Search for: All records

As IoT device adoption grows, ensuring cybersecurity compliance with IoT standards, like National Institute of Standards and Technology Interagency (NISTIR) 8259A, has become increasingly complex. These standards are typically presented in lengthy, text-based formats that are difficult to process and query automatically. We built a knowledge graph to address this challenge to represent the key concepts, relationships, and references within NISTIR 8259A. We further integrate this knowledge graph with Retrieval-Augmented Generation (RAG) techniques that can be used by large language models (LLMs) to enhance the accuracy and contextual relevance of information retrieval. Additionally, we evaluate the performance of RAG using both graph-based queries and vector database embeddings. Our framework, implemented in Neo4j, was tested using multiple LLMs, including LLAMA2, Mistral-7B, and GPT-4. Our findings show that combining knowledge graphs with RAG significantly improves query precision and contextual relevance compared to unstructured vector-based retrieval methods. While traditional rule-based compliance tools were not evaluated in this study, our results demonstrate the advantages of structured, graph driven querying for security standards like NISTIR 8259A. 

The dramatic surge of health misinformation on social media platforms poses a significant threat to public health, contributing to hesitancy in vaccines, delayed medical interventions, and the adoption of untested or harmful treatments. We present a novel, hybrid AI-driven framework designed for the real-time detection of health misinformation on social media platforms while prioritizing user privacy. The framework integrates the strengths of Large Language Models (LLMs), such as DistilBERT, with domain-specific Knowledge Graphs (KGs) to enhance the detection of nuanced and contextually dependent misinformation. LLMs excel at understanding the complexities of human language, while KGs provide a structured representation of medical knowledge, allowing factual verification and identification of inconsistencies. Furthermore, the framework incorporates robust privacy-preserving mechanisms, including differential privacy and secure data pipelines, to address user privacy concerns and comply with healthcare data protection regulations. Our experimental results on a dataset of Reddit posts related to chronic health conditions demonstrate the performance of this hybrid approach compared to models that only use text or KG, highlighting the synergistic effect of combining LLMs and KGs for improved misinformation detection.