Research map / Continual learning / Online and replay-based
Prompt-augmented Temporal Point Process for Streaming Event Sequence
Siqiao Xue, Yan Wang, Zhixuan Chu, Xiaoming Shi, Caigao JIANG, Hongyan Hao, Gangwei Jiang, Xiaoyun Feng, James Y. Zhang, JUN ZHOU
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add CtRetroPromptPool to AttNHP/NHP (Pt-NHP, Pt-ANHP) vs CL-NHP/CL-ANHP | Pt-NHP / Pt-ANHP · Taobao, Amazon, StackOverflow | type error rate and time RMSE: relative 4% - 6% | Improved |
| replace NHP with stronger AttNHP base in retraining (ablation) | Re-ANHP vs Re-NHP · Taobao, Amazon | average CL performance | No clear effect |
| replace NHP with AttNHP base in prompt method (ablation) | Pt-ANHP vs Pt-NHP · Taobao, Amazon | error rate and RMSE | No clear effect |
| use temporal prompt instead of standard prompt (Pt-ANHP-std) (ablation) | Pt-ANHP · Taobao, Amazon | error rate: p-value <0.05 | Improved |
| use temporal prompt instead of standard prompt (Pt-ANHP-std) (ablation) | Pt-ANHP · Taobao, Amazon | time RMSE: p-value approx 0.08 | Improved |
| use prefix-tuning attachment instead of naive prompting or prompt tuning (ablation) | Pt-ANHP · Taobao, Amazon | error rate and RMSE | Improved |
| remove the prompt pool, use a single temporal prompt (w/o CtRroPP) (ablation) | PromptTPP · Amazon | performance | Hurt |
| remove learnable keys, use mean of prompts as keys (w/o k-v) (ablation) | PromptTPP · Amazon | performance | Hurt |
| asynchronous refresh C=2 instead of C=1 (ablation) | Pt-ANHP · Taobao | performance | No clear effect |
| asynchronous refresh C=2 instead of C=1 (ablation) | Pt-ANHP · Amazon | convergence | Improved |
| asynchronous refresh C=4 instead of C=1 (ablation) | Pt-ANHP · Taobao, Amazon | convergence | No clear effect |
| too small prompt length L_p (ablation) | Pt-ANHP · Amazon | error rate and RMSE | Hurt |
Closest papers by meaning
- EasyTPP: Towards Open Benchmarking Temporal Point Processes (ICLR 2024 · similarity 0.89)
- Deep Continuous-Time State-Space Models for Marked Event Sequences (NeurIPS 2025 · similarity 0.86)
- Residual TPP: A Unified Lightweight Approach for Event Stream Data Analysis (ICML 2025 · similarity 0.86)
- Interacting Diffusion Processes for Event Sequence Forecasting (ICML 2024 · similarity 0.85)
- Decoupled Marked Temporal Point Process using Neural Ordinary Differential Equations (ICLR 2024 · similarity 0.85)
- ExPERT: Modeling Human Behavior Under External Stimuli Aware Personalized MTPP (AAAI 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
