2019 · 46 citations · 12 references
ComputingComputer SciencePhysicsArtificial IntelligenceDynamical SystemsMachine LearningComputational ModelingDeep LearningDynamic Systems ModelingNeural ModelingNeural NetworksState-space ModelsData-driven ModelingHybrid Dynamical SystemsLimited Data SetsLong Short-term Memory NetworksModel FusionNeural Dynamical SystemsRecurrent Neural NetworksTemporal Memory
We introduce a novel machine learning-based fusion model termed as PI-LSTM (Physics-Infused Long Short-Term Memory Networks) that integrates first principle Physics-Based Models and Long Short-Term Memory (LSTM) network. Our architecture aims at combining equation-based models with data-driven machine learning models to enable accurate predictions of complex dynamic systems. In this hybrid architecture, recurrency aids the temporal memory of the inputs and output of the partial physics model, in a way that facilitates generalization with scarce data sets. We illustrate the application of PI-LSTM on two dynamical systems namely Inverted Pendulum and Tumor Growth. Empirical results on both test problems stand witness to the effectiveness of using physics in guiding machine learning models and the superiority of the outlined hybrid model over purely data-driven models.
12
Ask Me Anything: Dynamic Memory Networks for Natural Language Processing
Ankit Kumar, Ozan İrsoy, Peter Ondrúška et al. · arXiv (Cornell University) · 2015 · 631 citations · Full text
Mathematical Models of Avascular Tumor Growth
Tiina Roose, S. Jonathan Chapman, Philip K. Maini · SIAM Review · 2007 · 565 citations
Medicine, Cancer Biology, Oncology +17
Physics-guided Neural Networks (PGNN): An Application in Lake Temperature Modeling
Arka Daw, Anuj Karpatne, William J. Watkins et al. · arXiv (Cornell University) · 2017 · 316 citations · Full text
Computing, Computer Science, Physics +17