Machine Learning (ML) systems are a building part of the modern tools which impact our daily life in several application domains. Due to their black-box nature, those systems are hardly adopted in application domains (e.g. health, finance) where understanding the decision process is of paramount importance. Explanation methods were developed to explain how the ML model has taken a specific decision for a given case/instance. Graph Counterfactual Explanations (GCE) is one of the explanation techniques adopted in the Graph Learning domain. The existing works of Graph Counterfactual Explanations diverge mostly in the problem definition, application domain, test data, and evaluation metrics, and most existing works do not compare exhaustively against other counterfactual explanation techniques present in the literature. We present GRETEL, a unified framework to develop and test GCE methods in several settings. GRETEL is a highly extensible evaluation framework which promotes the Open Science and the evaluations reproducibility by providing a set of well-defined mechanisms to integrate and manage easily: both real and synthetic datasets, ML models, state-of-the-art explanation techniques, and evaluation measures. To present GRETEL, we show the experiments conducted to integrate and test several synthetic and real datasets with several existing explanation techniques and base ML models.
相關內容
Various studies among side-channel attacks have tried to extract information through leakages from electronic devices to reach the instruction flow of some appliances. However, previous methods highly depend on the resolution of traced data. Obtaining low-noise traces is not always feasible in real attack scenarios. This study proposes two deep models to extract low and high-level features from side-channel traces and classify them to related instructions. We aim to evaluate the accuracy of a side-channel attack on low-resolution data with a more robust feature extractor thanks to neural networks. As inves-tigated, instruction flow in real programs is predictable and follows specific distributions. This leads to proposing a LSTM model to estimate these distributions, which could expedite the reverse engineering process and also raise the accuracy. The proposed model for leakage classification reaches 54.58% accuracy on average and outperforms other existing methods on our datasets. Also, LSTM model reaches 94.39% accuracy for instruction prediction on standard implementation of cryptographic algorithms.
Citations in science are being studied from several perspectives. On the one hand, there are approaches such as scientometrics and the science of science, which take a more quantitative perspective. In this chapter I briefly review some of the literature on citations, citation distributions and models of citations. These citations feature prominently in another part of the literature which is dealing with research evaluation and the role of metrics and indicators in that process. Here I briefly review part of the discussion in research evaluation. This also touches on the subject of how citations relate to peer review. Finally, I try to integrate the two literatures with the aim of clarifying what I believe the two can learn from each other. The fundamental problem in research evaluation is that research quality is unobservable. This has consequences for conclusions that we can draw from quantitative studies of citations and citation models. The term "indicators" is a relevant concept in this context, which I try to clarify. Causality is important for properly understanding indicators, especially when indicators are used in practice: when we act on indicators, we enter causal territory. Even when an indicator might have been valid, through its very use, the consequences of its use may invalidate it. By combining citation models with proper causal reasoning and acknowledging the fundamental problem about unobservable research quality, we may hope to make progress.
Explaining the decisions of an Artificial Intelligence (AI) model is increasingly critical in many real-world, high-stake applications. Hundreds of papers have either proposed new feature attribution methods, discussed or harnessed these tools in their work. However, despite humans being the target end-users, most attribution methods were only evaluated on proxy automatic-evaluation metrics (Zhang et al. 2018; Zhou et al. 2016; Petsiuk et al. 2018). In this paper, we conduct the first user study to measure attribution map effectiveness in assisting humans in ImageNet classification and Stanford Dogs fine-grained classification, and when an image is natural or adversarial (i.e., contains adversarial perturbations). Overall, feature attribution is surprisingly not more effective than showing humans nearest training-set examples. On a harder task of fine-grained dog categorization, presenting attribution maps to humans does not help, but instead hurts the performance of human-AI teams compared to AI alone. Importantly, we found automatic attribution-map evaluation measures to correlate poorly with the actual human-AI team performance. Our findings encourage the community to rigorously test their methods on the downstream human-in-the-loop applications and to rethink the existing evaluation metrics.
The production, shipping, usage, and disposal of consumer goods have a substantial impact on greenhouse gas emissions and the depletion of resources. Machine Learning (ML) can help to foster sustainable consumption patterns by accounting for sustainability aspects in product search or recommendations of modern retail platforms. However, the lack of large high quality publicly available product data with trustworthy sustainability information impedes the development of ML technology that can help to reach our sustainability goals. Here we present GreenDB, a database that collects products from European online shops on a weekly basis. As proxy for the products' sustainability, it relies on sustainability labels, which are evaluated by experts. The GreenDB schema extends the well-known schema.org Product definition and can be readily integrated into existing product catalogs. We present initial results demonstrating that ML models trained with our data can reliably (F1 score 96%) predict the sustainability label of products. These contributions can help to complement existing e-commerce experiences and ultimately encourage users to more sustainable consumption patterns.
