In an era where artificial intelligence and machine learning algorithms increasingly impact human life, it is crucial to develop models that account for potential discrimination in their predictions. This paper tackles this problem by introducing a new classification tree algorithm using a novel splitting criterion that incorporates fairness adjustments into the tree-building process. The proposed method integrates a fairness-aware impurity measure that balances predictive accuracy with fairness across protected groups. By ensuring that each splitting node considers both the gain in classification error and the fairness, our algorithm encourages splits that mitigate discrimination. Importantly, in penalizing unfair splits, we account for the uncertainty in the fairness metric by utilizing its confidence interval instead of relying on its point estimate. Experimental results on benchmark and synthetic datasets illustrate that our method effectively reduces discriminatory predictions compared to traditional classification trees, without significant loss in overall accuracy.

The paper presents two nice ways for improving the usual gradient boosting algorithm where weak classifiers are decision trees. It is a paper oriented towards efficient (less costful) implementation of the usual algorithm in order to speed up the learning of decision trees by taking into account previous computations and sparse data. The approaches are interesting and smart. A risk bound is given for one of the improvements (GOSS), which seems sound but still quite loose: according to the experiments, a tighter bound could be obtained, getting rid of the "max" sizes of considered sets. No garantee is given for the second improvement (EFB) although is seems to be quite efficient in practice.

The article tackles the problem of variable importance in regression trees. The strategy is to select the variables based on the impurity reduction they induce on label Y. The main feature of this strategy is that the impurity reduction measure is based on the ordering of Y according to the ranking of the X variable under consideration, therefore it measures the relationship between Y and any variable in a more robust way than simple correlation would. The authors prove that this strategy is consistent (i.e. the true explanatory variables are selected) in a range of settings. This is then illustrated on a simulated example where the results displayed are somewhat the ones one could have expected: the proposed procedure is able to account for monotone but non linear relationships between X and Y so it yields better results than simple correlations.

The authors of this paper present a new decision tree algorithm for the interval regression problem. Leaves are partitioned using a margin based hinge loss similar to the L1-regularized hinge loss in Rigaill et al, Proc ICML 2013. However, the regression tree algorithm presented in this work is not limited to modeling linear patterns as the L1-regularized linear models in Rigaill et al. For training the non linear tree model, a sequence of convex optimization subproblems are optimally solved in log-linear time by Dynamic Programming (DP). The new maximum margin interval tree (MMIT) algorithm is compared with state-of-the-art margin-based and non-margin-based methods in several real and simulated datasets.

Their method requires an initial decision tree. The topology of this tree will be fixed, and only the decision rules at each node will be adjusted. The idea behind the proposed adjustment is based on the observation that, fixing all of the parameters of all the nodes except the parameters of node i, the likelihood function for the whole tree reduces to the likelihood function of a simple K-classes classifier. This simple classifier can be trained efficiently (using existing techniques) and doing so will always guarantee that the overall loss will decrease when compared to the loss for the initial decision tree.

Decision trees are powerful for predictive modeling but often suffer from high variance when modeling continuous relationships. While algorithms like Multivariate Adaptive Regression Splines (MARS) excel at capturing such continuous relationships, they perform poorly when modeling discontinuities. To address the limitations of both approaches, we introduce Spline-based Multivariate Adaptive Regression Trees (SMART), which uses a decision tree to identify subsets of data with distinct continuous relationships and then leverages MARS to fit these relationships independently. Unlike other methods that rely on the tree structure to model interaction and higher-order terms, SMART leverages MARS's native ability to handle these terms, allowing the tree to focus solely on identifying discontinuities in the relationship. We test SMART on various datasets, demonstrating its improvement over state-of-the-art methods in such cases. Additionally, we provide an open-source implementation of our method to be used by practitioners.

Adversarial attacks are major threats to the deployment of machine learning (ML) models in many applications. Testing ML models against such attacks is becoming an essential step for evaluating and improving ML models. In this paper, we report the design and development of an interactive system for aiding the workflow of Testing Against Adversarial Attacks (TA3). In particular, with TA3, human-in-the-loop (HITL) enables human-steered attack simulation and visualization-assisted attack impact evaluation. While the current version of TA3 focuses on testing decision tree models against adversarial attacks based on the One Pixel Attack Method, it demonstrates the importance of HITL in ML testing and the potential application of HITL to the ML testing workflows for other types of ML models and other types of adversarial attacks.

Protecting the intellectual property of machine learning models is a hot topic and many watermarking schemes for deep neural networks have been proposed in the literature. Unfortunately, prior work largely neglected the investigation of watermarking techniques for other types of models, including decision tree ensembles, which are a state-of-the-art model for classification tasks on non-perceptual data. In this paper, we present the first watermarking scheme designed for decision tree ensembles, focusing in particular on random forest models. We discuss watermark creation and verification, presenting a thorough security analysis with respect to possible attacks. We finally perform an experimental evaluation of the proposed scheme, showing excellent results in terms of accuracy and security against the most relevant threats.

The research investigates how the application of a machine-learning random forest model improves the accuracy and precision of a Delphi model. The context of the research is Azerbaijani SMEs and the data for the study has been obtained from a financial institution which had gathered it from the enterprises (as there is no public data on local SMEs, it was not practical to verify the data independently). The research used accuracy, precision, recall and F-1 scores for both models to compare them and run the algorithms in Python. The findings showed that accuracy, precision, recall and F- 1 all improve considerably (from 0.69 to 0.83, from 0.65 to 0.81, from 0.56 to 0.77 and from 0.58 to 0.79, respectively). The implications are that by applying AI models in credit risk modeling, financial institutions can improve the accuracy of identifying potential defaulters which would reduce their credit risk. In addition, an unfair rejection of credit access for SMEs would also go down having a significant contribution to an economic growth in the economy. Finally, such ethical issues as transparency of algorithms and biases in historical data should be taken on board while making decisions based on AI algorithms in order to reduce mechanical dependence on algorithms that cannot be justified in practice.

Random forests utilize bootstrap sampling to create an individual training set for each component tree. This involves sampling with replacement, with the number of instances equal to the size of the original training set (N). Research literature indicates that drawing fewer than N observations can also yield satisfactory results. The ratio of the number of observations in each bootstrap sample to the total number of training instances is called the bootstrap rate (BR). Sampling more than N observations (BR > 1) has been explored in the literature only to a limited extent and has generally proven ineffective. In this paper, we re-examine this approach using 36 diverse datasets and consider BR values ranging from 1.2 to 5.0. Contrary to previous findings, we show that such parameterization can result in statistically significant improvements in classification accuracy compared to standard settings (BR 1). Furthermore, we investigate what the optimal BR depends on and conclude that it is more a property of the dataset than a dependence on the random forest hyperparameters. Finally, we develop a binary classifier to predict whether the optimal BR is 1 or > 1 for a given dataset, achieving between 81.88% and 88.81% accuracy, depending on the experiment configuration. Random forest (RF) algorithm, introduced by Breiman (2001), is an ensemble of decision trees (DTs) that collectively make decisions using either majority or soft voting. RF reduces variance, sometimes at the cost of slightly increasing bias, by introducing two sources of randomness.