Advancing Time Series Classification with Multimodal Language Modeling

arXiv:2403.12371v1 Announce Type: new
Abstract: For the advancements of time series classification, scrutinizing previous studies, most existing methods adopt a common learning-to-classify paradigm - a time series classifier model tries to learn the relation between sequence inputs and target label encoded by one-hot distribution. Although effective, this paradigm conceals two inherent limitations: (1) encoding target categories with one-hot distribution fails to reflect the comparability and similarity between labels, and (2) it is very difficult to learn transferable model across domains, …

abstract arxiv classification classifier cs.lg distribution hot inputs language learn limitations modeling multimodal paradigm series studies time series type