Prior knowledge meets Neural ODEs: a two-stage training method for improved explainability.
C. CoelhoM. Fernanda P. CostaLuis L. FerrásPublished in: Tiny Papers @ ICLR (2023)
Keyphrases
- prior knowledge
- significant improvement
- synthetic data
- prior information
- cost function
- feature set
- dynamic programming
- detection method
- clustering method
- improved algorithm
- high accuracy
- computational cost
- experimental evaluation
- artificial neural networks
- objective function
- similarity measure
- classification method
- high precision
- training phase
- network architecture
- data sets
- mathematical model
- back propagation
- training samples
- model selection
- computational complexity
- decision trees