Unsupervised Accuracy Estimation of Deep Visual Models using Domain-Adaptive Adversarial Perturbation without Source Samples.
JoonHo LeeJae Oh WooHankyu MoonKwonho LeePublished in: CoRR (2023)
Keyphrases
- data driven
- unsupervised learning
- high accuracy
- data sets
- statistical models
- domain independent
- complex systems
- domain specific
- semi supervised
- domain models
- probabilistic model
- parameter estimation
- models built
- deep learning
- conceptual models
- estimation algorithm
- parametric models
- accurate models
- unseen data
- deep architectures
- estimation accuracy
- monte carlo simulation
- highly accurate
- computational models
- prediction accuracy
- domain knowledge
- multi agent
- high level
- neural network