1. Federated Multi-View K-Means with Enhanced Distance (Fed-MVKM-ED)

Comprehensive Implementation & Publication Demonstration
Developed and demonstrated a complete federated multi-view clustering algorithm for privacy-preserving distributed learning. The implementation features rectified Gaussian kernels, differential privacy mechanisms, and adaptive view weight learning.

Key Technical Achievements:

Performance: NMI: 0.8925, ARI: 0.6999 (32.7% improvement over local models)
Privacy-Preserving: Differential privacy with 0.9 privacy level
Multi-View Integration: Depth (6144D) + RGB (110D) feature fusion
Federated Setup: 3-site collaboration (Research Hospital, Rehab Center, Sports Clinic)
Comprehensive Evaluation: 12+ visualizations, convergence analysis, performance metrics

Dataset: DHA (Depth-included Human Action) simulation with 23 action categories across 600 samples. Applications: Healthcare collaboration, IoT networks, multi-institutional research.

Complete implementation available in A Comprehensive Tutorial on Federated Multi-View K-Means Clustering with Rectified Gaussian Kernel and GitHub Archive with publication-ready results and analysis.

2. Multi-View Clustering Framework (Foundational Work)

Designed and implemented a modular, scalable framework for multi-view clustering algorithms. Built to support dynamic feature integration, interpretability, and performance tracking across federated environments. This foundational work enabled the advanced Fed-MVKM-ED implementation showcased above.

3. Decentralized Anomaly Detection System

Led the development of a privacy-preserving anomaly detection model optimized for industrial edge devices. Deployed models across simulated sensor networks with minimal communication cost.

4. Clown Network Chronicles (Observational Series)

A satirical-analytical blog project that documents the absurdities of shared spaces, noise patterns, and unauthorized psychological phenomena through poetic sarcasm and digital logging.

5. Visualization of Federated Learning Topologies

Created animated visual models of peer-to-peer learning dynamics across edge devices, using real-time traffic simulation data and graph overlays.

6. Personal Website Infrastructure (2025)