Research

My research focuses on robustness and explainability of large multimodal models — understanding how microscopic input perturbations cascade through neural representations to cause failures in vision-language understanding, embodied navigation, and LLM reasoning.

ActivationGuard: Prompt Injection Detection in LLM Agents
Under review

Frozen LLM hidden states, logistic regression · 2026

Built a lightweight detector for indirect prompt injection in network-integrated LLM agents using logistic regression on frozen LLM hidden states. Expanded training coverage raised attack recall from 17.2% to 93.8% under a 5% false-positive-rate validation budget, achieving 0.979 AUROC.

Adversarial Attacks on LLMs via Regularized Relaxation
Information Sciences 2026

Co-developed regularized continuous optimization of adversarial suffix embeddings with mapping to valid vocabulary tokens. Improved attack success across five LLMs while running approximately 100 times faster than greedy coordinate gradient (GCG) search.

Paper Codebase
Instruction Strategies for Tool-Grounded AI Agents
AgentSkills 2026 · Poster

MCP, agent evaluation · ACM CAIS 2026 AgentSkills Workshop

Evaluated instruction strategies in 180 cybersecurity-agent runs across four guidance levels. Increasing context from 591 to 36,001 tokens produced no significant success-rate gain, suggesting that additional procedural guidance may have limited value when the tool environment already supplies useful feedback.

arXiv
Neural Expansion: A Unified Mechanism for How Deep Neural Networks Generalize
NeurIPS 2026

Accepted at NeurIPS 2026

Proposed Neural Expansion, a geometric account of generalization based on Generalization Intervals, directions along which predictions remain unchanged. Stable regions grow during training; a theoretical bound connects interval size to classification margin and directional sensitivity.

Paper Codebase
Numerical Instability and Chaos in Large Language Models
IJCNN 2026 · Oral

International Joint Conference on Neural Networks (IJCNN 2026), part of IEEE WCCI 2026

Discovered that microscopic floating-point input perturbations cause up to 10-million-fold amplification through LLM layers, collapsing logit margins and triggering unpredictable token flips. Quantified regimes of stability and rounding-induced chaos to characterize when LLM predictions remain stable or become unpredictable.

arXiv Codebase
NeuroShield-ViT: Robust Vision Transformers via Representation Vulnerability Mitigation
ISVC 2025 · Oral

International Symposium on Visual Computing (ISVC 2025)

Examined how small input changes affect layer-wise representations in Vision Transformers (ViTs). Proposed NeuroShield-ViT to neutralize vulnerable neurons and enhance adversarial robustness without retraining, raising PGD-100 accuracy from 15.6% to 72.3%. Classification accuracy also improved by 71.6% against IGO attacks on ImageNet1K.

Paper
Adversarial Exploitation in Robot Vision-Language Navigation
IROS 2024 · Oral

IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2024)

Developed algorithms that exploit semantic representation gaps in vision-language navigation (VLN) models to redirect robotic agents along malicious paths. Achieved 91.7% landmark-conditioned path redirection and proposed a noise-sensitivity-based detection mechanism improving navigation policy stability by 96.3%.

Paper Codebase
Spatial-ViLT: Enhancing Visual Spatial Reasoning through Multi-Task Learning
ISVC 2025 · Oral

International Symposium on Visual Computing (ISVC 2025) · LNCS, Springer

Developed SpatialViLT, a vision-language model integrating depth maps, 3D coordinates, and edge representations via multi-task learning to improve spatial grounding for embodied intelligence. Improved the Visual Spatial Reasoning (VSR) benchmark by 2.1%.

Paper
Vision Embeddings and Hallucination Vulnerabilities in Multimodal LLMs
ISVC 2025 · Poster

International Symposium on Visual Computing (ISVC 2025)

Analyzed how vision encoder embeddings propagate hallucination vulnerabilities through multimodal LLMs, identifying representation-level failure modes that cause systematic visual misperception in deployed systems such as DeepSeek.

Paper Related arXiv
Leveraging Superresolution for Enhanced Deep Fake Detection

Xception, super-resolution · 2022

Evaluated state-of-the-art models for identifying low-resolution deep fakes in video data. Added super-resolution to improve test accuracy by 19% on the FaceShifter dataset using Xception Net.

Codebase