On the explanatory power of Boolean decision trees

Audemard, Gilles; Bellart, Steve; Bounia, Louenas; Koriche, Frédéric; Lagniez, Jean-Marie; Marquis, Pierre

doi:10.1016/j.datak.2022.102088

Cited by 20 publications

(10 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Later works showed the same complexity for decision graphs [47] and some classes of tractable circuits [14], [48]. The generation of sufficient reasons for decision trees was also studied in [49], including the generation of shortest sufficient reasons which was shown to be hard even for a single reason. The generation of shortest sufficient reasons was also studied in a broader context that includes decision graphs and SDDs [4].…”

Section: Discussionmentioning

confidence: 98%

“…13 The complexity of shortest sufficient reasons was studied in [55] for Boolean classifiers which correspond to decision graphs and for neural networks with ReLU activation functions. It was further shown that the number of necessary reasons is linear in the decision tree size [4], [47], that all such reasons can be computed in polynomial time [4], [49], and that the shortest necessary reasons can be enumerated with polynomial delay if the classifier satisfies some conditions as stated in [56]. Further complexity results were shown in [14], [48], where classifiers where categorized based on the tractable circuits that represent them [48] or the kinds of processing they permit in polynomial time [14].…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

On Quantifying Literals in Boolean Logic and its Applications to Explainable AI

Darwiche

Marquis²

2021

jair

View full text Add to dashboard Cite

Quantified Boolean logic results from adding operators to Boolean logic for existentially and universally quantifying variables. This extends the reach of Boolean logic by enabling a variety of applications that have been explored over the decades. The existential quantification of literals (variable states) and its applications have also been studied in the literature. In this paper, we complement this by introducing and studying universal literal quantification and its applications, particularly to explainable AI. We also provide a novel semantics for quantification, discuss the interplay between variable/literal and existential/universal quantification, and identify some classes of Boolean formulas and circuits on which quantification can be done efficiently. Literal quantification is more fine-grained than variable quantification as the latter can be defined in terms of the former, leading to a refinement of quantified Boolean logic with literal quantification as its primitive.

show abstract

Section: Discussionmentioning

confidence: 98%

Section: Discussionmentioning

confidence: 99%

On Quantifying Literals in Boolean Logic and its Applications to Explainable AI

Darwiche

Marquis²

2021

jair

View full text Add to dashboard Cite

show abstract

“…Besides the clustering, determining both the impact of various features on the clusters, and their impact on the estimation of the environmental performance is essential to determine the correlation of the features. Decision tree approaches have performed well in providing an explanatory framework for deep learning frameworks for the resulting outputs (Audemard, et al, 2022).…”

Section: Methodsmentioning

confidence: 99%

Environmental Compliance and Financial Performance of Shariah-Compliant Enterprises – a Data-Driven Analysis

Katterbauer¹,

Syed²,

Genç³

et al. 2022

RGSA

View full text Add to dashboard Cite

Purpose: Environmental, social and governance factors have gained significant traction for investors to evaluate their investment decisions and achieve higher impact investments. Societies have had a growing impact on investors to ensure that besides profitability factors, the impact on the environment, society and internal governance criteria are taken into account when allocating funds. This has led to a growing divestment from highly polluting industries and corporations that are not diverse in their workforce, or have a negative impact on the society such as using forced labor. Islamic finance encompasses the principles of Shariah law that put a strong focus on preserving the environment and support the society. Evaluating whether corporations are Shariah compliant with respect to the environment is challenging, as environmental ESG scores may not adequately represent the entire impact on the environment and the Shariah-environment compliance of corporations. Method: This article presents a new data-driven framework for the assessment of Shariah – environmental compliance for corporations in addition to their financial performance. Results and conclusion: An analysis based on large Islamic compliant US listed enterprises indicate strong clustering performance, and differentiation in terms of Islamic environmental compliance versus non-compliance. Originality: The deep learning framework incorporates an unsupervised-random forest learning approach at categorizing environmental compliance while simultaneously estimating the financial performance of these corporations.

show abstract

“…Its objective is to create a model for predicting a given value by learning simple rules deduced from the characteristics of the data. The classification process is applied through a set of rules or conditions that determine the path followed from the root node and ends with one of the final nodes representing the final decision [19]. For all non-final nodes, a decision should be made about the next node.…”

Section: Decision Treementioning

confidence: 99%

“…The decision is to choose a solution from several solutions to a particular problem. Therefore, decision-making is the choice of one of the available alternatives, so The decision-making process is a series of stages and procedures that lead to selecting the best alternative options in the end [19,20].…”

Section: Decision Treementioning

confidence: 99%

Network Traffic Prediction Based on Boosting Learning

Kashmoola

Ahmed

Alsaleem

2022

eijs

View full text Add to dashboard Cite

Classification of network traffic is an important topic for network management, traffic routing, safe traffic discrimination, and better service delivery. Traffic examination is the entire process of examining traffic data, from intercepting traffic data to discovering patterns, relationships, misconfigurations, and anomalies in a network. Between them, traffic classification is a sub-domain of this field, the purpose of which is to classify network traffic into predefined classes such as usual or abnormal traffic and application type. Most Internet applications encrypt data during traffic, and classifying encrypted data during traffic is not possible with traditional methods. Statistical and intelligence methods can find and model traffic patterns that can be categorized based on statistical characteristics. These methods help determine the type of traffic and protect user privacy at the same time. To classify encrypted traffic from end to end, this paper proposes using (XGboost) algorithms, finding the highest parameters using Bayesian optimization, and comparing the proposed model with machine learning algorithms (Nearest Neighbor, Logistic Regression, Decision Trees, Naive Bayes, Multilayer Neural Networks) to classify traffic from end to end. Network traffic has two classifications: whether the traffic is encrypted or not, and the target application. The research results showed the possibility of classifying dual and multiple traffic with high accuracy. The proposed model has a higher classification accuracy than the other models, and finding the optimal parameters increases the model accuracy.

show abstract

On the explanatory power of Boolean decision trees

Cited by 20 publications

References 45 publications

On Quantifying Literals in Boolean Logic and its Applications to Explainable AI

On Quantifying Literals in Boolean Logic and its Applications to Explainable AI

Environmental Compliance and Financial Performance of Shariah-Compliant Enterprises – a Data-Driven Analysis

Network Traffic Prediction Based on Boosting Learning

Contact Info

Product

Resources

About