Posted Content•
Towards Ground Truth Evaluation of Visual Explanations
16 Mar 2020-
TL;DR: This work generates a CLEVR-alike visual question answering benchmark with around 40,000 questions, and introduces two straightforward metrics to evaluate explanations in this setup, and compares their outcomes to standard pixel perturbation using a Relation Network model and three decomposition-based explanation methods.
Abstract: Several methods have been proposed to explain the decisions of neural networks in the visual domain via saliency heatmaps (aka relevances/feature importance scores). Thus far, these methods were mainly validated on real-world images, using either pixel perturbation experiments or bounding box localization accuracies. In the present work, we propose instead to evaluate explanations in a restricted and controlled setup using a synthetic dataset of rendered 3D shapes. To this end, we generate a CLEVR-alike visual question answering benchmark with around 40,000 questions, where the ground truth pixel coordinates of relevant objects are known, which allows us to validate explanations in a fair and transparent way. We further introduce two straightforward metrics to evaluate explanations in this setup, and compare their outcomes to standard pixel perturbation using a Relation Network model and three decomposition-based explanation methods: Gradient x Input, Integrated Gradients and Layer-wise Relevance Propagation. Among the tested methods, Layer-wise Relevance Propagation was shown to perform best, followed by Integrated Gradients. More generally, we expect the release of our dataset and code to support the development and comparison of methods on a well-defined common ground.
Citations
More filters
Posted Content•
TL;DR: This work aims to provide a timely overview of this active emerging field of machine learning and explain its theoretical foundations, put interpretability algorithms to a test both from a theory and comparative evaluation perspective using extensive simulations, and outline best practice aspects.
Abstract: With the broader and highly successful usage of machine learning in industry and the sciences, there has been a growing demand for explainable AI. Interpretability and explanation methods for gaining a better understanding about the problem solving abilities and strategies of nonlinear Machine Learning such as Deep Learning (DL), LSTMs, and kernel methods are therefore receiving increased attention. In this work we aim to (1) provide a timely overview of this active emerging field and explain its theoretical foundations, (2) put interpretability algorithms to a test both from a theory and comparative evaluation perspective using extensive simulations, (3) outline best practice aspects i.e. how to best include interpretation methods into the standard usage of machine learning and (4) demonstrate successful usage of explainable AI in a representative selection of application scenarios. Finally, we discuss challenges and possible future directions of this exciting foundational field of machine learning.
75 citations
Proceedings Article•
01 Jan 2020
TL;DR: This work adapt commonly-used attribution methods for GNNs and quantitatively evaluate them using the axes of attribution accuracy, stability, faithfulness and consistency, and makes concrete recommendations for which attribution methods to use.
Abstract: Interpretability of machine learning models is critical to scientific understanding, AI safety, and debugging. Attribution is one approach to interpretability, which highlights input dimensions that are influential to a neural network’s prediction. Evaluation of these methods is largely qualitative for image and text models, because acquiring ground truth attributions requires expensive and unreliable human judgment. Attribution has been comparatively understudied for graph neural networks (GNNs), a model class of growing importance that makes predictions on arbitrarily-sized graphs. Graph-valued data offer an opportunity to quantitatively benchmark attribution methods, because challenging synthetic graph problems have computable ground-truth attributions. In this work we adapt commonly-used attribution methods for GNNs and quantitatively evaluate them using the axes of attribution accuracy, stability, faithfulness and consistency. We make concrete recommendations for which attribution methods to use, and provide the data and code for our benchmarking suite. Rigorous and open source benchmarking of attribution methods in graphs could enable new methods development and broader use of attribution in real-world ML tasks.
51 citations
Cites background from "Towards Ground Truth Evaluation of ..."
...Efforts to quantify the utility of attribution methods or apply sanity checks have been undertaken in input domains where human intuition is usually used to evaluate attribution quality [4, 50, 7, 32, 21, 5], like images and text....
[...]
Posted Content•
TL;DR: Results indicate that several deep learning models, and in particular WILDCAT and deep MIL can provide a high level of classification accuracy, although pixel-wise localization of cancer regions remains an issue for such images.
Abstract: Using state-of-the-art deep learning models for cancer diagnosis presents several challenges related to the nature and availability of labeled histology images. In particular, cancer grading and localization in these images normally relies on both image- and pixel-level labels, the latter requiring a costly annotation process. In this survey, deep weakly-supervised learning (WSL) models are investigated to identify and locate diseases in histology images, without the need for pixel-level annotations. Given training data with global image-level labels, these models allow to simultaneously classify histology images and yield pixel-wise localization scores, thereby identifying the corresponding regions of interest (ROI). Since relevant WSL models have mainly been investigated within the computer vision community, and validated on natural scene images, we assess the extent to which they apply to histology images which have challenging properties, e.g. very large size, similarity between foreground/background, highly unstructured regions, stain heterogeneity, and noisy/ambiguous labels. The most relevant models for deep WSL are compared experimentally in terms of accuracy (classification and pixel-wise localization) on several public benchmark histology datasets for breast and colon cancer -- BACH ICIAR 2018, BreaKHis, CAMELYON16, and GlaS. Furthermore, for large-scale evaluation of WSL models on histology images, we propose a protocol to construct WSL datasets from Whole Slide Imaging. Results indicate that several deep learning models can provide a high level of classification accuracy, although accurate pixel-wise localization of cancer regions remains an issue for such images. Code is publicly available.
