arXiv subjects

Lev V. Utkin

Publications and source records attributed to Lev V. Utkin.

At least 37 records · Page 2Linked to original sources

arXiv2023

Neural Attention Forests: Transformer-Based Forest Improvement

A new approach called NAF (the Neural Attention Forest) for solving regression and classification tasks under tabular training data is proposed. The main idea behind the proposed NAF model is to introduce the attention mechanism into the random forest by assigning attention weights calculated by neural networks of a specific form to data in leaves of decision trees and to the random forest itself in the framework of the Nadaraya-Watson kernel regression. In contrast to the available models like the attention-based random forest, the attention weights and the Nadaraya-Watson regression are represented in the form of neural networks whose weights can be regarded as trainable parameters. The first part of neural networks with shared weights is trained for all trees and computes attention weights of data in leaves. The second part aggregates outputs of the tree networks and aims to minimize the difference between the random forest prediction and the truth target value from a training set. The neural network is trained in an end-to-end manner. The combination of the random forest and neural networks implementing the attention mechanism forms a transformer for enhancing the forest predictions. Numerical experiments with real datasets illustrate the proposed method. The code implementing the approach is publicly available.

cs.LG↗

arXiv2023

Interpretable Ensembles of Hyper-Rectangles as Base Models

A new extremely simple ensemble-based model with the uniformly generated axis-parallel hyper-rectangles as base models (HRBM) is proposed. Two types of HRBMs are studied: closed rectangles and corners. The main idea behind HRBM is to consider and count training examples inside and outside each rectangle. It is proposed to incorporate HRBMs into the gradient boosting machine (GBM). Despite simplicity of HRBMs, it turns out that these simple base models allow us to construct effective ensemble-based models and avoid overfitting. A simple method for calculating optimal regularization parameters of the ensemble-based model, which can be modified in the explicit way at each iteration of GBM, is considered. Moreover, a new regularization called the "step height penalty" is studied in addition to the standard L1 and L2 regularizations. An extremely simple approach to the proposed ensemble-based model prediction interpretation by using the well-known method SHAP is proposed. It is shown that GBM with HRBM can be regarded as a model extending a set of interpretable models for explaining black-box models. Numerical experiments with real datasets illustrate the proposed GBM with HRBMs for regression and classification problems. Experiments also illustrate computational efficiency of the proposed SHAP modifications. The code of proposed algorithms implementing GBM with HRBM is publicly available.

cs.LG↗

arXiv2023

Multiple Instance Learning with Trainable Decision Tree Ensembles

A new random forest based model for solving the Multiple Instance Learning (MIL) problem under small tabular data, called Soft Tree Ensemble MIL (STE-MIL), is proposed. A new type of soft decision trees is considered, which is similar to the well-known soft oblique trees, but with a smaller number of trainable parameters. In order to train the trees, it is proposed to convert them into neural networks of a specific form, which approximate the tree functions. It is also proposed to aggregate the instance and bag embeddings (output vectors) by using the attention mechanism. The whole STE-MIL model, including soft decision trees, neural networks, the attention mechanism and a classifier, is trained in an end-to-end manner. Numerical experiments with tabular datasets illustrate STE-MIL. The corresponding code implementing the model is publicly available.

cs.LG↗

arXiv2022

BENK: The Beran Estimator with Neural Kernels for Estimating the Heterogeneous Treatment Effect

A method for estimating the conditional average treatment effect under condition of censored time-to-event data called BENK (the Beran Estimator with Neural Kernels) is proposed. The main idea behind the method is to apply the Beran estimator for estimating the survival functions of controls and treatments. Instead of typical kernel functions in the Beran estimator, it is proposed to implement kernels in the form of neural networks of a specific form called the neural kernels. The conditional average treatment effect is estimated by using the survival functions as outcomes of the control and treatment neural networks which consists of a set of neural kernels with shared parameters. The neural kernels are more flexible and can accurately model a complex location structure of feature vectors. Various numerical simulation experiments illustrate BENK and compare it with the well-known T-learner, S-learner and X-learner for several types of the control and treatment outcome functions based on the Cox models, the random survival forest and the Nadaraya-Watson regression with Gaussian kernels. The code of proposed algorithms implementing BENK is available in https://github.com/Stasychbr/BENK.

