Digital Commons Network^™

Risk Manager Predictive Fraud Futures, Gayathri U, Sitaram Panigrahi

Defensive Publications Series

The present disclosure relates to a system and method for predicting merchant-level fraud spikes in payment transaction networks using Temporal Graph Neural Networks (TGNN). The present disclosure suggests receiving historical payment transaction data associated with a plurality of transactions. The present disclosure also suggests generating, from the historical payment transaction data, a plurality of heterogeneous transaction graphs representing transaction relationships across one or more time intervals. Node feature representations and edge feature representations are generated for the heterogeneous transaction graphs based on the transaction data. Subsequently, the present disclosure suggests determining node-specific temporal context information using transaction timestamps associated with …

Root Cause Investigation Memory For Artificial Intelligence (Ai) Agents, Necati Cehreli

Defensive Publications Series

Artificial intelligence (AI) troubleshooting agents can lose root cause analysis (RCA) investigation state when conversation history is compacted or an investigation is interrupted and resumed. To address the challenges of maintaining RCA investigation state, the techniques presented herein implement a memory system that maintains structured RCA memory outside the chat transcript. The memory tracks diagnostic coverage, unexplored evidence, and candidate root causes; refreshes stale state before compaction; and updates the investigation when new evidence is added. A candidate-cause ledger supports final causal adjudication, reducing the likelihood that a downstream event is mistaken for the underlying root cause.

Novel Crystalline Polymorph Of Janus Kinase Inhibitor, Lee Pharma Limited

Defensive Publications Series

The present disclosure relates to {1-(ethyl sulfonyl)-3-[4-(7H-pyrrolo[2,3-d] pyrimidin-4-yl)-1H-pyrazol-1-yl] azetidin-3-yl} acetonitrile, is represented by following structural formula (1) and its polymorph thereof.

User Authorization Of Ai Generated Code Based On Pre-Execution Telemetry, Timo Hoyer

Defensive Publications Series

Generative artificial intelligence (genAI) technology enables users to specify natural language commands and obtain executable code and/or database queries as AI output. While such code can help users accomplish goals quickly, in enterprise computing environments, it can introduce operational issues/vulnerabilities, waste computational resources, lead to platform failures, and/or administrative friction. This disclosure describes a pre-execution orchestration architecture and dynamic user interface for user authorization of code prior to execution. A user can use genAI to generate code. Prior to execution, such code (executable code, database queries, or infrastructural commands) is subjected to analysis to generate pre-execution telemetry that includes factors …

Determinants Of Ai Adoption In Banking: Evidence From Tunisia, Rahma Khattab, Sami Bacha

Arab Economic and Business Journal

This study investigates the factors influencing customers’ intentions to adopt Artificial Intelligence (AI) banking services in Tunisa. Building on an integrated framework combining the Technology Acceptance Model, Unified Theory of Acceptance and Use of Technology, Theory of Planned Behavior, Perceived Risk Theory, and Diffusion of Innovation Theory, the study investigates how perceived usefulness, perceived ease of use (, awareness, knowledge, attitude, subjective norms, trust, and perceived risk influence AI adoption intentions. Using survey data from 400 Tunisian retail banking customers and applying regression analysis, the findings show that perceived usefulness, perceived ease of use, subjective norms, and trust positively and …

Forecasting Demand At Physical Facilities Using Diverse Heterogeneous Real-World Data, Kapil Parakh, Daniel Klein

Defensive Publications Series

Unpredictable fluctuations in customer demand at physical facilities can have substantial operational impacts. Current demand forecasting approaches tend to be simplistic and reactive as they rely primarily on internal historical data, which provides inadequate accuracy to support future resource allocation. This disclosure describes techniques to provide accurate medium-term demand forecasts for physical operational facilities by synthesizing a diverse set of relevant real-world factors obtained from a fusion of numerous heterogeneous data sources related to human practices and needs. The factors can be continually aggregated and input to a suitably trained machine learning model trained to generate specific actionable output designed …

