Digital Commons Network^™

Crowdsourced Biometric Synchronization For Video Intensity Mapping And Benchmarking, Alan She

Defensive Publications Series

This paper describes a biometric-to-video synchronization engine that pairs wearable physiological sensors with a distributed video platform for crowdsourced intensity mapping and normalized effort-based benchmarking. Fitness video content on open platforms typically relies on subjective difficulty labels assigned by creators, which fail to account for the physiological diversity of viewers. The synchronization engine establishes a temporal sync anchor between a wearable device's biometric data stream and the video playback timeline, employing clock-drift compensation to align UTC-stamped biometric packets with the relative playback head and a latency correction algorithm to account for physiological response delay. As anonymized, normalized heart rate data …

Dynamic Information-Density-Aware (Dida) Video Playback Scaling, Alan She

Defensive Publications Series

This paper describes a Dynamic Information-Density-Aware (DIDA) Playback System, which automatically modulates video speed on a per-segment basis according to measured content density. Video content exhibits asymmetric information density, where some segments contain dense instructional material while others contain filler, repetitive transitions, or silence. Static playback speed forces viewers to choose between missing details during information spikes in dense segments or enduring disengagement during sparse ones. DIDA uses a Smart Speed engine combining visual frame classification, audio speech-rate measurement, and silence detection with crowdsourced viewer engagement data to compute an optimal playback rate for each timestamp.

An Algorithmic Density Scorer …

Dynamic Viewport Slicing And Client-Side Compositing For Bandwidth-Optimized Live Video Broadcasting, Nikita Agarwal

Defensive Publications Series

This paper describes a dual-stream video delivery system for live broadcasting that reduces transmitted bandwidth by encoding only the region of active motion at full resolution. In typical live sports broadcasts, the area of viewer interest occupies only 15-20% of the video frame, while the surrounding background consumes bandwidth at uniform quality. The described system repurposes raw motion vectors generated during the AOMedia Video 1 (AV1) encoding pass to compute an action centroid coordinate, bypassing neural network-based analysis.

The encoder splits the broadcast into two concurrent layers: a full-frame background aggressively compressed to low resolution and a bounding-box action crop …

Investigation Of A Shielded Magnetic Switching Architecture And Magnetic Energy Transfer Mechanisms, Peter Branton

Defensive Publications Series

This paper presents a magnetic switching architecture based upon the controlled exposure and shielding of permanent magnetic fields.

The system combines:

• Permanent magnets

• Movable magnetic shielding

• Attraction phases

• Repulsion phases

• Force-balanced switching mechanisms

• Mechanical momentum recovery

The architecture was developed through the progressive elimination, reduction, or counterbalancing of apparent energy costs commonly associated with magnetic switching systems.

The resulting design raises a fundamental question:

Can the complete behavior of the disclosed system be fully explained using conventional magnetic energy accounting, or does the geometry reveal aspects of magnetic energy transfer that warrant further investigation?

This document presents a specific engineering …

Optical Circuit Switches For Grouped Port Switching, N/A

Defensive Publications Series

Keywords: optical circuit switch, grouped switching, liquid crystal, polarization, optical networking, data center, photonic switching, bulk switching, parallel switching

Abstract

Optical circuit switches are a vital part of datacenter networking and machine learning infrastructure because of their ability to switch large amounts of bandwidth in an energy efficient, low latency, and data/format signaling agnostic manner. However, common approaches for switching groups of ports, such as parallelizing multiple optical circuit switches or optically multiplexing signals, encounter scalability limitations as bandwidth demands increase and per-fiber bandwidth saturates. To address these challenges, this publication describes a system and method for switching groups of …

System And Method For User-Owned Longitudinal Memory And Interpretation System With A Provenance-Tagged Interpretive Ledger, Engine-Interchangeable Inference, Federated Validation, And Generational Maintenance, Joseph Jm Walker

Defensive Publications Series

Contemporary personal-AI systems preserve continuity in the weakest possible way: the provider owns the model, the memory, and the interface, and the user retains accumulated context only so long as the service remains available and willing to keep it. When a model changes, a platform resets memory, or a provider alters policy, the user's interpretive history can vanish. This disclosure presents the Continuity Node Framework, a user-owned, local-first, longitudinal memory-and-interpretation architecture in which the durable asset is not the AI model but the preserved, provenance-tagged map of a person's reasoning, values, patterns, and tested interpretations over time, and in which …

Autonomous Infrastructure Orchestration Via Natural Language Narratives, N/A

Defensive Publications Series

Keywords: Cloud Computing, Infrastructure as Code, Multi-Agent Systems, Natural Language Processing, Autonomous Orchestration, Shadow Infrastructure, Canonical Documentation.

