Research & Projects
Exploring the frontiers of artificial intelligence through innovative research and practical applications
Publications
Improved Energy Valley Optimizer for Wildfire Prediction Under review
Environmental Modelling & Software (2025)
Benchmarks the Energy Valley Optimizer and spiral variants for province-specific wildfire forecasting across Canada.
Spiral-Enhanced Liver Cancer Algorithm for Feature Selection Under review
Journal of Computational Science (2025)
Introduces spiral-based feature-selection method improving sensitivity in imbalanced datasets.
Advanced Wildfire Prediction with Machine Learning Published
MSc Thesis, Thompson Rivers University Library (2025)
Comprehensive thesis exploring machine learning approaches for wildfire prediction using metaheuristic optimization and spiral-based algorithms.
Research Projects
Spiral Optimization Framework
Development of spiral-enhanced metaheuristic algorithms for feature selection and optimization in imbalanced wildfire and environmental datasets.
Technologies:
| Algorithm | Accuracy | F1-Score | Time (s) |
|---|---|---|---|
| Spiral-Enhanced | 94.2% | 0.91 | 2.3 |
| PSO | 89.1% | 0.85 | 1.8 |
| GA | 87.3% | 0.82 | 3.1 |
Simulation-Driven Reinforcement Learning
Design of multi-fidelity simulation environments that bridge physical data and AI models, improving generalization and decision-making in safety-critical domains.
Technologies:
| Environment | Success Rate | Episodes | Reward |
|---|---|---|---|
| High-Fidelity | 87.3% | 1000 | 0.92 |
| Medium-Fidelity | 82.1% | 500 | 0.88 |
| Low-Fidelity | 76.8% | 250 | 0.84 |
Vision-Language Grounding for Environmental AI
Exploration of multimodal models integrating satellite imagery and textual reports through large language models for explainable geospatial reasoning.