cs.LG↗

arXiv2022

LARF: Two-level Attention-based Random Forests with a Mixture of Contamination Models

New models of the attention-based random forests called LARF (Leaf Attention-based Random Forest) are proposed. The first idea behind the models is to introduce a two-level attention, where one of the levels is the "leaf" attention and the attention mechanism is applied to every leaf of trees. The second level is the tree attention depending on the "leaf" attention. The second idea is to replace the softmax operation in the attention with the weighted sum of the softmax operations with different parameters. It is implemented by applying a mixture of the Huber's contamination models and can be regarded as an analog of the multi-head attention with "heads" defined by selecting a value of the softmax parameter. Attention parameters are simply trained by solving the quadratic optimization problem. To simplify the tuning process of the models, it is proposed to make the tuning contamination parameters to be training and to compute them by solving the quadratic optimization problem. Many numerical experiments with real datasets are performed for studying LARFs. The code of proposed algorithms can be found in https://github.com/andruekonst/leaf-attention-forest.

cs.LG↗

arXiv2022

Improved Anomaly Detection by Using the Attention-Based Isolation Forest

A new modification of Isolation Forest called Attention-Based Isolation Forest (ABIForest) for solving the anomaly detection problem is proposed. It incorporates the attention mechanism in the form of the Nadaraya-Watson regression into the Isolation Forest for improving solution of the anomaly detection problem. The main idea underlying the modification is to assign attention weights to each path of trees with learnable parameters depending on instances and trees themselves. The Huber's contamination model is proposed to be used for defining the attention weights and their parameters. As a result, the attention weights are linearly depend on the learnable attention parameters which are trained by solving the standard linear or quadratic optimization problem. ABIForest can be viewed as the first modification of Isolation Forest, which incorporates the attention mechanism in a simple way without applying gradient-based algorithms. Numerical experiments with synthetic and real datasets illustrate outperforming results of ABIForest. The code of proposed algorithms is available.

cs.LG↗

arXiv2022

Heterogeneous Treatment Effect with Trained Kernels of the Nadaraya-Watson Regression

A new method for estimating the conditional average treatment effect is proposed in the paper. It is called TNW-CATE (the Trainable Nadaraya-Watson regression for CATE) and based on the assumption that the number of controls is rather large whereas the number of treatments is small. TNW-CATE uses the Nadaraya-Watson regression for predicting outcomes of patients from the control and treatment groups. The main idea behind TNW-CATE is to train kernels of the Nadaraya-Watson regression by using a weight sharing neural network of a specific form. The network is trained on controls, and it replaces standard kernels with a set of neural subnetworks with shared parameters such that every subnetwork implements the trainable kernel, but the whole network implements the Nadaraya-Watson estimator. The network memorizes how the feature vectors are located in the feature space. The proposed approach is similar to the transfer learning when domains of source and target data are similar, but tasks are different. Various numerical simulation experiments illustrate TNW-CATE and compare it with the well-known T-learner, S-learner and X-learner for several types of the control and treatment outcome functions. The code of proposed algorithms implementing TNW-CATE is available in https://github.com/Stasychbr/TNW-CATE.

cs.LG↗

arXiv2022

Attention and Self-Attention in Random Forests

New models of random forests jointly using the attention and self-attention mechanisms are proposed for solving the regression problem. The models can be regarded as extensions of the attention-based random forest whose idea stems from applying a combination of the Nadaraya-Watson kernel regression and the Huber's contamination model to random forests. The self-attention aims to capture dependencies of the tree predictions and to remove noise or anomalous predictions in the random forest. The self-attention module is trained jointly with the attention module for computing weights. It is shown that the training process of attention weights is reduced to solving a single quadratic or linear optimization problem. Three modifications of the general approach are proposed and compared. A specific multi-head self-attention for the random forest is also considered. Heads of the self-attention are obtained by changing its tuning parameters including the kernel parameters and the contamination parameter of models. Numerical experiments with various datasets illustrate the proposed models and show that the supplement of the self-attention improves the model performance for many datasets.