Abstract

Managing and modifying infrastructure, especially when systems have been manually configured or have drifted from their original Infrastructure-as-Code (IaC) definitions, is complex, risky, and expensive. This situation can result in a fragmented understanding of the infrastructure's actual state, a loss of institutional knowledge, and increased manual operational burden, leading to fragility and availability risks. To address this, a multi-agent system implements an Infrastructure-as-Narrative (IaN) paradigm. Autonomous agents discover the current infrastructure and document its state in a high-level, human-readable Canonical Infrastructure Document …

Adhesive-Sealed Two-Part Cold Plate Assembly With Narrow Flange, N/A, N/A

Defensive Publications Series

A cold plate assembly uses an adhesive sealant to create a leak-proof seal between a cold plate base and a top cover. The sealant adheres to the mating surfaces of the two components and eliminates the need for an O-ring and its associated compression screws. Removing the mechanical fasteners reduces the perimeter flange width from a typical 14-17 mm to approximately 2-3 mm. This assembly approach avoids the high temperatures of brazing, preventing thermal damage or weakening of the cold plate materials. The assembly maintains a two-part design, which facilitates manufacturing and inspection workflows. In one configuration, the cold plate …

Blade-Integrated Triply Periodic Minimal Surface Micro-Cooling Distribution Unit For High-Density Data Centers, N/A, N/A

Defensive Publications Series

A decentralized liquid cooling architecture for high-density data center racks integrates a micro-cooling distribution unit (mCDU) directly onto each server blade or chassis wall. The mCDU, an additively manufactured monolithic block of titanium or aluminum, replaces centralized rack-level heat exchangers and their associated plumbing. Its internal structure uses a triply periodic minimal surface (TPMS) lattice, such as a Schwarz-P or Schwarz-D geometry, to create two intertwined, co-continuous fluid paths. One path circulates hot coolant from the server components while the other circulates cold facility water. This TPMS geometry has a large surface-area-to-volume ratio for thermal exchange and results in low …

Detachable Stator Support For Gas Turbine, Anonymous

Defensive Publications Series

Compressor casings are machined with slots for every compressor stage to facilitate insertion of compressor stator blades. In the same way, turbine casings are machined with the slots for assembling the required hot gas path parts. These slots are machined with the hook interface at forward and aft of every slot, making the manufacturing very complex and time intensive.

Programmable Error Severity Setting And Firmware Status Reporting For Testing Accelerator Platforms, N/A, N/A

Defensive Publications Series

A system and method are described for a statistical testing and analysis framework configured for computational accelerators. The framework facilitates programmable error severity settings, firmware status reporting, and configuration percentage rollouts. Techniques include using a management controller to append firmware versions to error logs, employing error mask maps provided by a user-space driver to mask firmware-level errors, and utilizing configurable error severity levels to manage how exceptions are handled without stopping active machine learning workloads. Keywords: computational accelerator, firmware testing, error masking, configuration rollout, automated analytics.