System For Orchestrating Multi-Modal Generative Pipelines For State-Consistent Interactive Environments, Parnika Singhal, Chirag Saini, Nilanjana Thakur

Defensive Publications Series

Generative systems for dynamic interactive environments may face technical challenges related to maintaining logical consistency and narrative coherence, sometimes referred to as state drift. A computer-implemented system, such as a server or cloud computing platform, can utilize an orchestration layer to coordinate multiple specialized generative pipelines, for example, for narrative, visuals, and audio. A technique can involve constraining a generative language model to output a structured data object that separates a persistent world state vector from an ephemeral narrative payload. The persistent state vector can be reinjected into the model on subsequent turns to help maintain long-term coherence, while metadata …

Cable Coiling Detection Using Coupling-Based Sensing, Hp Inc

Defensive Publications Series

A smart cable monitoring system detects cable coiling conditions before cable degradation occurs. The system utilizes dedicated sensing conductors integrated within the cable to monitor changes in coupling characteristics caused by cable geometry variations. By comparing the measured sensing signature against a baseline condition, the system can identify prolonged coiled states and provide user feedback or preventive actions. Thermal sensing may also be incorporated to improve detection robustness and support cable health assessment.

Host No-Op Mode, Hp Inc

Defensive Publications Series

There are various attack vectors over USB such as injecting malware, HID spoofing, juice jacking etc. This issue can be mitigated by disabling the USB for good, leaving the functionality over USB unavailable to the user. This problem is more prominent for gadgets/peripherals like chargers/charging cables, storage devices (thumb drives/hard disks), mice/Keyboard etc. connected to the host via USB.

Orientation-Based Eye Safety Control For Floor-Facing Optical Sensors In Mobile Robots, Hp Inc

Defensive Publications Series

An orientation-dependent control system for floor-facing optical modules in a mobile ground vehicle is described. An orientation sensor supplies pitch and roll, or equivalent gravity-vector information, to a controller. If the vehicle remains outside a permitted orientation for a validation period, the controller disables, disconnects, or reduces optical emission. Normal operation is restored after the vehicle returns to a permitted orientation for a recovery period. The arrangement reduces unintended direct exposure during lifting, carrying, servicing, or transport and can reduce power use while the modules are not needed.

Hybrid Reaction Bridge, Hp Inc

Defensive Publications Series

Hybrid meeting platforms give remote attendees structured side-channel mechanisms—virtual hand raises, emoji reactions, poll responses—that are entirely absent for in-room participants, whose non-verbal behaviour remains visible only inside the physical room. This paper describes a bridging system that observes in-room participants through a shared room camera, detects a defined vocabulary of natural non-verbal signals (nodding, head-shaking, raised hands, and postural cues), and translates each qualified signal into a corresponding structured reaction event within the conferencing platform’s existing reaction model. The result is a unified reaction channel in which in-room and remote participants are represented with equivalent expressive bandwidth, without requiring …

Intent-Aware Human Motion Interpretation, Hp Inc

Defensive Publications Series

The central challenge of camera-based gesture input in video meetings is distinguishing deliberate interaction gestures from the natural hand motion that accompanies speech. This paper describes a system that meets that challenge through an intent classification engine evaluating kinematic, temporal, postural, spatial, and speech-aligned features over sliding temporal windows. Qualified gestures pass to a gesture-to-structure transformation pipeline that produces clean, labelled diagram elements—not freeform ink—and renders them as first-class shared content within the conferencing platform, visible to all participants in real time, without requiring specialised hardware.

