Never Train from Scratch: Fair Comparison of Long-Sequence Models Requires Data-Driven Priors.

Ido Amos Jonathan Berant Ankit Gupta

Published in: ICLR (2024)

Keyphrases

data driven
probabilistic model
bayesian framework
complex systems
bayesian models
statistical models
statistical model
neural network
pairwise
real time
three dimensional
genetic algorithm
prior information
computational models
parametric models
modeling framework
databases