arXiv：cs.LG· Dmitry Osin, Egor Surkov, Petr Mokrov, Igor Udovichenko, Ivan Karpukhin, Evgeny Burnaev, Egor Shvetsov, Alexander Korotin·· 11 小时前
真实感 vs 准确率：从生成建模视角看事件序列预测
Realism VS Accuracy: Event Sequence Forecasting from a Generative Modeling Perspective
AI 导读
研究将事件序列的多步预测建模为条件生成问题，并构建了同时评估预测准确率与分布真实感的基准，涵盖统计基线、邻近领域模型与真实数据集。评测显示没有方法能同时在准确率和真实感上表现最优，两者存在根本性权衡。作者提出 Wasserstein BaryBooster，把真实感强的扩散模型样本聚合进 SOTA 预测器，证明分布真实感可转化为预测准确率。
正文
Abstract:Event sequences are chains of random events, each described by heterogeneous (numerical and categorical) features. The problem of event sequence forecasting occurs in finance, healthcare, business process intelligence, and other fields. In this work, we study multi-horizon forecasting of event sequences as a conditional generative modeling problem. We evaluate forecasts not only for prediction accuracy, but also for distributional realism. To this end, we present a benchmark comprising metrics for both quality aspects, strong statistical baselines, recent models adapted from adjacent fields, and curated real-world datasets. Careful evaluation reveals that no method excels at both accuracy and realism, highlighting a fundamental trade-off between the two. Finally, we introduce Wasserstein BaryBooster, an algorithm that aggregates samples from a realism-strong diffusion model into a state-of-the-art predictor, demonstrating that distributional realism can be converted into prediction accuracy. Code is available at this https URL.
| Comments: | Spotlight at the NeurIPS 2026 Track on Evaluations and Datasets. Code: this https URL |
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.12586 [cs.LG] |
| (or arXiv:2610.12586v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.12586 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Dmitry Osin [view email]
[v1]
Thu, 8 Oct 2026 18:07:47 UTC (699 KB)
来源：arXiv：cs.LG · arxiv.org