arXiv research

Saliency Map, the gradient of the score function with respect to the input, is the most basic technique for interpreting deep neural network decisions. However, saliency maps are often visually noisy. Although several hypotheses were proposed to account for this phenomenon, there are few works that provide rigorous ana…

Saliency maps are a popular approach to creating post-hoc explanations of image classifier outputs. These methods produce estimates of the relevance of each pixel to the classification output score, which can be displayed as a saliency map that highlights important pixels. Despite a proliferation of such methods, littl…

New nonlinear saliency maps improve deep neural network interpretability.

Evaluating, explaining, and visualizing high-level concepts in generative models, such as variational autoencoders (VAEs), is challenging in part due to a lack of known prediction classes that are required to generate saliency maps in supervised learning. While saliency maps may help identify relevant features (e.g., p…

Study validates saliency maps of GNNs using selective inference.

New method improves interpretability of fMRI decoding models.

Study evaluates saliency maps on artificial data with different backgrounds.

Saliency maps are often used in computer vision to provide intuitive interpretations of what input regions a model has used to produce a specific prediction. A number of approaches to saliency map generation are available, but most require access to model parameters. This work proposes an approach for saliency map gene…

Saliency methods can make deep neural network predictions more interpretable by identifying a set of critical features in an input sample, such as pixels that contribute most strongly to a prediction made by an image classifier. Unfortunately, recent evidence suggests that many saliency methods poorly perform, especial…

1-Lipschitz neural networks produce clearer, more focused Saliency Maps for explainable AI.

Enhances counterfactual explanations with more valid and informative saliency maps.

Proposes a method to quantify the reliability of salient regions in deep learning models using p-values.

Recent studies on the adversarial vulnerability of neural networks have shown that models trained to be more robust to adversarial attacks exhibit more interpretable saliency maps than their non-robust counterparts. We aim to quantify this behavior by considering the alignment between input image and saliency map. We h…

Currently available methods for extracting saliency maps identify parts of the input which are the most important to a specific fixed classifier. We show that this strong dependence on a given classifier hinders their performance. To address this problem, we propose classifier-agnostic saliency map extraction, which fi…

With recent advances in deep learning, neuroimaging studies increasingly rely on convolutional networks (ConvNets) to predict diagnosis based on MR images. To gain a better understanding of how a disease impacts the brain, the studies visualize the salience maps of the ConvNet highlighting voxels within the brain major…

In this work we develop a fast saliency detection method that can be applied to any differentiable image classifier. We train a masking model to manipulate the scores of the classifier by masking salient parts of the input image. Our model generalises well to unseen images and requires a single forward pass to perform …

Saliency methods can aid understanding of deep neural networks. Recent years have witnessed many improvements to saliency methods, as well as new ways for evaluating them. In this paper, we 1) present a novel region-based attribution method, XRAI, that builds upon integrated gradients (Sundararajan et al. 2017), 2) int…

Combining global and local explanations improves user understanding of RL agents.

New saliency evaluations focus on completeness and soundness, improving explanations.

This article reports on an investigation of the use of convolutional neural networks to predict the visual attention of chess players. The visual attention model described in this article has been created to generate saliency maps that capture hierarchical and spatial features of chessboard, in order to predict the pro…

We introduce a new tool for interpreting neural net responses, namely full-gradients, which decomposes the neural net response into input sensitivity and per-neuron sensitivity components. This is the first proposed representation which satisfies two key properties: completeness and weak dependence, which provably cann…

The paper benchmarks saliency methods for time series predictions.

A novel method using saliency maps detects adversarial attacks effectively.

Improved saliency maps for deep neural networks with reduced noise.