48 citations
Posted Content•
TL;DR: High uncertainty is introduced as a criterion to localize non-discriminative regions that do not affect classifier decision, and is described with original Kullback-Leibler (KL) divergence losses evaluating the deviation of posterior predictions from the uniform distribution.
Abstract: Weakly supervised learning (WSL) has recently triggered substantial interest as it mitigates the lack of pixel-wise annotations, while enabling interpretable models. Given global image labels, WSL methods yield pixel-level predictions (segmentations). Despite their recent success, mostly with natural images, such methods could be seriously challenged when the foreground and background regions have similar visual cues, yielding high false-positive rates in segmentations, as is the case of challenging histology images. WSL training is commonly driven by standard classification losses, which implicitly maximize model confidence and find the discriminative regions linked to classification decisions. Therefore, they lack mechanisms for modeling explicitly non-discriminative regions and reducing false-positive rates. We propose new regularization terms, which enable the model to seek both non-discriminative and discriminative regions, while discouraging unbalanced segmentations. We introduce high uncertainty as a criterion to localize non-discriminative regions that do not affect classifier decision, and describe it with original Kullback-Leibler (KL) divergence losses evaluating the deviation of posterior predictions from the uniform distribution. Our KL terms encourage high uncertainty of the model when the latter takes the latent non-discriminative regions as input. Our loss integrates: (i) a cross-entropy seeking a foreground, where model confidence about class prediction is high; (ii) a KL regularizer seeking a background, where model uncertainty is high; and (iii) log-barrier terms discouraging unbalanced segmentations. Comprehensive experiments and ablation studies over the public GlaS colon cancer data show substantial improvements over state-of-the-art WSL methods, and confirm the effect of our new regularizers. Our code is publicly available.
34 citations
Cites background from "Towards Ground Truth Evaluation of ..."
...It is worth noting that such interpretability aspects are also attracting wide interest in computer vision (Bach et al., 2015; Bau et al., 2017; Bhatt et al., 2020; Dabkowski and Gal, 2017; Escalante et al., 2018; Fong et al., 2019; Fong and Vedaldi, 2017; Goh et al., 2020; Osman et al., 2020; Murdoch et al., 2019; Petsiuk et al., 2020; 2018; Ribeiro et al., 2016; Samek et al., 2020; 2017; Zhang et al., 2020; Belharbi et al., 2021) and medical imaging (de La Torre et al....
[...]
Posted Content•
TL;DR: This paper develops a procedure for learning an aggregate explanation function with lower complexity and then derive a new aggregate Shapley value explanation function that minimizes sensitivity.
Abstract: A feature-based model explanation denotes how much each input feature contributes to a model's output for a given data point. As the number of proposed explanation functions grows, we lack quantitative evaluation criteria to help practitioners know when to use which explanation function. This paper proposes quantitative evaluation criteria for feature-based explanations: low sensitivity, high faithfulness, and low complexity. We devise a framework for aggregating explanation functions. We develop a procedure for learning an aggregate explanation function with lower complexity and then derive a new aggregate Shapley value explanation function that minimizes sensitivity.
25 citations
Cites methods from "Towards Ground Truth Evaluation of ..."
...Note we omit evaluation criteria that assume access to ground-truth explanations for training points; for a thorough treatment on this topic, see [Hind et al., 2019; Osman et al., 2020]....
[...]
References
More filters
TL;DR: The ImageNet Large Scale Visual Recognition Challenge (ILSVRC) as mentioned in this paper is a benchmark in object category classification and detection on hundreds of object categories and millions of images, which has been run annually from 2010 to present, attracting participation from more than fifty institutions.
Abstract: The ImageNet Large Scale Visual Recognition Challenge is a benchmark in object category classification and detection on hundreds of object categories and millions of images. The challenge has been run annually from 2010 to present, attracting participation from more than fifty institutions. This paper describes the creation of this benchmark dataset and the advances in object recognition that have been possible as a result. We discuss the challenges of collecting large-scale ground truth annotation, highlight key breakthroughs in categorical object recognition, provide a detailed analysis of the current state of the field of large-scale image classification and object detection, and compare the state-of-the-art computer vision accuracy with human accuracy. We conclude with lessons learned in the 5 years of the challenge, and propose future directions and improvements.
30,811 citations
11 Oct 2018
TL;DR: BERT as mentioned in this paper pre-trains deep bidirectional representations from unlabeled text by jointly conditioning on both left and right context in all layers, which can be fine-tuned with just one additional output layer to create state-of-the-art models for a wide range of tasks.