Interaction Claim Standard, Brian Harber

Defensive Publications Series

Independently implemented systems increasingly produce, exchange, and interpret claims about AI interactions: what participated, which interaction the supporting evidence concerns, and how far an assertion reaches. Where no shared requirements govern these claims, agreement on their meaning across implementations remains incidental rather than normative—each implementation may interpret the same assertion differently.   This document defines the Interaction Claim Standard, a standard for expressing and interpreting claims about AI interactions across independently implemented systems. It standardizes four normative objects: claim semantics, interaction-specific evidentiary binding, claim boundaries, and normative interpretation requirements. Together, these establish cross-implementation agreement on the meaning of claims, their evidentiary …

Integrated Clamp Interface For Joining Mjf Additively Manufactured Components To Non‑Mjf Structural Members, Hp Inc

Defensive Publications Series

This disclosure describes an integrated fastening architecture that enables direct attachment of Multi Jet Fusion (MJF) additively manufactured components to structural elements manufactured from other materials. The solution uses a clamp geometry incorporated into the MJF component, a guided fastening mechanism, and controlled mating faces. The design enables rapid assembly, repeatable retention performance, and non-permanent attachment while minimizing additional hardware. Unlike adhesive bonding, interference fits, or external clamps, the approach separates assembly torque requirements from retention-force generation, improving manufacturing robustness and reducing sensitivity to torque variation. Figure 1a, Figure 1b, and Figure 2 illustrate representative fastening concepts and the screw-held …

Rubber Foot Speaker, Hp Inc

Defensive Publications Series

This disclosure proposed on repositioning the speaker from internal (inside the chassis) to the external surface. The design incorporates the use of rubber foot with speaker into a single component.  The intention of this idea is to reduce the impact of vibration from the speakers and eliminating the rattling issue when installed internally. The idea also promotes sound quality, flexibility and cost reduction.

Server Tray Guide Tool With Anti-Tilt And Winged-Edge Features For Rack Installation, N/A

Defensive Publications Series

Abstract

A guide tool assists with the installation of server trays into data center racks. Constructed from, for example, stainless steel, peek polymer, alumina, and ceramic, the tool is inserted into a rack slot before the server tray to spread the rack’s internal separators. This action creates an unobstructed path, enabling the tray to be seated smoothly and with reduced force. The tool’s design includes winged edges that function as both an insertion stop and a handle for removal. A stepped feature on one end of the tool prevents it from tilting during installation in slim shelf racks. The design …

Expand Iserialnumber In Usb Device Descriptor To Include Configuration Information, Hp Inc

Defensive Publications Series

This disclosure presents a modification to the iSerialNumber field in the USB Device Descriptor to include additional configuration information that manufacturing can use. Printer configuration frequently requires devices to be fully booted up for the manufacturing software to query the configuration.  This approach improves the manufacturing process by providing the needed device configuration at USB device enumeration time.  The manufacturing software can then take the appropriate actions without the need to wait for fully boot up.

Intent-Driven Orchestration For Enterprise Cyber-Physical Systems: Bridging Probabilistic Planning And Deterministic Actuation, Hp Inc

Defensive Publications Series

This disclosure describes an intent-driven orchestration architecture for enterprise cyber-physical systems that bridges generative, probabilistic, or agentic planning and deterministic device actuation. A non-executing, domain-aware planner converts high-level intent into a candidate graph of typed, idempotent Atomic Tool calls, but does not invoke tools or control devices. A deterministic orchestrator executes only graphs that pass four deterministic validation gates covering schema, governance, semantic correctness, and Device Twin-based simulation. Device Twins provide cached state for deterministic simulation, guard evaluation, binding resolution, and observability, but are not consumed by the Planner and are not authoritative for write-path success. Atomic Tools perform semantic …

Domain N Plus X Redundancy And Dynamic Thermal Management For High-Density Compute Domains, N/A

Defensive Publications Series

This paper describes a system architecture for managing high-density compute domains to address thermal and power constraints while facilitating continuous maintenance operations. The system utilizes an N plus x redundancy model for compute units and network switches, where supplemental units are managed as spares in varying low-power states. A central management controller orchestrates power steering and workload migration, aiming to adhere to designated power budgets while avoiding thermal hardware damage. This approach allows for ongoing service availability during hardware repairs and software updates without performance degradation.

