Neural Spectral Marked Point Processes
This work addresses the problem of modeling complex discrete event data with marks and non-stationary dependencies for applications in machine learning and statistics, representing a novel method for a known bottleneck rather than an incremental improvement.
The paper tackled the need for more general point process models that incorporate contextual marks and non-stationary spatio-temporal dependence, introducing a neural network-based non-stationary influence kernel that demonstrated superior performance on synthetic and real data compared to state-of-the-art methods.
Self- and mutually-exciting point processes are popular models in machine learning and statistics for dependent discrete event data. To date, most existing models assume stationary kernels (including the classical Hawkes processes) and simple parametric models. Modern applications with complex event data require more general point process models that can incorporate contextual information of the events, called marks, besides the temporal and location information. Moreover, such applications often require non-stationary models to capture more complex spatio-temporal dependence. To tackle these challenges, a key question is to devise a versatile influence kernel in the point process model. In this paper, we introduce a novel and general neural network-based non-stationary influence kernel with high expressiveness for handling complex discrete events data while providing theoretical performance guarantees. We demonstrate the superior performance of our proposed method compared with the state-of-the-art on synthetic and real data.