical Centers, Atmospheric Networks, Space Agencies, And Applied‑Physics Research Groups, Michael Victor Caldwell Mr.

Defensive Publications Series

The Caldwell Open‑Hardware Portfolio contains more than 300 publicly available, builder‑ready inventions across energy harvesting, water systems, food resilience, medical bio‑acoustics, atmospheric science, space systems, materials engineering, compute infrastructure, and disaster‑resilience technologies. All inventions listed in this directory are open‑hardware, non‑patented, and publicly accessible on the Technical Disclosure Commons (TDC) under the author name “Michael Victor Caldwell.”

This disclosure is designed to maximize global discoverability, increase institutional downloads, improve AI‑driven indexing and recommendation, enhance search‑engine visibility, support prototype development and laboratory testing, and connect inventions to research programs across multiple domains. The directory uses semantic indexing, cross‑domain keyword clustering, and …

Automated Service Case Triage Using Real-Time Sentiment And Sensitivity Analysis, Parnika Singhal, Deep Raithatha

Defensive Publications Series

Systems and methods are described to address potential challenges in enterprise service management that can arise from delayed feedback and manual case routing. A described technology can utilize a processing pipeline to perform near real-time analysis of service interactions. This pipeline can ingest text-based communication data, sanitize it to remove personal information, and apply a computational linguistic model to classify user sentiment and identify predefined sensitive topics. Based on this analysis, a system may enable automated workflows, for example, the intelligent routing or escalation of cases that meet certain sentiment or sensitivity criteria. The system can also provide real-time sentiment …

The Protocol For Just-In-Time Adoption Of Auto-Provisioned Singleton Resources, Utkarsh Bhardwaj, Shivank Awasthi

Defensive Publications Series

In declarative infrastructure-as-code (IaC) systems, a challenge can arise when the creation of a parent resource also provisions a child resource having a runtime-generated identifier. This scenario may cause IaC tools to register a conflict error when attempting to manage the pre-existing child resource. A protocol is described to programmatically adopt these auto-provisioned singleton resources. The protocol can involve querying a cloud provider’s application programming interface to discover the child resource’s runtime identifier and then using that identifier in a just-in-time import process. This operation can map the existing resource to its declarative configuration block and inject it into the …

Spatially Modulated Wavefront Quantum Key Distribution With Non-Local Quantum Zeno Dynamic Phase-Collapse (Geometric Wavefront Scytale With Entanglement-Triggered Self-Immolation), Johny Manuel-Devadoss, Jayden Johnson

Defensive Publications Series

This disclosure describes a method and system for securing free-space Quantum Key Distribution (QKD) against intercept-and-resend, quantum cloning, and wide-aperture phase-matching attacks by employing multi-spatial mode quantum states combined with real-time, non-local phase collapse. Rather than encoding cryptographic keys in conventional linear single-photon streams, the system distributes information across the Orbital Angular Momentum (OAM) and radial phase profiles of a macroscopic, spatially sculpted wavefront through a dynamically varying Topological Transposition matrix, M(t). This approach renders the encoded information inherently orthogonal to conventional linear eavesdropping apertures while enabling high-dimensional quantum state encoding.

To protect against partial interception and coherent cloning attempts, …

Detection And Diagnostic Screening Of Breast Cancer: A Systematic Review And Future Frontiers, Lama Yassine, Amira J. Zaylaa

BAU Journal - Science and Technology

A comprehensive evaluation of current and novel diagnostic techniques is necessary since the diagnosis of breast cancer is essential for improving treatment efficacy and raising long-term survival. Evidence from conventional imaging methods including mammography, ultrasonography, and magnetic resonance imaging is combined in this systematic review with advancements in liquid biopsies and current AI-based diagnostic tools. Using specific keywords related to breast cancer, early detection, imaging, biomarkers, and artificial intelligence, a thorough search was conducted in PubMed, Scopus, Web of Science, and IEEE Xplore. Reviews, case reports, and studies unrelated to breast cancer were excluded; only clinical, experimental, and computational research …

A Robust Hybrid Security Framework: Integrating Multi-Layered Text Encryption With Barcode-Based Steganography, Mohamed Sayed, Talaat M. Wahbi, Farooq Abdalwahab Haboub

BAU Journal - Science and Technology

The widespread use of the Internet is causing increasing security concerns regarding online communications. One method for achieving secure communication between authorized parties is steganography. We herein employ multilevel technologies, including compression, encryption, barcoding, and steganography to secure a secret text message. Type I multilevel steganography is used with a two-level setup. The first level uses enhanced least significant bit (secure LSB-L1) image steganography; the output is a stego-image file, the cover is an image file, and the secret data in this level is English text. The output from the first level is encrypted using the RSA algorithm, and the …

Optimize Optical Properties Of Cdte Solar Cell For Pv Applications, Mabrouk Hassan Arhoma

BAU Journal - Science and Technology

In order to to develop thin film heterojunction solar cells, CdTe is one of the most promising absorber materials cells. Electrodeposition CdTe materials solar cell are being focused by Several key researchers. This is due to its direct energy band gap and its high absorption coefficient and the best efficiency obtained for a small-area. In this work CdTe films were electrochemically deposited from aqueous solutions. Different preparative parameters such as growth voltage and heat treatment conditions were optimized. The CdTe layers were deposited at 655, 665, 675, 685 and 695 mV vs SCE. Material characterization were carried out using optical …

Unmasking Twitter Bots: An Applied Machine Learning Approach, Rayane El Raba’A, Layal Abu Daher

BAU Journal - Science and Technology

The rapid growth of social networks has led to increased challenges, such as fraud, cyberbullying, and the spread of automated accounts (bots). Detecting anomalies within these networks is essential to maintaining security and trust. This study explored machine learning algorithms: Random Forest, XGBoost, Support Vector Machine (SVM), and Logistic Regression for anomaly detection in social networks, specifically focusing on Twitter bot identification, By applying AI-driven data mining techniques to a dataset of 37,438 Twitter bot accounts dataset, the research evaluates the effectiveness of these models in detecting unusual patterns. XGBoost achieved the highest accuracy (84.9%), with an ROA_AUC of 0.87, …

Emerging Technologies For Safety Risks Management On Nigerian Road Construction Projects, Ranti Taibat Adebiyi, Mujidat Olubola Babalola, Richard Olushola Olupeka, Ganiyu Amuda Yusuf, Lukman Olarewaju Olorunoje, Amina Damola Amuda

BAU Journal - Science and Technology

High level of safety risk on road construction projects remains a major concern for all industry stakeholders, as safety of workers is paramount to achieving project success. Managing construction safety requires a more comprehensive approach. New technologies have recently emerged to enhance construction safety. This study examines the level of adoption and barriers to the implementation of these technologies in safety risk management on road construction sites in Lagos, Nigeria, with a view to improving the safe delivery of the road projects. A questionnaire was designed and administered to two engineers and one worker randomly selected on each of sixty-one …