Keywords: high-density compute domains, dynamic thermal management, workload migration, power steering, system redundancy, …

Dynamic Exponent-Mantissa Repurposing For Adaptive Microscaling (Demr-Am) Architecture, N/A

Defensive Publications Series

This publication details a hardware architecture and data processing methodology for dynamic exponent and mantissa repurposing, tailored to facilitate adaptive microscaling within machine learning accelerators. The described methodology addresses the quantization errors and memory bandwidth constraints that frequently emerge during the deployment of large-scale artificial intelligence models. By dynamically repurposing redundant exponent bits into additional mantissa bits specifically for outlier values within a given tensor block, the system dramatically enhances numerical precision without expanding the core memory footprint of the data block. A dedicated hardware decoder can be positioned immediately prior to a dot-product engine to re-route electrical signals and …

Agentic Knowledge Graph Engine With Continuous Evolution And Hybrid Semantic-Structural Embeddings For Intelligent Data Extraction, Knowledge Creation, And Retrieval, Hp Inc

Defensive Publications Series

An Agentic Knowledge Graph Engine (AKGE) is disclosed that provides a complete, continuously evolving knowledge lifecycle managed by a multi-agent autonomous architecture. The system comprises a data extraction layer receiving heterogeneous input sources, a schema-enforced knowledge graph storage layer with native temporal versioning and audit lineage, a hybrid embedding module fusing semantic, structural, contextual, and schema-constrained representations into dynamically updated entity embeddings, a continuous evolution engine responding to event triggers including query failure, embedding drift, conflict detection, and new data arrival, and a hybrid retrieval engine combining multi-hop graph traversal with approximate nearest-neighbor vector search. A specialized multi-agent ensemble — …

System And Method For Structured Queue Folder Management Of Print Jobs, Hp Inc

Defensive Publications Series

This disclosure presents a system and method for managing print queues as a persistent, hierarchical structure of folders, subfolders, and job records within production printing environments. Unlike conventional flat-list queue models, the disclosed approach preserves semantic relationships among related print jobs throughout the entire execution lifecycle, enabling group-level operations—such as print, pause, resume, cancel, and prioritize—to be applied coherently to collections of jobs without requiring individual job manipulation.

A central queue manager maintains a hierarchical model composed of folder nodes, subfolder nodes, and job nodes, each carrying metadata such as priority, state, sequencing policy, media requirements, and device constraints. A …

Hybrid Proactive And Reactive Voltage Set-Point Adjustment For Processor Power Maximization, N/A

Defensive Publications Series

The disclosure is directed to a hybrid control system that adjusts a processor’s voltage set-point by combining proactive workload analysis with reactive hardware feedback. A power performance policy uses feedback from an analog droop detector to reactively adjust the operating voltage, increasing it for stability or gradually decreasing it during stable operation to reduce power use. Concurrently, workload recognition software establishes a minimum voltage floor based on the predicted demands of upcoming tasks, setting the floor higher for high dynamic range workloads and lower for sustained workloads. A hardware-level mechanism also directly throttles the compute element to prevent timing violations …

Hardware-Level Protocol Proxying For Dynamic-Capacity Memory Systems, Vadlamudi Venkata Girish

Defensive Publications Series

Hardware-level protocol proxying (HLPP) resolves latency conflicts between a host’s transaction timers and the analog wake-up of power-gated, dynamic-capacity memory in compute express link (CXL) systems. The architecture decouples the host’s error-containment domain from the memory’s physical wake-up domain using one of three distinct hardware mechanisms. One approach involves a CXL switch performing dynamic topological hot-swapping, which routes initial memory requests through an isolated, high-latency virtual path and atomically swaps to a low-latency production path after the memory awakens. In another method, a CXL memory controller performs protocol spoofing by sending synthetic shadow completion packets to the host to repeatedly …

Imsa Summer Symposium 2026, Student Inquiry And Research (Sir)

Summer Symposium

For nearly four decades, IMSA has worked to advance the human condition through educational innovation. The Summer Symposium is our annual showcase for the research projects that were conducted in IMSA’s experiential learning programs between May and August. These programs include Student Inquiry and Research (SIR), IN2 Business Internships, and the Student Research and Experiential Learning Opportunity (SRELO) program. For the first time, we are also excited to welcome IMSA Research, Inquiry Skills, and Experimentation (RISE) students to present at the Summer Symposium.

