# Towards better understanding of gradient-based attribution methods for Deep Neural Networks

arXiv: https://arxiv.org/abs/1711.06104 · HTML: https://mlsanity.com/paper/1711.06104/

## Key facts

- Cited by 54 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 16 Nov 2017.

## Built on by (most cited first, 30 of 54)

- [From Anecdotal Evidence to Quantitative Evaluation Methods: A Systematic Review on Evaluating Explainable AI](https://mlsanity.com/paper/2201.08164/) (2022)
- [Inspect, Understand, Overcome: A Survey of Practical Methods for AI Safety](https://mlsanity.com/paper/2104.14235/) (2021)
- [First is Better Than Last for Language Data Influence](https://mlsanity.com/paper/2202.11844/) (2022)
- [Interpretable Semantic Photo Geolocation](https://mlsanity.com/paper/2104.14995/) (2021)
- [Disentangled Explanations of Neural Network Predictions by Finding Relevant Subspaces](https://mlsanity.com/paper/2212.14855/) (2022)
- [Peering into the Mind of Language Models: An Approach for Attribution in Contextual Question Answering](https://mlsanity.com/paper/2405.17980/) (2024)
- [Human-Centered Concept Explanations for Neural Networks](https://mlsanity.com/paper/2202.12451/) (2022)
- [LLM-CAS: Dynamic Neuron Perturbation for Real-Time Hallucination Correction](https://mlsanity.com/paper/2512.18623/) (2025)
- [Morphological feature visualization of Alzheimer's disease via Multidirectional Perception GAN](https://mlsanity.com/paper/2111.12886/) (2021)
- [Enhancing Post-Hoc Attributions in Long Document Comprehension via Coarse Grained Answer Decomposition](https://mlsanity.com/paper/2409.17073/) (2024)
- [Saliency Cards: A Framework to Characterize and Compare Saliency Methods](https://mlsanity.com/paper/2206.02958/) (2022)
- [MAEA: Multimodal Attribution for Embodied AI](https://mlsanity.com/paper/2307.13850/) (2023)
- [Sanity checks for patch visualisation in prototype-based image classification](https://mlsanity.com/paper/2311.16120/) (2023)
- [Evaluating Input Perturbation Methods for Interpreting CNNs and Saliency Map Comparison](https://mlsanity.com/paper/2101.10977/) (2021)
- [Explainable Adversarial Attacks in Deep Neural Networks Using Activation Profiles](https://mlsanity.com/paper/2103.10229/) (2021)
- [Threading the Needle of On and Off-Manifold Value Functions for Shapley Explanations](https://mlsanity.com/paper/2202.11919/) (2022)
- [On the explainable properties of 1-Lipschitz Neural Networks: An Optimal Transport Perspective](https://mlsanity.com/paper/2206.06854/) (2022)
- [Neuron Patching: Semantic-based Neuron-level Language Model Repair for Code Generation](https://mlsanity.com/paper/2312.05356/) (2023)
- [Counterfactual Generation with Knockoffs](https://mlsanity.com/paper/2102.00951/) (2021)
- [Promises and pitfalls of deep neural networks in neuroimaging-based psychiatric research](https://mlsanity.com/paper/2301.08525/) (2023)
- [What Sketch Explainability Really Means for Downstream Tasks](https://mlsanity.com/paper/2403.09480/) (2024)
- [BEE: Metric-Adapted Explanations via Baseline Exploration-Exploitation](https://mlsanity.com/paper/2412.17512/) (2024)
- [Detection Accuracy for Evaluating Compositional Explanations of Units](https://mlsanity.com/paper/2109.07804/) (2021)
- [Sanity checks and improvements for patch visualisation in prototype-based image classification](https://mlsanity.com/paper/2302.08508/) (2023)
- [XAI-CLASS: Explanation-Enhanced Text Classification with Extremely Weak Supervision](https://mlsanity.com/paper/2311.00189/) (2023)
- [Towards Unified Attribution in Explainable AI, Data-Centric AI, and Mechanistic Interpretability](https://mlsanity.com/paper/2501.18887/) (2025)
- [Looking deeper into interpretable deep learning in neuroimaging: a comprehensive survey](https://mlsanity.com/paper/2307.09615/) (2023)
- [Challenging the Black Box: A Comprehensive Evaluation of Attribution Maps of CNN Applications in Agriculture and Forestry](https://mlsanity.com/paper/2402.11670/) (2024)
- [The Conformer Encoder May Reverse the Time Dimension](https://mlsanity.com/paper/2410.00680/) (2024)
- [Audio Explanation Synthesis with Generative Foundation Models](https://mlsanity.com/paper/2410.07530/) (2024)

## Abstract (arXiv)

Understanding the flow of information in Deep Neural Networks (DNNs) is a challenging problem that has gain increasing attention over the last few years. While several methods have been proposed to explain network predictions, there have been only a few attempts to compare them from a theoretical perspective. What is more, no exhaustive empirical comparison has been performed in the past. In this work, we analyze four gradient-based attribution methods and formally prove conditions of equivalence and approximation between them. By reformulating two of these methods, we construct a unified framework which enables a direct comparison, as well as an easier implementation. Finally, we propose a novel evaluation metric, called Sensitivity-n and test the gradient-based attribution methods alongside with a simple perturbation-based attribution method on several datasets in the domains of image and text classification, using various network architectures.