Evaluating Agent Cognitive Resilience Using Controlled Environmental Inconsistencies, Isaac Leimgruber, Nicolas Zimmermann

Defensive Publications Series

Conventional evaluation methods for autonomous agents may prioritize task completion without assessing an agent's ability to handle systemic discrepancies between its instructions and its operational environment. A dynamic evaluation framework is described that can assess an agent's cognitive resilience. The framework may operate by introducing controlled inconsistencies, for example, planted traps, into a sandboxed environment, which can create a conflict between the agent's instructional data and the runtime reality. The system can then evaluate the agent's response, focusing on its ability to detect the discrepancy, classify its root cause, and propose a structural resolution rather than focusing on task completion. …

Predictive Hardware State Mitigation Using Telemetry And Feedback, Mohsen Heidarinejad, Farid Kenarangi, Arpit Mittal, Amin Khajeh

Defensive Publications Series

Varying computational tasks introduce thermal and power dependencies within modern computing devices. Static threshold policies might not fully capture cumulative operational stress, potentially causing over-throttling or delayed hardware responses. A closed-loop machine learning framework is disclosed that dynamically evaluates multidimensional telemetry data to compute a predictive risk score. A feature selector extracts direct measurements, system activity indicators, temporal features, and health indicators. Based on the calculated risk score, an arbitrator dynamically applies mitigation actions (e.g., scaling component voltage and frequency states). A model corrector operates during hardware idle periods to iteratively refine predictive modeling using observed performance metrics. The disclosed …

Fused Arithmetic Datapath For Bit‑Slicing Multiply‑And‑Accumulate Units, Dongseok Im, Hee Jun Park

Defensive Publications Series

Many modern processing systems utilizing bit‑scalable architectures may experience performance limitations due to dynamic power consumption from heavy integer arithmetic operations. Executing higher‑precision quantized matrix multiplications on lower‑precision multiply‑and‑accumulate (MAC) hardware typically requires calculating multiple independent partial products, which may lower computational efficiency. The disclosed processing system derives cross‑product terms simultaneously through a fused mathematical operation. Specifically, it uses pre‑multiplication addition on decomposed operand segments and executes a post‑multiplication subtraction to isolate the cross‑products. This approach reduces the total number of hardware multiplication operations utilized per calculation. By reducing these redundant steps, the processing system may lower dynamic power consumption, …

Generative Media Repair For Uninterrupted Transcoding With An Asynchronous Approval, Nikita Agarwal

Defensive Publications Series

The process of ingesting large-scale media files may be interrupted by data corruption, which can necessitate a re-upload and re-transcode, consuming network and computational resources. Described systems and methods can address this by detecting corruption during ingestion, generating a provisional synthetic patch for the corrupted portion, and allowing the transcoding process to proceed. A cryptographic manifest can be created to document the change, and the provisionally repaired asset may be held in a quarantined state. An asynchronous approval workflow can then allow a content partner to review, approve, or reject the patch, with a fallback for providing a corrected segment. …

Cross-Layer Detection, Quarantine, And Safe Reclamation Of Stale Communication State In Ai Fabrics, Pradeep Bhagya Govindraj

Defensive Publications Series

Artificial intelligence (AI) fabrics for distributed training and inference often retain stale communication state after a job, worker, endpoint, or session has failed, completed, or stopped making progress. This stale state can waste resources, preserve bad path choices, and reduce accelerator utilization. Proposed herein are techniques to facilitate lifecycle-aware management of communication state in AI fabrics. The proposed techniques use cross-layer telemetry to classify communication state as active, dormant, degraded, or orphaned, then safely quarantine and reclaim only the state that is truly invalid. Unlike generic telemetry analytics or rerouting, the proposed techniques add workload-aware validation, post-remediation verification, and rollback …

Fusing Uwb Spatial Data And Imu Orientation For Deterministic Device Targeting, Nikita Agarwal

Defensive Publications Series

Systems for routing interactive content, for example, from a primary display to a specific user's device in a multi-user environment, can present challenges in identifying the intended recipient. This disclosure describes systems and methods that can combine spatial data with orientation data to help ascertain user intent. For example, spatial positioning of devices may be determined using ultra-wideband (UWB) radio signals, while an inertial measurement unit in a handheld control device (e.g., a remote control, smartphone, or wearable device) can track its orientation. By fusing these data streams, a directional vector can be projected from the control device. A target …

Ai Output Enrichment System With Data-Shape-Aware Routing, Enrich Directive Protocol, And Context-Aware Block Absorption, Bhautik Kishorchandra Doshi, Mark Goho, Zac Lovoy, Juan Carlos Pontaza, Arpita Mehta