cs.LG↗

arXiv2022

Attention-based Random Forest and Contamination Model

A new approach called ABRF (the attention-based random forest) and its modifications for applying the attention mechanism to the random forest (RF) for regression and classification are proposed. The main idea behind the proposed ABRF models is to assign attention weights with trainable parameters to decision trees in a specific way. The weights depend on the distance between an instance, which falls into a corresponding leaf of a tree, and instances, which fall in the same leaf. This idea stems from representation of the Nadaraya-Watson kernel regression in the form of a RF. Three modifications of the general approach are proposed. The first one is based on applying the Huber's contamination model and on computing the attention weights by solving quadratic or linear optimization problems. The second and the third modifications use the gradient-based algorithms for computing trainable parameters. Numerical experiments with various regression and classification datasets illustrate the proposed method.

cs.LG↗

arXiv2021

Multi-Attention Multiple Instance Learning

A new multi-attention based method for solving the MIL problem (MAMIL), which takes into account the neighboring patches or instances of each analyzed patch in a bag, is proposed. In the method, one of the attention modules takes into account adjacent patches or instances, several attention modules are used to get a diverse feature representation of patches, and one attention module is used to unite different feature representations to provide an accurate classification of each patch (instance) and the whole bag. Due to MAMIL, a combined representation of patches and their neighbors in the form of embeddings of a small dimensionality for simple classification is realized. Moreover, different types of patches are efficiently processed, and a diverse feature representation of patches in a bag by using several attention modules is implemented. A simple approach for explaining the classification predictions of patches is proposed. Numerical experiments with various datasets illustrate the proposed method.

cs.LG↗

arXiv2021

Attention-like feature explanation for tabular data

A new method for local and global explanation of the machine learning black-box model predictions by tabular data is proposed. It is implemented as a system called AFEX (Attention-like Feature EXplanation) and consisting of two main parts. The first part is a set of the one-feature neural subnetworks which aim to get a specific representation for every feature in the form of a basis of shape functions. The subnetworks use shortcut connections with trainable parameters to improve the network performance. The second part of AFEX produces shape functions of features as the weighted sum of the basis shape functions where weights are computed by using an attention-like mechanism. AFEX identifies pairwise interactions between features based on pairwise multiplications of shape functions corresponding to different features. A modification of AFEX with incorporating an additional surrogate model which approximates the black-box model is proposed. AFEX is trained end-to-end on a whole dataset only once such that it does not require to train neural networks again in the explanation stage. Numerical experiments with synthetic and real data illustrate AFEX.

cs.LG↗

arXiv2021

An Imprecise SHAP as a Tool for Explaining the Class Probability Distributions under Limited Training Data

One of the most popular methods of the machine learning prediction explanation is the SHapley Additive exPlanations method (SHAP). An imprecise SHAP as a modification of the original SHAP is proposed for cases when the class probability distributions are imprecise and represented by sets of distributions. The first idea behind the imprecise SHAP is a new approach for computing the marginal contribution of a feature, which fulfils the important efficiency property of Shapley values. The second idea is an attempt to consider a general approach to calculating and reducing interval-valued Shapley values, which is similar to the idea of reachable probability intervals in the imprecise probability theory. A simple special implementation of the general approach in the form of linear optimization problems is proposed, which is based on using the Kolmogorov-Smirnov distance and imprecise contamination models. Numerical examples with synthetic and real data illustrate the imprecise SHAP.

cs.LG↗

arXiv2021

SurvNAM: The machine learning survival model explanation

A new modification of the Neural Additive Model (NAM) called SurvNAM and its modifications are proposed to explain predictions of the black-box machine learning survival model. The method is based on applying the original NAM to solving the explanation problem in the framework of survival analysis. The basic idea behind SurvNAM is to train the network by means of a specific expected loss function which takes into account peculiarities of the survival model predictions and is based on approximating the black-box model by the extension of the Cox proportional hazards model which uses the well-known Generalized Additive Model (GAM) in place of the simple linear relationship of covariates. The proposed method SurvNAM allows performing the local and global explanation. A set of examples around the explained example is randomly generated for the local explanation. The global explanation uses the whole training dataset. The proposed modifications of SurvNAM are based on using the Lasso-based regularization for functions from GAM and for a special representation of the GAM functions using their weighted linear and non-linear parts, which is implemented as a shortcut connection. A lot of numerical experiments illustrate the SurvNAM efficiency.

