Initializing LSTM internal states via manifold learning
Sign Up to like & getrecommendations! Published in 2021 at "Chaos"
DOI: 10.1063/5.0055371
Abstract: We present an approach, based on learning an intrinsic data manifold, for the initialization of the internal state values of long short-term memory (LSTM) recurrent neural networks, ensuring consistency with the initial observed input data.… read more here.
Keywords: lstm internal; initializing lstm; via manifold; states via ... See more keywords
Exploring Predictive States via Cantor Embeddings and Wasserstein Distance
Sign Up to like & getrecommendations! Published in 2022 at "Chaos"
DOI: 10.1063/5.0102603
Abstract: Predictive states for stochastic processes are a nonparametric and interpretable construct with relevance across a multitude of modeling paradigms. Recent progress on the self-supervised reconstruction of predictive states from time-series data focused on the use… read more here.
Keywords: via cantor; exploring predictive; states via; predictive states ... See more keywords