Research contributions in LLM security, explainable AI, and machine learning
Google Scholar ProfileIEEE Big Data 2025 - Workshop on Knowledge Graphs and Big Data
Presents an efficient approach for multi-hop question answering over knowledge graphs by combining LLM-based query planning with embedding-guided graph search, enabling accurate reasoning across complex knowledge structures.
AAAI Fall Symposium 2025 - SECURE-AI4H
Applies conformal prediction methods to verify and quantify uncertainty in LLM-based extractions from electronic health records, enabling safer and more reliable clinical AI applications.
Workshop on Deployable AI at AAAI-2025
Research on privacy-preserving methods for collaborative generative AI, enabling multiple parties to jointly leverage LLMs without exposing sensitive data. Addresses secure inference and aggregation challenges.
UMAP Workshops 2025
Introduces a comprehensive benchmark for evaluating LLM security through automated penetration testing. Analyzes vulnerabilities across multiple LLM architectures and proposes improvements for building more robust systems.
Conference on Computer Vision and Pattern Recognition (CVPR) 2024 - Workshop
Proposes an ensemble approach for detecting AI-generated images that can adapt to new generator architectures with minimal training data. Addresses the challenge of rapidly evolving synthetic image generation.
IEEE Big Data 2024 - Workshop on Knowledge Graphs and Big Data
Develops LLM-based pipelines for extracting structured medical knowledge from clinical text. Enables construction of medical knowledge graphs for improved search and summarization in healthcare applications.
ICCBR Workshops 2021
Explores how case-based reasoning can enhance explainable AI by providing explanations that go beyond what models alone can offer. Demonstrates the value of domain knowledge in generating meaningful explanations.
ICCBR Workshops, 52-62
Introduces explanation containers for biomedical question-answering systems, providing structured explanations that combine case-based reasoning with modern NLP techniques.
2021
Comprehensive tutorial on explainable AI methods, covering interpretability techniques, evaluation approaches, and practical implementation guidance.