Cybersecurity
Intrusion detection, vulnerability prioritization, e-commerce risk, and cyber policy.
Research
Research across
security and AI.
Faculty and students investigate secure systems, adversarial learning, neural computing, and visual intelligence.
Featured collaboration
EEG to text, from model to infrastructure
Saydul Akbar Murad and Nick Rahimi connect brain signal decoding with secure, scalable peer-to-peer fog computing.
Explore Murad’s researchFocus
Research areas
Trustworthy AI
Adversarial robustness, safe learning, explainability, and model evaluation.
Neural & distributed systems
EEG decoding, privacy, fog architectures, and secure peer-to-peer computing.
Visual intelligence
Video reasoning, remote sensing, and dynamic scene representation.
From the lab
Recent projects
Multimodal evaluation · 2026
Redemption Score
Evaluating image captions through perceptual, linguistic, and distributional signals.
Ashim Dahal · Ankit Ghimire · Saydul Akbar Murad · Nick Rahimi
Read accepted manuscript
Graph learning · 2026
Bitcoin transaction graph learning
Comparing Euclidean and hyperbolic graph models on Bitcoin transaction networks.
Ankit Ghimire · Saydul Akbar Murad · Nick Rahimi
Read preprint
Social engineering · 2026
Filter-then-Verify
A graph filter and language model work together to detect email social engineering.
Barsat Khadka · Prasant Koirala · Kshitiz Neupane · Nick Rahimi
Read preprint
Video understanding · 2026
POVQA
Video question answering with rationales and fewer input tokens.
Ashim Dahal · Ankit Ghimire · Saydul Akbar Murad · Nick Rahimi
Project website
Robotics · 2026
TwinSim
A Gaussian scene digital twin for simulation to real world robot policy training.
Ashim Dahal · Nick Rahimi
Watch demonstration below
Dynamic scenes · 2026
4D Gaussian streaming
Testing how anchor selection changes reconstruction quality and efficiency.
Ashim Dahal · Rabab Abdelfattah · Nick Rahimi
Read preprint
Vision-language models · 2025
Embedding Shift Dissection on CLIP
Tracking how image augmentations change vision-language representations.
Ashim Dahal · Saydul Akbar Murad · Nick Rahimi
Read paper
Remote sensing · 2025
Vision Transformers versus CNNs
Comparing segmentation models on remote sensing imagery.
Ashim Dahal · Saydul Akbar Murad · Nick Rahimi
Read paper
Model efficiency · 2025
Convolutional Kolmogorov-Arnold Networks
Measuring practical efficiency bottlenecks across image and tabular tasks.
Ashim Dahal · Saydul Akbar Murad · Nick Rahimi
Read preprint
Cyber defense · 2026
Attack-aware intrusion detection
Adaptive multi-stage defenses against adversarial network traffic.
Oluseyi Olukola · Nick Rahimi
Read preprint
Neural computing
EEG-to-text and secure fog architecture
Decoding brain signals with scalable and secure peer-to-peer infrastructure.
Saydul Akbar Murad · Nick Rahimi
Explore researchAcross the lab
More work
AI safety · 2026
Mastery-conditioned policy optimization for AI tutoring
Oluseyi Olukola · Nick Rahimi
AI safety
Pedagogical safety in AI tutoring
Oluseyi Olukola · Nick Rahimi
E-commerce security
Securing e-commerce: risk management and compliance
Derek Taylor · Mehdi Barati · Nick Rahimi
Password security
Adversarial password strength estimation
Pappu Jha · Hanzla Hamid · Oluseyi Olukola · Ashim Dahal · Nick Rahimi
Side-channel analysis
Machine learning for AES key recovery
Mukesh Poudel · Nick Rahimi
Insider threat
DANTE.v2 scenario-specific detection
Alabhya Pahari · Aryan Adhikari · Oluseyi Olukola · Ahmed Sherif · Nick Rahimi
Cyber resilience
WBTS vulnerability prioritization
Sindhuja Penchala · Nick Rahimi
Threat intelligence
Dark web traffic analysis under attack
Nyzaireyus Harrison · Heather Broome · Yaju Shrestha · Alexander Robles · Aayush Gautam · Nick Rahimi
Knowledge systems
Cache-augmented generation in RAG
Yaju Shrestha · Nick Rahimi
Language models
Historical text optimization
Heather Broome · Nick Rahimi
News classification
Rethinking news classification
Luana De Jesus Ferreira · Nick Rahimi
Cyber threat detection
Multilingual threat detection in Tweets/X
Ashim Dahal · Saydul Akbar Murad · Nick Rahimi
Handwriting recognition
Bangla character recognition with multi-head attention
M. Raquib · A. Polok · K. Biswas · F. Prity · Saydul Akbar Murad · Nick Rahimi
Explainable AI
Zero-day attack detection using MLP and XAI
Ashim Dahal · P. Bajgai · Nick Rahimi
Watch the work
Research demonstrations
Robotics
TwinSim
A silent visual walkthrough of the digital twin scene used for robot policy training.
Video description
The demonstration moves through a simulated environment and shows the robot alongside its reconstructed surroundings.
Video understanding
POVQA
A visual comparison of questions, human references, and model responses over sampled video frames.
Video description
The video cycles through movie clips. Each example shows time-stamped frames and subtitles, then a question, human reasoning and answer beside the model response.
Dynamic scenes
4D Gaussian streaming
A side-by-side comparison of reconstructed scenes under different anchor policies.
Video description
The video compares three anchor-selection policies. Each column shows a reconstruction, selected 3D Gaussians, and an influence blend beside ground truth.
Vision-language models
CLIP embedding shift
An animation showing changes in CLIP attention across an image sequence.
Video description
The sequence displays changing attention overlays to illustrate how image transformations affect the model representation.
Methods and prototypes
Research figures
