Institution profile

Gonzaga University

Academic institutionnorthamerica · us
Official website
Research library2linked papers
Opportunities0open roles
Selected work

Representative Papers

Topological Attribution Distance (TAD): Revealing Segment-Level RAG Influence on LLM Output Geometry for Incident Log Analysis

Aug 17, 2026

This study addresses the critical challenge of trustworthy provenance and evidence attribution for LLM-generated content in cybersecurity applications. To this end, we propose Topological Attribution Distance (TAD), a novel segment-level attribution mechanism grounded in topological geometry. By leveraging embedding space modeling and hidden state analysis, TAD quantifies the global geometric influence of retrieval logs on generated outputs. This approach adaptively localizes critical logs to enable segment-wise evidence verification and interpretable decision tracing. Consequently, TAD effectively elucidates the intrinsic mechanisms of Retrieval-Augmented Generation, providing both theoretical foundations and technical pathways for enhancing the trustworthiness of model outputs in security-critical domains.

0 citationsRead paper

AoI-Aware Multi-Robot Sensing and Transport on Connected Graphs

May 03, 2026

This work addresses cooperative sensing and data dissemination in multi-robot systems operating over connected graphs, aiming to minimize the Age of Information (AoI) while accounting for stochastic sensing delays and hop-based communication delays. By decomposing AoI into sensing and propagation components, the authors separately optimize robot resource allocation and sample transmission paths. The sensing component is formulated as a separable discrete convex resource allocation problem, solved optimally via a greedy water-filling algorithm, while the propagation component leverages shortest-path trees combined with Eulerian tours to construct a full-delivery mechanism. Theoretical analysis demonstrates that this mechanism achieves the derived network-wide lower bound on AoI, and simulations confirm its effectiveness in significantly reducing AoI compared to baseline approaches.

0 citationsRead paper
Recent publications

Latest Papers

Topological Attribution Distance (TAD): Revealing Segment-Level RAG Influence on LLM Output Geometry for Incident Log Analysis

Aug 17, 2026

This study addresses the critical challenge of trustworthy provenance and evidence attribution for LLM-generated content in cybersecurity applications. To this end, we propose Topological Attribution Distance (TAD), a novel segment-level attribution mechanism grounded in topological geometry. By leveraging embedding space modeling and hidden state analysis, TAD quantifies the global geometric influence of retrieval logs on generated outputs. This approach adaptively localizes critical logs to enable segment-wise evidence verification and interpretable decision tracing. Consequently, TAD effectively elucidates the intrinsic mechanisms of Retrieval-Augmented Generation, providing both theoretical foundations and technical pathways for enhancing the trustworthiness of model outputs in security-critical domains.

0 citationsRead paper

AoI-Aware Multi-Robot Sensing and Transport on Connected Graphs

May 03, 2026

This work addresses cooperative sensing and data dissemination in multi-robot systems operating over connected graphs, aiming to minimize the Age of Information (AoI) while accounting for stochastic sensing delays and hop-based communication delays. By decomposing AoI into sensing and propagation components, the authors separately optimize robot resource allocation and sample transmission paths. The sensing component is formulated as a separable discrete convex resource allocation problem, solved optimally via a greedy water-filling algorithm, while the propagation component leverages shortest-path trees combined with Eulerian tours to construct a full-delivery mechanism. Theoretical analysis demonstrates that this mechanism achieves the derived network-wide lower bound on AoI, and simulations confirm its effectiveness in significantly reducing AoI compared to baseline approaches.

0 citationsRead paper