Abstract: We introduce a new language representation model called BERT, which stands for Bidirectional Encoder Representations from Transformers. Unlike recent language representation models (Peters et al., 2018a; Radford et al., 2018), BERT is designed to pre-train deep bidirectional representations from unlabeled text by jointly conditioning on both left and right context in all layers. As a result, the pre-trained BERT model can be fine-tuned with just one additional output layer to create state-of-the-art models for a wide range of tasks, such as question answering and language inference, without substantial task-specific architecture modifications. BERT is conceptually simple and empirically powerful. It obtains new state-of-the-art results on eleven natural language processing tasks, including pushing the GLUE score to 80.5 (7.7 point absolute improvement), MultiNLI accuracy to 86.7% (4.6% absolute improvement), SQuAD v1.1 question answering Test F1 to 93.2 (1.5 point absolute improvement) and SQuAD v2.0 Test F1 to 83.1 (5.1 point absolute improvement).
24,672 citations
06 Sep 2014
TL;DR: A novel visualization technique is introduced that gives insight into the function of intermediate feature layers and the operation of the classifier in large Convolutional Network models, used in a diagnostic role to find model architectures that outperform Krizhevsky et al on the ImageNet classification benchmark.
Abstract: Large Convolutional Network models have recently demonstrated impressive classification performance on the ImageNet benchmark Krizhevsky et al. [18]. However there is no clear understanding of why they perform so well, or how they might be improved. In this paper we explore both issues. We introduce a novel visualization technique that gives insight into the function of intermediate feature layers and the operation of the classifier. Used in a diagnostic role, these visualizations allow us to find model architectures that outperform Krizhevsky et al on the ImageNet classification benchmark. We also perform an ablation study to discover the performance contribution from different model layers. We show our ImageNet model generalizes well to other datasets: when the softmax classifier is retrained, it convincingly beats the current state-of-the-art results on Caltech-101 and Caltech-256 datasets.
12,783 citations
"Towards Ground Truth Evaluation of ..." refers methods in this paper
...Methods that provide such heatmaps in a direct and unambiguous way include, amongst others, Class Saliency Map [21], Occlusion [24], Gradient × Input, Integrated Gradients [23], Layer-wise Relevance Propagation [6], Excitation Backpropagation [26], Guided Backpropagation [22]....
[...]
13 Aug 2016
TL;DR: In this article, the authors propose LIME, a method to explain models by presenting representative individual predictions and their explanations in a non-redundant way, framing the task as a submodular optimization problem.
Abstract: Despite widespread adoption, machine learning models remain mostly black boxes. Understanding the reasons behind predictions is, however, quite important in assessing trust, which is fundamental if one plans to take action based on a prediction, or when choosing whether to deploy a new model. Such understanding also provides insights into the model, which can be used to transform an untrustworthy model or prediction into a trustworthy one. In this work, we propose LIME, a novel explanation technique that explains the predictions of any classifier in an interpretable and faithful manner, by learning an interpretable model locally varound the prediction. We also propose a method to explain models by presenting representative individual predictions and their explanations in a non-redundant way, framing the task as a submodular optimization problem. We demonstrate the flexibility of these methods by explaining different models for text (e.g. random forests) and image classification (e.g. neural networks). We show the utility of explanations via novel experiments, both simulated and with human subjects, on various scenarios that require trust: deciding if one should trust a prediction, choosing between models, improving an untrustworthy classifier, and identifying why a classifier should not be trusted.
11,104 citations
TL;DR: This work demonstrates an artificial intelligence capable of classifying skin cancer with a level of competence comparable to dermatologists, trained end-to-end from images directly, using only pixels and disease labels as inputs.
Abstract: Skin cancer, the most common human malignancy, is primarily diagnosed visually, beginning with an initial clinical screening and followed potentially by dermoscopic analysis, a biopsy and histopathological examination. Automated classification of skin lesions using images is a challenging task owing to the fine-grained variability in the appearance of skin lesions. Deep convolutional neural networks (CNNs) show potential for general and highly variable tasks across many fine-grained object categories. Here we demonstrate classification of skin lesions using a single CNN, trained end-to-end from images directly, using only pixels and disease labels as inputs. We train a CNN using a dataset of 129,450 clinical images-two orders of magnitude larger than previous datasets-consisting of 2,032 different diseases. We test its performance against 21 board-certified dermatologists on biopsy-proven clinical images with two critical binary classification use cases: keratinocyte carcinomas versus benign seborrheic keratoses; and malignant melanomas versus benign nevi. The first case represents the identification of the most common cancers, the second represents the identification of the deadliest skin cancer. The CNN achieves performance on par with all tested experts across both tasks, demonstrating an artificial intelligence capable of classifying skin cancer with a level of competence comparable to dermatologists. Outfitted with deep neural networks, mobile devices can potentially extend the reach of dermatologists outside of the clinic. It is projected that 6.3 billion smartphone subscriptions will exist by the year 2021 (ref. 13) and can therefore potentially provide low-cost universal access to vital diagnostic care.
8,424 citations