Defensive Publications Series

Current artificial intelligence (AI) systems can generate semantically relevant markdown responses, but they do not reliably generate high-quality interactive user interfaces. Techniques described herein address that gap through a two-layer downstream enrichment system that transforms arbitrary AI-agent markdown into rich user interface components without modifying the originating model.  Layer 1 is a deterministic, data-shape-aware pipeline that converts markdown blocks into charts, Key Performance Indicator (KPI) cards, tables, and titled sections using specific routing and absorption heuristics.  Layer 2 is an optional AI presentation layer that enhances the structured output through semantic highlighting, chart refinement, thread-context-aware adaptation, and layout optimization, without …

Agentic Graph Querying On A Sharded Architecture Using A Large Language Model, Sai Charan Tej Kommuri

Defensive Publications Series

Using large language models (LLMs) to query large, scale, graph, structured data can present challenges, as encoding an entire graph into a single prompt may exceed context window limits and can lead to reduced accuracy on multi-hop queries. Systems and methods can partition a graph into a collection of smaller, human-readable documents, or shards. An agentic orchestration process may use a global symbolic index, contained within a summary shard, to locate and retrieve relevant shards on demand. This process can decompose a complex query into a sequence of single-hop LLM calls, where each call can operate on a context comprising …

The Hybrid Path Architecture For Live Application Telemetry And Control, Stephen Norum, Christopher Burns, Leila Al-Muhtasib

Defensive Publications Series

Standard tools for analyzing software often slow down the programs they are testing and create data logjams, making it hard to catch fast-moving glitches. This system solves that problem by splitting communication into different "paths." It creates a high-speed lane for data to flow directly from the app to a dashboard, skipping the usual middleman tools that cause delays. Developers can also add new diagnostic tools to an app while it is already running without needing to change the code or restart it. This allows for real-time monitoring and fixing of apps with much less impact on their performance.

Dynamic Modification Of Digital Content Based On Real-Time Environmental And Geo-Temporal Signals, Indu Ramamurthi, Aravind Balakrishnan

Defensive Publications Series

The present disclosure relates to dynamically modifying digital content, such as digital advertising and e-commerce recommendations, based on real-time environmental and geo-temporal signals. Systems and methods are provided for utilizing an asynchronous environmental caching layer integrated with generative artificial intelligence to mutate product metadata during live content serving auctions. By mapping a user's local environmental conditions against a product's attribute graph, the system suppresses irrelevant product features and elevates contextually relevant attributes. This enables content adaptation to the immediate physical reality of a user, enhancing contextual relevance and user engagement. Keywords: dynamic content generation, generative AI, environmental signals, geo-temporal data, …

Context-Aware Generative Differentiation Of Digital Content Slates, Indu Ramamurthi, Siva Kumar Gorantla

Defensive Publications Series

A system and method for generating differentiated digital content slates are described. The technology utilizes a multi-pass, context-aware architecture to analyze and modify the text of multiple items presented together on a user interface. By sequentially rewriting product descriptions or titles based on the semantic overlap of the group, the system maximizes information entropy and provides users with distinct attributes for each item. This approach mitigates cognitive fatigue and semantic redundancy often associated with generic point-wise content generation. Keywords: Generative AI, Context-Aware Architecture, Content Slate, Semantic Differentiation, Dynamic Rewriting, Information Entropy.

Minimizing The Gap Between Church And Disability In Churches Of Christ, Paris Freeman

Discernment: Theology and the Practice of Ministry

Church members with intellectual and developmental disabilities often experience disconnection from their churches. This qualitative study examined this disconnect in three suburban Church of Christ congregations within the North Texas area. Using an ethnographic approach, data was collected through interviews with ministry leaders, participatory observation, and a literature review. The analysis of the data, in conjunction with family systems theory, was utilized to construct a model for disability ministry which validates personhood and promotes awareness of self and others. The impact of the church growth movement on disability was also examined, but no definitive correlation could be determined. Further research …

Transmission Of Firm-Level Demand Shocks To Stock Returns: Evidence From 10-K Text Analysis, Dylan Gunjan Parikh

