Grant: Research Tools and Instruments

The Data Mining Lab and the LAB-HA at York University are leading multiple NSERC-funded projects aimed at developing innovative, efficient, and interpretable machine learning and data mining models. These projects, often in collaboration with industries in health, transportation, media, and high-tech sectors, leverage large datasets to address real-world challenges. However, the current hardware limitations are impeding progress and competitiveness in this fast-evolving field. The equipment requested in this RTI grant will enhance research capabilities, enabling the processing of large datasets and training of deep learning models, which can lead to the development of new products or improvements for Canadian industry partners, giving them a global competitive edge. Advances in both machine learning algorithms and computer hardware have significantly improved the efficiency of training deep neural networks. GPUs, often equipped with AI-specific features, have largely replaced CPUs for training large-scale ML models. As modern ML strategies evolve in academia and industry, keeping pace with computational demands is crucial for training high-quality personnel (HQP), sustaining competitive research output, and supporting industry partnerships. The RTI grant will enable us to purchase a high-performance GPU server and refurbish outdated workstations, ensuring we have the necessary hardware to meet these challenges. The equipment will be utilized by postdoctoral fellows, graduate students, and undergraduate research assistants involved in the following NSERC-funded projects: (1) ML based optimization for database workloads, (2) deep learning models for mobility data mining, (3) deep learning models for spatiotemporal graphs/networks, (4) multi-agent reinforcement learning based routing of autonomous vehicles, (5) streaming algorithms for online mining and learning of IoT sensor data, (6) knowledge-based neural question generation from text, (7) a deep generative network for ultra-resolving images, (8) GPU job scheduling using deep reinforcement learning, (9) wearable platforms for self-monitoring of biomarkers, and (10) portable, sensitive platform technologies for rapid detection of critical minerals in fluidics. Our research program aligns with the Pan-Canadian AI Strategy by fostering world-class machine learning and data science research, which has the potential to attract top international talent while retaining the best domestic researchers. This program also plays a key role in training HQP with advanced analytical and problem-solving skills, enabling them to leverage AI for boosting innovation and productivity in knowledge-based industries, driving economic growth, and enhancing quality of life in Canada and beyond.