Cover Design

Luke Berryman

Combined Effects Of Space-Relevant Thermal & Radiation Stress On Wide-Gap Mixed-Halide Perovskite Solar Cells, Veran Stanek

Theses

Mixed-halide wide-band gap perovskites are attractive for space power due to their high open-circuit voltages and compatibility with tandem device architectures, but their response to combined orbital stressors remains under-explored. Here, we study 1.86 eV band gap all-inorganic CsPbI2Br solar cells across a space-relevant thermal window of -80°C to +100°C before and after 1.0 M eV proton irradiation. n-i-p devices exhibit notable radiation tolerance retaining approximately 60% of initial power conversion efficiency (PCE) upon irradiation with a fluence of 5 × 10^13 cm^−2. Temperature-dependent hyperspectral photoluminescence imaging and spectroscopy uncover a reversible low-temperature emissive heterogeneity in the perovskite absorber, manifested …

Zineth, Colin Wilcox

Theses

This thesis asks whether a brand can generate real engagement without surrendering to an algorithm, and whether web design can recover the color, mystery, and sense of discovery it traded away as the internet corporatized. Daily internet use has risen every year on record, yet the spaces users inhabit have flattened into grey buttons and endless feeds. This question is answered through Zineth, a hybrid brand universe told across three coordinated forms: a Risograph-printed zine, a dense and layered website, and limited clothing drops. The zine operates as a key. Codes printed in its pages unlock gated areas of the …

Ppt-1 Regulates Lipid Droplet Homeostasis And Promotes Oxidative-Stress Resilience In C. Elegans., Chukwuadinula Deborah Ashiedu

Master’s Theses and Projects

No abstract provided.

Low-Temperature Co2-Enhanced Reductive Bioleaching For Selective Nickel Recovery From Iron-Rich Silicate Tailings, With Concurrent Carbon Sequestration, Kwabena Boafo, Palas Kamlakar Borkar, Atharva Rane, Timothy Eisele

Michigan Tech Publications

Recovering nickel from iron-rich mine tailings by bioleaching is held back by the co-dissolution of iron, which gives pregnant leach solutions with high Fe/Ni mass ratios and heavy downstream purification costs. This study shows that running bioleaching at 0 °C under a CO2-enriched atmosphere, with no post-leach iron removal step, lowers the Fe/Ni mass ratio in the leachate from 13.87 at 25 °C to 2.25 at 0 °C, a 6.2-fold improvement against a feed Fe/Ni ratio of 119:1. Nickel recovery stayed near 45% at both temperatures, showing that the fen-derived psychrotolerant consortium remained fully active near freezing. The mechanism rests …

Profiling Arbutin As A Potent Cholinesterase And Gsk-3Β Inhibitor And Evaluating Its Cytotoxicity In Sh-Sy5y Cells For Alzheimer’S Disease Therapy, Shirin Tarbiat, Tuğba Bal, Deni̇z Gülmez

Turkish Journal of Biology

Background/aim: Alzheimer’s disease (AD) is a neurodegenerative disorder linked to cognitive decline and memory loss, marked by hyperphosphorylated tau tangles and decreased acetylcholine levels, which are affected by acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE) activity. Glycogen synthase kinase-3 beta (GSK-3β) also plays a role in AD through abnormal activation. In this study, the potential of β-arbutin to inhibit cholinesterases and GSK-3β was investigated, utilizing in vitro assays, cytotoxicity analysis, and molecular docking simulations to explore its therapeutic implications.

Materials and methods: Ellman’s technique was used to test the inhibitory effects of arbutin on cholinesterase enzymes. An ATP-Glo luminescent assay kit was …