Non-autoregressive diffusion-based temporal point processes for continuous-time long-term event prediction

Wang-Tao Zhou, Zhao Kang, Ling Tian, Jinchuan Zhang, Yumeng Liu

Abstract: Highlights•Efficient end-to-end framework for long-term continuous-time event prediction.•A recurrent residual denoising network that captures complex long-term patterns.•Extensive experiments that validate its superiority over state-of-the-art models.