Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].
The LRP Toolbox for Artificial Neural Networks
Authors: Sebastian Lapuschkin, Alexander Binder, Grégoire Montavon, Klaus-Robert Müller, Wojciech Samek
JMLR 2016 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | With the LRP Toolbox we provide platform-agnostic implementations for explaining the predictions of pre-trained state of the art Caffe networks and stand-alone implementations for fully connected Neural Network models. The implementations for Matlab and python shall serve as a playing field to familiarize oneself with the LRP algorithm and are implemented with readability and transparency in mind. Models and data can be imported and exported using raw text formats, Matlab s .mat files and the .npy format for numpy or plain text. |
| Researcher Affiliation | Academia | Sebastian Lapuschkin EMAIL Fraunhofer Heinrich Hertz Institute, Video Coding and Analytics 10587 Berlin, Germany Alexander Binder EMAIL Singapore University of Technology, ISTD Singapore 487372, Singapore Gr egoire Montavon EMAIL Berlin Institute of Technology, Machine Learning Group 10623 Berlin, Germany Klaus-Robert M uller EMAIL Berlin Institute of Technology, Machine Learning Group 10623 Berlin, Germany Korea University, Department of Brain and Cognitive Engineering Seoul 02841, Korea Wojciech Samek EMAIL Fraunhofer Heinrich Hertz Institute, Video Coding and Analytics 10587 Berlin, Germany |
| Pseudocode | No | The paper describes the LRP algorithm conceptually and its implementation within the toolbox but does not include any structured pseudocode or algorithm blocks. |
| Open Source Code | Yes | The source code is provided under Free BSD (2-Clause) License and is available in a separate archive for each released version. The latest official release of the toolbox code is available from http://heatmapping.org. More recent and work-in-progress versions can be found at https://github.com/sebastian-lapuschkin/lrp_toolbox. |
| Open Datasets | No | The paper introduces a toolbox and describes its functionality, mentioning that it works with "pre-trained state of the art Caffe networks" and that "Models and data can be imported and exported". It also states that the "Caffe version works out of the box with the BVLC reference, the Goog Le Net model and the VGG CNN models from Chatfield et al. (2014)". While these models are typically trained on public datasets (like ImageNet), the paper itself does not provide concrete access information (link, DOI, specific citation) for a dataset it used for its own experimental results or makes available. |
| Dataset Splits | No | The paper introduces a toolbox and does not present new experimental results on a specific dataset. Therefore, no dataset splits are provided. |
| Hardware Specification | No | The paper does not provide specific hardware details (exact GPU/CPU models, processor types with speeds, or memory amounts) used for running its experiments or developing the toolbox. |
| Software Dependencies | Yes | The Caffe version is based on the caffe-master branched on 3rd October 2015. |
| Experiment Setup | No | The paper describes a software toolbox and its usage through demo scripts and configuration files, but it does not provide specific experimental setup details such as hyperparameter values, training configurations, or system-level settings within the main text. These details are indicated to be available in demo files or a manual. |