Rethinking the Role of Gradient-based Attribution Methods for Model Interpretability

Authors: Suraj Srinivas, Francois Fleuret

ICLR 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our experiments show that improving the alignment of the implicit density model with the data distribution enhances gradient structure and explanatory power while reducing this alignment has the opposite effect.
Researcher Affiliation Academia Suraj Srinivas Idiap Research Institute & EPFL suraj.srinivas@idiap.ch Franc ois Fleuret University of Geneva francois.fleuret@unige.ch
Pseudocode No The paper describes algorithms and approximations but does not include any clearly labeled 'Pseudocode' or 'Algorithm' blocks.
Open Source Code No The paper does not contain any explicit statement about releasing source code for the methodology or a link to a code repository.
Open Datasets Yes For experiments, we shall consider the CIFAR100 dataset. We present experiments with CIFAR10 in the supplementary section.
Dataset Splits No The paper refers to a 'test set' but does not specify the full train/validation/test dataset splits, percentages, or methodology used for partitioning the data.
Hardware Specification No The paper mentions running experiments but does not provide specific hardware details such as GPU models, CPU models, or memory specifications.
Software Dependencies No The paper does not list specific software dependencies with version numbers, such as Python versions, deep learning frameworks (e.g., PyTorch, TensorFlow) with their versions, or other libraries.
Experiment Setup Yes Unless stated otherwise, the network structure we use shall be a 18-layer Res Net... and the optimizer used shall be SGD with momentum. All models use the softplus non-linearity with β = 10... For this, we use a regularization constant λ = 1e 3. ... We use a threshold of τ = 1000, and regularization constant λ = 1e 4.