Digital Commons Network^™

Method For Trust-Aware And Language Model–Driven Email Threat Detection And Mitigation, Niranjan M M

Defensive Publications Series

Email remains the primary communication channel for enterprises and continues to be the most exploited attack vector for cyber threats. Modern email attacks increasingly rely on impersonation, social engineering, and contextual manipulation rather than traditional malware or malicious links, allowing them to evade existing detection mechanisms. At the same time, email security systems are adopting large language models (LLMs) to improve intent detection and contextual analysis, introducing new risks where email content itself can manipulate or degrade automated reasoning.

The proposal introduces a secure, trust-aware, and LLM-safe framework for automated email threat detection and remediation. The proposed system integrates sender …

Techniques For Preventing Anchoring Bias In Ai Investigative Analysis Through Staged Data Redaction And Hypothesis-Gated Revelation, Avneet Singh Chhabra

Defensive Publications Series

Presented herein is a system for preventing anchoring bias in large language model (LLM) investigative analysis. Bias in LLMs is not typically addressed by instruction-level approaches; rather, it is most effectively mitigated by physically separating the biasing data from an LLM's input during the hypothesis formation phase. The proposed system includes five core components: a Data Redaction Engine, a Staged Analysis Protocol, a Structural Gate, a Corrections Bridge, and a Procedural Attestation Record. The Data Redaction Engine operates to separate source data into a redacted evidence set and an attribution set, validated by a multi-layer validation pass. The Staged Analysis …

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 …

National Ai Forensic Grid: Federated Multi-Modal Deepfake Detection And Cryptographic Evidence Infrastructure For Law Enforcement, Pranav Bhatnagar Mr

Defensive Publications Series

This disclosure presents a federated forensic infrastructure designed for detection of malicious synthetic media and preservation of evidentiary integrity across law enforcement jurisdictions. The system integrates multi-modal deepfake detection, cross-node artifact intelligence exchange, and cryptographic chain-of-custody preservation. The disclosed architecture enables scalable national coordination, standardized forensic scoring, and tamper-resistant evidence management suitable for judicial proceedings.

Generating Financial Risk Tiers For Geo-Located Entities Using Monetary Claims From User-Generated Content And Transactional Signals, Mithun Kumar S R

Defensive Publications Series

Moderation systems for local business platforms may have difficulty identifying certain types of financial fraud, such as overcharging or fake deposit scams, because of challenges in validating monetary claims found in user-generated content (UGC). A disclosed method can address this potential limitation by synthesizing two different data streams. A system can use technologies, for example natural language processing and computer vision, to extract unstructured financial claims and pricing information from UGC, which may include content such as text reviews and photos of menus or receipts. The system can then correlate these claims with structured, verified transactional signals from integrated payment …

Halo: Heterophily-Aware Label-Free Ordering For Unsupervised Graph Fraud Detection

Defensive Publications Series

Graph fraud detection (GFD) has rapidly advanced in protecting online services by identifying malicious fraudsters. Recent supervised GFD research highlights that heterophilic connections between fraudsters and users can greatly impact detection performance, since fraudsters tend to camouflage themselves by building more connections to benign users. Despite the promising performance of supervised GFD methods, the reliance on labels limits their applications to unsupervised scenarios; Additionally, accurately capturing complex and diverse heterophily patterns without labels poses a further challenge. To fill the gap, we propose a Heterophily-guided Unsupervised Graph fraud dEtection approach (HUGE) for unsupervised GFD, which contains two essential components: a …

Halo: Heterophily-Aware Label-Free Ordering For Unsupervised Graph Fraud Detection

Defensive Publications Series

Graph fraud detection (GFD) has rapidly advanced in protecting online services by identifying malicious fraudsters. Recent supervised GFD research highlights that heterophilic connections between fraudsters and users can greatly impact detection performance, since fraudsters tend to camouflage themselves by building more connections to benign users. Despite the promising performance of supervised GFD methods, the reliance on labels limits their applications to unsupervised scenarios; Additionally, accurately capturing complex and diverse heterophily patterns without labels poses a further challenge. To fill the gap, we propose a Heterophily-guided Unsupervised Graph fraud dEtection approach (HUGE) for unsupervised GFD, which contains two essential components: a …

Refining Multi-Relational Graph Anomaly Detection With Learned Structures And Scalable Meta-Path Aggregation

Defensive Publications Series

Graph Neural Networks (GNNs) have recently achieved remarkable success in various learning tasks involving graph-structured data. However, their application to multi-relational graph anomaly detection problems on real-world datasets presents several challenges that significantly hinder performance: the structural noise and inconsistencies inherent in real-world graph data, difficulties in aggregating information across multiple relation types, and the imbalanced class labels. To address these limitations, we introduce a graph structure learning layer designed to refine the original, noisy graph structure, enhancing the representation of node relationships. This enables our model to effectively handle inconsistencies and structural noise present in the original graph data …

Ai-Enabled Federated Learning System For Privacy-Preserving Health Insurance Underwriting And Fraud Detection, Vijayalaxmi Methuku

Defensive Publications Series

This disclosure presents an AI-powered federated learning system designed to enhance health insurance underwriting and fraud detection while maintaining strict privacy standards. The proposed system enables healthcare providers, insurers, and regulatory bodies to collaboratively train machine learning models without sharing sensitive patient data. The federated architecture ensures secure data processing, bias mitigation, and real-time fraud detection. Key innovations include distributed AI training, differential privacy, blockchain-based claim tracking, reinforcement learning for policy adjustments, and NLP-driven prior authorization automation. This system optimizes risk assessment, reduces fraudulent claims, and ensures transparency in health insurance operations.

Security And Privacy Risks Of Multiple Payment Systems, Miguel Silva

Defensive Publications Series

The proliferation of multiple payment systems has introduced significant security and privacy risks, as users navigate a complex landscape of digital transactions. These risks arise from vulnerabilities in system integration, data breaches, and insufficient encryption protocols. Additionally, the lack of standardized security practices across platforms exacerbates the challenges of ensuring user privacy. Addressing these concerns requires comprehensive solutions that prioritize robust encryption, secure data sharing, and unified regulatory frameworks.

Skimmer Detection On Payment Terminals, Adam Ratica

Defensive Publications Series

The present disclosure relates to an apparatus and a method for efficiently detecting skimmers on ATM/POS machines. The apparatus and method sends images of ATMs/ POS machines to a central service. LIDAR is used to take a detailed 3D image of the ATMs/ POS machine and the service generate a complete 3D model. The service include a database of different vendor ATMs/ POS machines, known skimmers, and a history of GPS coordinates and LIDAR 3D information. In an instance in which the LIDAR and images doesn’t match a known Vendor and/or past LIDAR/Image and GPS data do not match the …

Method And System For Secure Credential Generation, Rhidian John

Defensive Publications Series

One embodiment includes a method performed by an authentication provider. The method includes receiving a sensitive information from a user. For example, the sensitive information may include payment data or a healthcare data of the user. The method includes transmitting the sensitive information to a processing server and receiving a long living token associated with the sensitive information from the processing server. The authentication provider stores the long living token. The method further includes the authentication provider receiving, from a resource provider, a request for a subset of the sensitive information. The method includes transmitting the subset of the sensitive …