cs.LG↗

arXiv2021

Ensembles of Random SHAPs

Ensemble-based modifications of the well-known SHapley Additive exPlanations (SHAP) method for the local explanation of a black-box model are proposed. The modifications aim to simplify SHAP which is computationally expensive when there is a large number of features. The main idea behind the proposed modifications is to approximate SHAP by an ensemble of SHAPs with a smaller number of features. According to the first modification, called ER-SHAP, several features are randomly selected many times from the feature set, and Shapley values for the features are computed by means of "small" SHAPs. The explanation results are averaged to get the final Shapley values. According to the second modification, called ERW-SHAP, several points are generated around the explained instance for diversity purposes, and results of their explanation are combined with weights depending on distances between points and the explained instance. The third modification, called ER-SHAP-RF, uses the random forest for preliminary explanation of instances and determining a feature probability distribution which is applied to selection of features in the ensemble-based procedure of ER-SHAP. Many numerical experiments illustrating the proposed modifications demonstrate their efficiency and properties for local explanation.

cs.LG↗

arXiv2020

Interpretable Machine Learning with an Ensemble of Gradient Boosting Machines

A method for the local and global interpretation of a black-box model on the basis of the well-known generalized additive models is proposed. It can be viewed as an extension or a modification of the algorithm using the neural additive model. The method is based on using an ensemble of gradient boosting machines (GBMs) such that each GBM is learned on a single feature and produces a shape function of the feature. The ensemble is composed as a weighted sum of separate GBMs resulting a weighted sum of shape functions which form the generalized additive model. GBMs are built in parallel using randomized decision trees of depth 1, which provide a very simple architecture. Weights of GBMs as well as features are computed in each iteration of boosting by using the Lasso method and then updated by means of a specific smoothing procedure. In contrast to the neural additive model, the method provides weights of features in the explicit form, and it is simply trained. A lot of numerical experiments with an algorithm implementing the proposed method on synthetic and real datasets demonstrate its efficiency and properties for local and global interpretation.

cs.LG↗

arXiv2020

A Generalized Stacking for Implementing Ensembles of Gradient Boosting Machines

The gradient boosting machine is one of the powerful tools for solving regression problems. In order to cope with its shortcomings, an approach for constructing ensembles of gradient boosting models is proposed. The main idea behind the approach is to use the stacking algorithm in order to learn a second-level meta-model which can be regarded as a model for implementing various ensembles of gradient boosting models. First, the linear regression of the gradient boosting models is considered as a simplest realization of the meta-model under condition that the linear model is differentiable with respect to its coefficients (weights). Then it is shown that the proposed approach can be simply extended on arbitrary differentiable combination models, for example, on neural networks which are differentiable and can implement arbitrary functions of gradient boosting models. Various numerical examples illustrate the proposed approach.

cs.LG↗

arXiv2020

Counterfactual explanation of machine learning survival models

A method for counterfactual explanation of machine learning survival models is proposed. One of the difficulties of solving the counterfactual explanation problem is that the classes of examples are implicitly defined through outcomes of a machine learning survival model in the form of survival functions. A condition that establishes the difference between survival functions of the original example and the counterfactual is introduced. This condition is based on using a distance between mean times to event. It is shown that the counterfactual explanation problem can be reduced to a standard convex optimization problem with linear constraints when the explained black-box model is the Cox model. For other black-box models, it is proposed to apply the well-known Particle Swarm Optimization algorithm. A lot of numerical experiments with real and synthetic data demonstrate the proposed method.

cs.LG↗

arXiv2020

Gradient boosting machine with partially randomized decision trees

The gradient boosting machine is a powerful ensemble-based machine learning method for solving regression problems. However, one of the difficulties of its using is a possible discontinuity of the regression function, which arises when regions of training data are not densely covered by training points. In order to overcome this difficulty and to reduce the computational complexity of the gradient boosting machine, we propose to apply the partially randomized trees which can be regarded as a special case of the extremely randomized trees applied to the gradient boosting. The gradient boosting machine with the partially randomized trees is illustrated by means of many numerical examples using synthetic and real data.

cs.LG↗