Real-time Intelligent Systems: Making Wise Decisions

This lecture explores the challenges of making wise decisions in data-intensive applications and systems, focusing on real-time intelligence. Topics include data management challenges, query-driven data sanitization, evolving indexes, and scaling sharing-aware optimization. The instructor discusses Proteus, a system for building just-in-time databases, and addresses hardware considerations for performance optimization. The lecture also covers harnessing accelerator-level parallelism, decoupling data from control flow, and accessing fresh data on GPUs. Additionally, it delves into the HTAP design spectrum, the impact of hardware heterogeneity on intelligent systems, and the importance of incorporating change into system design.