Representation Learning of Daily Movement Data Using Text Encoders

arXiv:2405.04494v1 Announce Type: new
Abstract: Time-series representation learning is a key area of research for remote healthcare monitoring applications. In this work, we focus on a dataset of recordings of in-home activity from people living with Dementia. We design a representation learning method based on converting activity to text strings that can be encoded using a language model fine-tuned to transform data from the same participants within a $30$-day window to similar embeddings in the vector space. This allows for …

abstract applications arxiv cs.lg daily data dataset dementia design focus healthcare home key monitoring people representation representation learning research series strings text type work