Explore an in-depth presentation on keep-alive caching for the Hawkes process, delivered at the Uncertainty in Artificial Intelligence conference. Dive into the design of caching policies for applications like serverless computing, where cache size is not fixed but incurs costs based on item retention time. Examine the proposed model that balances these costs against cache miss expenses. Discover the characterization of optimal caching policies, with a focus on closed-form solutions for Hawkes processes. Learn about the development of history-independent policies that achieve near-optimal performance. Gain insights from performance evaluations using simulations and real-world data from Microsoft's Azure Functions platform. Access accompanying slides to enhance your understanding of this cutting-edge research in caching strategies for dynamic environments.