There is great interest in "saliency methods" (also called "attribution methods"), which give "explanations" for a deep net's decision, by assigning a "score" to each feature/pixel in the input. Their design usually involves credit-assignment via the gradient of the output with respect to input. Recently Adebayo et al.…

Recent efforts to improve the interpretability of deep neural networks use saliency to characterize the importance of input features to predictions made by models. Work on interpretability using saliency-based methods on Recurrent Neural Networks (RNNs) has mostly targeted language tasks, and their applicability to tim…

Deep learning models are now used in many different industries, while in certain domains safety is not a critical issue in the medical field it is a huge concern. Not only, we want the models to generalize well but we also want to know the models confidence respect to its decision and which features matter the most. Ou…

While machine learning systems show high success rate in many complex tasks, research shows they can also fail in very unexpected situations. Rise of machine learning products in safety-critical industries cause an increase in attention in evaluating model robustness and estimating failure probability in machine learni…

CRITS improves time series classification with interpretable local explanations.

Saliency methods have emerged as a popular tool to highlight features in an input deemed relevant for the prediction of a learned model. Several saliency methods have been proposed, often guided by visual appeal on image data. In this work, we propose an actionable methodology to evaluate what kinds of explanations a g…

The Jacobian-based Saliency Map Attack is a family of adversarial attack methods for fooling classification models, such as deep neural networks for image classification tasks. By saturating a few pixels in a given image to their maximum or minimum values, JSMA can cause the model to misclassify the resulting adversari…

SmoothGrad and VarGrad are techniques that enhance the empirical quality of standard saliency maps by adding noise to input. However, there were few works that provide a rigorous theoretical interpretation of those methods. We analytically formalize the result of these noise-adding methods. As a result, we observe two …

A method to combine saliency metrics for better CNN pruning decisions.

Study identifies and mitigates causes of image misclassifications in CNN models.

Deep learning interpretation is essential to explain the reasoning behind model predictions. Understanding the robustness of interpretation methods is important especially in sensitive domains such as medical applications since interpretation results are often used in downstream tasks. Although gradient-based saliency …

Deep learning models designed for visual classification tasks on natural images have become prevalent in medical image analysis. However, medical images differ from typical natural images in many ways, such as significantly higher resolutions and smaller regions of interest. Moreover, both the global structure and loca…

Convolutional-deconvolution networks can be adopted to perform end-to-end saliency detection. But, they do not work well with objects of multiple scales. To overcome such a limitation, in this work, we propose a recurrent attentional convolutional-deconvolution network (RACDNN). Using spatial transformer and recurrent …

Incorporating human domain knowledge for breast tumor diagnosis is challenging, since shape, boundary, curvature, intensity, or other common medical priors vary significantly across patients and cannot be employed. This work proposes a new approach for integrating visual saliency into a deep learning model for breast t…

Electron Cryo-Tomography (ECT) allows 3D visualization of subcellular structures at the submolecular resolution in close to the native state. However, due to the high degree of structural complexity and imaging limits, the automatic segmentation of cellular components from ECT images is very difficult. To complement an…

Two modified tests improve the reliability of evaluating explanation methods.

Novel method combines neural network features with survival models for ICU infections.

Study evaluates different saliency maps for CT image classification.

Pruning unimportant parameters can allow deep neural networks (DNNs) to reduce their heavy computation and memory requirements. A saliency metric estimates which parameters can be safely pruned with little impact on the classification performance of the DNN. Many saliency metrics have been proposed, each within the con…

Deep convolutional neural networks have become a key element in the recent breakthrough of salient object detection. However, existing CNN-based methods are based on either patch-wise (region-wise) training and inference or fully convolutional networks. Methods in the former category are generally time-consuming due to…

In this paper we propose a novel method for detecting adversarial examples by training a binary classifier with both origin data and saliency data. In the case of image classification model, saliency simply explain how the model make decisions by identifying significant pixels for prediction. A model shows wrong classi…

GGA improves untrustworthy prediction detection in neural networks without retraining.

A new LDA variant improves multi-label classification performance.

New approach makes adversarial examples less suspicious without changing perceptual salience.