Rethinking the Role of Gradient-based Attribution Methods for Model Interpretability.
Suraj SrinivasFrançois FleuretPublished in: ICLR (2021)
Keyphrases
- computational model
- qualitative and quantitative
- linear regression
- neural network
- hybrid method
- monte carlo simulation
- probabilistic model
- statistical methods
- machine learning methods
- experimental data
- mathematical model
- search engine
- classification models
- classification method
- computational cost
- cost function
- high level
- regression model
- benchmark datasets
- theoretical analysis
- image set
- face recognition
- learning models
- linear model
- social networks
- model free
- predictive model
- learned models