Economics Undergraduate Senior Theses

Industrial organization and corporate finance papers that measure firm-level demand characteristics employ price, quantity, and product characteristic data that is unavailable in public financial disclosures. As a result, cross-sectional asset pricing literature attention to demand-side factors is limited. This paper constructs a demand shock measure from natural language processing of 28,263 publicly disclosed 10-K filings from 2001 to 2020. I find that a 1SD demand shock increases contemporaneously predicts a 1.8% increase (SE = 0.4%) in returns. Long-short portfolio sorted on demand shock generates 8.5% alpha (SE = 1.9%) against combined Fama-French and Q-factor models and a 0.91 sharpe ratio …

Identifying Active Layer Dynamics Across Geomorphic Landforms In Longyeardalen, Central Svalbard, Madeline Kim Koelbel

Earth Sciences Undergraduate Senior Theses

The thermal state of the active layer and analysis of meteorological conditions provides a greater understanding of where active layer detachment slides can occur. This study provides an overview of the active layer ground thermal regime for 2024 and 2025 in Longyeardalen, Svalbard, where the largest settlement in Svalbard, Longyearbyen is located. Ground temperature data was extracted from 8 boreholes in Longyeardalen and meteorological data was extracted from one weather station in central Longyeardalen. Overall, the ALT thickness and the thawing period increased from 2024 to 2025 while the thaw rate remained the same or decreased. The zero-curtain effect was …

Fisheye Camera Object Detection For Traffic Surveillance, Hugo Yang, Robert Luo

Computer Science and Engineering Senior Theses

Effective traffic management is hampered by the physical limitations of traditional pinhole cameras, whose narrow fields of view create blind spots that require costly multi-camera infrastructure to overcome. Fisheye cameras offer a wide-angle alternative capable of single-camera intersection coverage, but the radial distortion inherent in their optics renders standard object detection algorithms unreliable, limiting their practical deployment in traffic surveillance.

This thesis investigates specialized object detection pipelines for fisheye traffic imagery by reproducing and building upon two top-ranked submissions from the 2025 AI City Challenge Track 4: the VNPT and SKKU-AutoLab pipelines. Both were successfully reconstructed outside their original containerized …

Interview With Rosa Elia Garza, Rosa Elia Garza, Logan Dovalina

Rio Grande Valley Oral Histories

Rosa Elia Garza (née Peña) was born on August 19, 1947, in Arcabuz, Tamaulipas, Mexico, and immigrated to the United States at four months of age with her parents, Porfirio Peña and Elodia González Peña. The youngest of eight sisters in a family with no sons, she grew up in Monte Alto, Texas, where her family built a life centered on hard work, family unity, and the pursuit of a better future after selling their property in Mexico. From age six, Rosa worked in the cotton fields alongside her mother and sisters, eventually traveling seasonally with other Valley families in …

Open Versus Locked Unit To Safely Foster Dementia Wandering, Marie Cummings Rn

Nurse Residency

How would the inclusion of a locked medical-surgical unit for patients with dementia who exhibit wandering behavior prevent elopement, reduce involuntary holds and restraint use, and high rates behavioral codes that are experienced on open units such as Mid Coast Hospital’s medical-surgical unit?

Plugging The Digital Reproductive Health Data Privacy Holes In The Post-Dobbs Era—A New Evidentiary Privilege, Fengting Liang

NULR Online

No abstract provided.

Long-Term Demography And Spatial Genetic Structure Reveal Mechanisms Of Sassafras Albidum Population Persistence Through Clonality, J.T. Michel, Jonathan P. Evans, Skyler J. Fox, Ashley B. Morris

Open Access Fund Publications

Premise: Vegetative regeneration is a key mechanism of woody plant persistence in forest ecosystems, and the coupled roles of basal sprouting and clonal growth in shaping long-term population dynamics have been understudied. Basal sprouting replaces stems at fixed ramet locations, whereas clonal growth via root suckering produces spatially distinct ramets that promote genet-level persistence across space. Sassafras albidum, a clonal understory tree native to eastern North America, employs both mechanisms, making it an excellent system for evaluating how sprouting and clonality interact to structure populations.

Methods: We combined over two decades of demographic data with spatially explicit microsatellite analyses …