Humans possess an innate ability to identify and differentiate instances that they are not familiar with, by leveraging and adapting the knowledge that they have acquired so far. Importantly, they achieve this without deteriorating the performance on their earlier learning. Inspired by this, we identify and formulate a new, pragmatic problem setting of NCDwF: Novel Class Discovery without Forgetting, which tasks a machine learning model to incrementally discover novel categories of instances from unlabeled data, while maintaining its performance on the previously seen categories. We propose 1) a method to generate pseudo-latent representations which act as a proxy for (no longer available) labeled data, thereby alleviating forgetting, 2) a mutual-information based regularizer which enhances unsupervised discovery of novel classes, and 3) a simple Known Class Identifier which aids generalized inference when the testing data contains instances form both seen and unseen categories. We introduce experimental protocols based on CIFAR-10, CIFAR-100 and ImageNet-1000 to measure the trade-off between knowledge retention and novel class discovery. Our extensive evaluations reveal that existing models catastrophically forget previously seen categories while identifying novel categories, while our method is able to effectively balance between the competing objectives. We hope our work will attract further research into this newly identified pragmatic problem setting.
Structural data well exists in Web applications, such as social networks in social media, citation networks in academic websites, and threads data in online forums. Due to the complex topology, it is difficult to process and make use of the rich information within such data. Graph Neural Networks (GNNs) have shown great advantages on learning representations for structural data. However, the non-transparency of the deep learning models makes it non-trivial to explain and interpret the predictions made by GNNs. Meanwhile, it is also a big challenge to evaluate the GNN explanations, since in many cases, the ground-truth explanations are unavailable. In this paper, we take insights of Counterfactual and Factual (CF^2) reasoning from causal inference theory, to solve both the learning and evaluation problems in explainable GNNs. For generating explanations, we propose a model-agnostic framework by formulating an optimization problem based on both of the two casual perspectives. This distinguishes CF^2 from previous explainable GNNs that only consider one of them. Another contribution of the work is the evaluation of GNN explanations. For quantitatively evaluating the generated explanations without the requirement of ground-truth, we design metrics based on Counterfactual and Factual reasoning to evaluate the necessity and sufficiency of the explanations. Experiments show that no matter ground-truth explanations are available or not, CF^2 generates better explanations than previous state-of-the-art methods on real-world datasets. Moreover, the statistic analysis justifies the correlation between the performance on ground-truth evaluation and our proposed metrics.
We present a novel counterfactual framework for both Zero-Shot Learning (ZSL) and Open-Set Recognition (OSR), whose common challenge is generalizing to the unseen-classes by only training on the seen-classes. Our idea stems from the observation that the generated samples for unseen-classes are often out of the true distribution, which causes severe recognition rate imbalance between the seen-class (high) and unseen-class (low). We show that the key reason is that the generation is not Counterfactual Faithful, and thus we propose a faithful one, whose generation is from the sample-specific counterfactual question: What would the sample look like, if we set its class attribute to a certain class, while keeping its sample attribute unchanged? Thanks to the faithfulness, we can apply the Consistency Rule to perform unseen/seen binary classification, by asking: Would its counterfactual still look like itself? If ``yes'', the sample is from a certain class, and ``no'' otherwise. Through extensive experiments on ZSL and OSR, we demonstrate that our framework effectively mitigates the seen/unseen imbalance and hence significantly improves the overall performance. Note that this framework is orthogonal to existing methods, thus, it can serve as a new baseline to evaluate how ZSL/OSR models generalize. Codes are available at //github.com/yue-zhongqi/gcm-cf.
Machine learning plays a role in many deployed decision systems, often in ways that are difficult or impossible to understand by human stakeholders. Explaining, in a human-understandable way, the relationship between the input and output of machine learning models is essential to the development of trustworthy machine-learning-based systems. A burgeoning body of research seeks to define the goals and methods of explainability in machine learning. In this paper, we seek to review and categorize research on counterfactual explanations, a specific class of explanation that provides a link between what could have happened had input to a model been changed in a particular way. Modern approaches to counterfactual explainability in machine learning draw connections to the established legal doctrine in many countries, making them appealing to fielded systems in high-impact areas such as finance and healthcare. Thus, we design a rubric with desirable properties of counterfactual explanation algorithms and comprehensively evaluate all currently-proposed algorithms against that rubric. Our rubric provides easy comparison and comprehension of the advantages and disadvantages of different approaches and serves as an introduction to major research themes in this field. We also identify gaps and discuss promising research directions in the space of counterfactual explainability.
We examine the problem of question answering over knowledge graphs, focusing on simple questions that can be answered by the lookup of a single fact. Adopting a straightforward decomposition of the problem into entity detection, entity linking, relation prediction, and evidence combination, we explore simple yet strong baselines. On the popular SimpleQuestions dataset, we find that basic LSTMs and GRUs plus a few heuristics yield accuracies that approach the state of the art, and techniques that do not use neural networks also perform reasonably well. These results show that gains from sophisticated deep learning techniques proposed in the literature are quite modest and that some previous models exhibit unnecessary complexity.
Visual Question Answering (VQA) models have struggled with counting objects in natural images so far. We identify a fundamental problem due to soft attention in these models as a cause. To circumvent this problem, we propose a neural network component that allows robust counting from object proposals. Experiments on a toy task show the effectiveness of this component and we obtain state-of-the-art accuracy on the number category of the VQA v2 dataset without negatively affecting other categories, even outperforming ensemble models with our single model. On a difficult balanced pair metric, the component gives a substantial improvement in counting over a strong baseline by 6.6%.
小貼士
登錄享
相關主題
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191