Parameter Sharing for Spatio-Temporal Process Models

Author: R. Fischer, N. Piatkowski, K. Morik
Journal: LWDA
Year: 2019

Citation information

R. Fischer, N. Piatkowski, K. Morik,
LWDA,
2019,
http://ceur-ws.org/Vol-2454/paper_61.pdf

While probabilistic models such as Markov random fields can be highly beneficial for spatio-temporal data, they often suffer from over- fitting and have limited use in memory-constrained systems. We present a novel method to compress trained models based on temporal parameter sharing, which reduces redundancies in the parameters.