Learning Temporal Point Processes for Efficient Retrieval of Continuous Time Event Sequences
This addresses the retrieval problem for continuous-time event sequences, which is largely unaddressed in literature, offering a solution with tradeoffs in accuracy and efficiency for applications involving such data.
The paper tackles the problem of retrieving continuous-time event sequences by proposing NEUROSEQRET, which uses a trainable unwarping function and MTPP-guided neural models to improve accuracy and efficiency, achieving significant accuracy boosts over baselines in experiments.
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.