arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2202.11485cs.IRcs.LG
收藏

学习时间点过程实现连续时间事件序列的高效检索

Learning Temporal Point Processes for Efficient Retrieval of Continuous Time Event Sequences

IIT Delhi(印度理工学院德里分校);IIT Bombay(印度理工学院孟买分校)

Vinayak Gupta, Srikanta Bedathur, Abir De

摘要:

最近的发展,在预测建模使用标记时间点过程(MTPP),使几个真实世界的应用,涉及连续时间事件序列(CTES)的准确表征。然而,这些序列的检索问题在文献中仍然很大程度上没有得到解决。为了解决这个问题,我们提出了NEUROSEQRET,它学习检索和排名一组相关的连续时间事件序列为一个给定的查询序列,从一个大型语料库的序列。更具体地说,NEUROSEQRET首先在查询序列上应用可训练的解扭曲函数,这使得它与语料库序列相当,特别是当相关的查询-语料库对具有单独不同的属性时。然后,它将未变形的查询序列和语料库序列馈送到MTPP引导的神经相关性模型中。我们开发了两种变体的相关性模型,提供了准确性和效率之间的权衡。我们还提出了一个优化框架来学习二进制序列嵌入的相关性分数,适合本地敏感的哈希导致显着的加速返回给定的查询序列的前K个结果。我们对几个数据集的实验表明,NEUROSEQRET的准确性显著提高,超过了几个基线,以及我们的哈希机制的有效性。

英文摘要:

Recent developments in predictive modeling using marked temporal point processes (MTPP) have enabled an accurate characterization of several real-world applications involving continuous-time event sequences (CTESs). However, the retrieval problem of such sequences remains largely unaddressed in literature. To tackle this, we propose NEUROSEQRET which learns to retrieve and rank a relevant set of continuous-time event sequences for a given query sequence, from a large corpus of sequences. More specifically, NEUROSEQRET first applies a trainable unwarping function on the query sequence, which makes it comparable with corpus sequences, especially when a relevant query-corpus pair has individually different attributes. Next, it feeds the unwarped query sequence and the corpus sequence into MTPP guided neural relevance models. We develop two variants of the relevance model which offer a tradeoff between accuracy and efficiency. We also propose an optimization framework to learn binary sequence embeddings from the relevance scores, suitable for the locality-sensitive hashing leading to a significant speedup in returning top-K results for a given query sequence. Our experiments with several datasets show the significant accuracy boost of NEUROSEQRET beyond several baselines, as well as the efficacy of our hashing mechanism.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