Advances in Temporal Point Processes: Bayesian, Neural, and LLM Approaches
A new survey paper reviews the latest advancements in temporal point processes (TPPs), which are models used to analyze event sequences. The paper covers traditional Bayesian methods, newer neural network approaches, and the emerging application of large language models (LLMs) in this field. It details model design and estimation techniques across these three frameworks and discusses their practical applications, while also identifying future research challenges. AI
IMPACT Provides a comprehensive overview of LLM applications in analyzing event sequences, potentially guiding future research and development in AI-driven temporal analysis.