Search Results - kaiqi+xiong

5 Results Sort By:
A Divide and Conquer Framework for Breaking Unlearnable Data
Advantages Bypasses unlearnable image data through label modifications Achieves over 85% accuracy on protected datasets Work across multiple types of unlearnable data Use both manual and automated (genetic algorithm) approaches Reveal hidden weaknesses in current image protection methods Help researchers test and improve future unlearnable defenses Summary Unlearnable...
Published: 8/21/2025   |   Inventor(s): Hapuarachchige Thushari Hapuarachchi, Kaiqi Xiong
Keywords(s): Cybersecurity, Data Security, Datasets
Category(s): Technology Classifications > Computer Science > Algorithms, Technology Classifications > Computer Science > C/S Cybersecurity, Technology Classifications > Computer Science > Datasets
Advancing Ensemble Learning Against Unlearnable Data
Advantages Integrates seamlessly into enterprise workflows Scales from edge devices to GPU clusters Guarantees compliance with data-protection standards Enhances ensemble learning with data augmentation and model diversity Recovers >89% test accuracy on datasets protected by 12 state-of-the-art unlearnable data methods Summary As AI adoption...
Published: 7/15/2025   |   Inventor(s): Hapuarachchige Thushari Hapuarachchi, Kaiqi Xiong
Keywords(s):  
Category(s): Technology Classifications > Data/AI > Cybersecurity - Data/AI, Technology Classifications > Computer Science > Algorithms
Improving Facial Recognition Systems Against Unlearnable Data
Advantages: Dramatically reduces the impact of unlearnable noise using smart data transformations Recovers facial recognition accuracy from nearly 0% to over 85% Designed for large facial datasets with limited number of images per person Summary: Some facial datasets are made "unlearnable" by adding small, hidden changes that stop...
Published: 4/17/2025   |   Inventor(s): Kaiqi Xiong, Long Kim Dang, Hapuarachchige Thushari Hapuarachchi
Keywords(s):  
Category(s): Technology Classifications > Data/AI > Cybersecurity - Data/AI
SYN Flood Attack Detection and Mitigation for Software-Defined Networks
Competitive Advantages • Dynamic and self-adjusting detection threshold • Better mitigation and response time • High accuracy and efficiency Summary USF inventors have developed a mechanism in which they first examine the impact of the SYN saturation against the data—to—control layer and then follow a novel approach...
Published: 7/18/2023   |   Inventor(s): Mohamed Rahouti, Kaiqi Xiong
Keywords(s): Computers Software and Information Technology, Environmental Sciences, Water Resources
Category(s): Technology Classifications > Computer Science, Technology Classifications > Computer Science > C/S Cybersecurity
LatencySmasher: A Software-Defined Networking-Based Framework for End-to-End Latency Optimization
Competitive Advantages Quality of Service-aware framework prioritizes delay sensitive applications Fast & adaptive optimal end-to-end path computation Reduces control overhead Computationally efficient Summary USF inventors have designed a mechanism that leverages both the global control and centralized view of the SDN to...
Published: 7/18/2023   |   Inventor(s): Kaiqi Xiong, Yufend Xin, Nasir Ghani, Mohamed Rahouti
Keywords(s): Computer Science, Computers Software and Information Technology
Category(s): Technology Classifications > Computer Science, Technology Classifications > Software