Explainable Deep Learning: A Field Guide for the Uninitiated
Authors: Gabrielle Ras, Ning Xie, Marcel van Gerven, Derek Doran
JAIR 2022 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Theoretical | This article offers a field guide to explore the space of explainable deep learning for those in the AI/ML field who are uninitiated. The field guide: i) Introduces three simple dimensions defining the space of foundational methods that contribute to explainable deep learning, ii) discusses the evaluations for model explanations, iii) places explainability in the context of other related deep learning research areas, and iv) discusses user-oriented explanation design and future directions. |
| Researcher Affiliation | Collaboration | Gabriëlle Ras EMAIL Donders Institute for Brain, Cognition and Behaviour Radboud University Nijmegen 6525 HR Nijmegen, the Netherlands Ning Xie EMAIL Amazon Seattle, Washington, USA Marcel van Gerven EMAIL Donders Institute for Brain, Cognition and Behaviour Radboud University Nijmegen 6525 HR Nijmegen, the Netherlands Derek Doran EMAIL Tenet3, LLC Dayton, Ohio, USA |
| Pseudocode | No | The paper provides mathematical equations and descriptions of methods, but it does not include any clearly labeled 'Pseudocode' or 'Algorithm' blocks with structured, code-like steps for a specific procedure. For example, methods are described using equations such as arg max X ai,j(X, θ) (1) or φi(f, x) = X |z |!(M |z | 1)! M! [fx(z ) fx(z \ i)] (20), but not in an algorithmic format. |
| Open Source Code | No | The paper is a survey and does not present new methodology for which source code would be released. It mentions third-party code in reference to a visualization tool: "PyTorch CNN visualizations. https://github.com/utkuozbulak/pytorch-cnn-visualizations." but this is not the authors' own code for the paper's main contribution. |
| Open Datasets | No | The paper is a review and does not conduct its own experiments requiring a specific dataset release. While it mentions well-known datasets in the context of discussing other research (e.g., MNIST in Deep LIFT description or Shape Stacks dataset in Model Debugging), it does not provide access information for datasets used in its own work. |
| Dataset Splits | No | The paper is a survey and does not report on original experiments, therefore it does not provide specific training/test/validation dataset splits. Any mention of dataset splits refers to those used in other research works being reviewed, not its own. |
| Hardware Specification | No | The paper is a survey and does not describe experimental results or the hardware used to obtain them. Therefore, no specific hardware specifications are provided. |
| Software Dependencies | No | The paper is a survey and does not describe experimental results or the specific software dependencies with version numbers that would be needed to reproduce its findings. Any software mentioned is in the context of tools or methods reviewed from other research. |
| Experiment Setup | No | The paper is a survey and does not present new experimental results or a specific experimental setup. Therefore, it does not contain details about hyperparameters, training configurations, or system-level settings. |