Fabrício Enembreck

A Survey on Ensemble Learning for Data Stream Classification

Ensemble-based methods are among the most widely used techniques for data stream classification. Their popularity is attributable to their good performance in comparison to strong single learners while being relatively easy to deploy in real-world applications. Ensemble algorithms are especially useful for data stream learning as they can be integrated with drift detection algorithms and incorporate dynamic updates, such as selective removal or addition of classifiers. This work proposes a taxonomy for data stream ensemble learning as derived from reviewing over 60 algorithms. Important aspects such as combination, diversity, and dynamic updates, are thoroughly discussed. Additional contributions include a listing of popular open-source tools and a discussion about current data stream research challenges and how they relate to ensemble learning (big data streams, concept evolution, feature drifts, temporal dependencies, and others).

Interaction Models for Multiagent Reinforcement Learning

This article proposes and compares different interaction models for reinforcement learning based on multi-agent system. The cooperation during the learning process is crucial to guarantee the convergence to a good policy. The exchange of rewards among the agents during the interaction is a complex task and if it is inadequate it may cause delays in learning or generate unexpected transitions, making the cooperation inefficient and con-verging to a non-satisfactory policy. In order to allow the interactive discovery of high quality policies we have developed several cooperation models based on the ex-change of action policies between the agents. Experimental results have shown that the proposed cooperation models are able to speed up the convergence of the agents while achieving optimal action policies even in high-dimensional environments (e.g. traffic), outperforming the standard Q-learning algorithm.

Distributed Constraint Optimization Problems: Review and perspectives

Intelligent agents is a research area of the Artificial Intelligence intensely studied since the 1980s. Multi-agent systems represent a powerful paradigm of analyzing, projecting, and developing complex systems. One of the main difficulties in modeling a multi-agent system is defining the coordination model, due to the autonomous behavior of the agents. Distributed Constraint Optimization Problems (DCOP) have emerged as one of most important formalisms for coordination and distributed problem solving in multi-agent systems and are capable of modeling a large class of real world problems naturally. This work aims to provide an overview and critical review of DCOP, addressing the most popular methods and techniques, the evolution and comparison of algorithms, and future perspectives on this promising research area.

Adaptive random forests for evolving data stream classification

Random forests is currently one of the most used machine learning algorithms in the non-streaming (batch) setting. This preference is attributable to its high learning performance and low demands with respect to input preparation and hyper-parameter tuning. However, in the challenging context of evolving data streams, there is no random forests algorithm that can be considered state-of-the-art in comparison to bagging and boosting based algorithms. In this work, we present the adaptive random forest (ARF) algorithm for classification of evolving data streams. In contrast to previous attempts of replicating random forests for data stream learning, ARF includes an effective resampling method and adaptive operators that can cope with different types of concept drifts without complex optimizations for different data sets. We present experiments with a parallel implementation of ARF which has no degradation in terms of classification performance in comparison to a serial implementation, since trees and adaptive operators are independent from one another. Finally, we compare ARF with state-of-the-art algorithms in a traditional test-then-train evaluation and a novel delayed labelling evaluation, and show that ARF is accurate and uses a feasible amount of resources.

Trust and Reputation Models for Multiagent Systems

Finding reliable partners to interact with in open environments is a challenging task for software agents, and trust and reputation mechanisms are used to handle this issue. From this viewpoint, we can observe the growing body of research on this subject, which indicates that these mechanisms can be considered key elements to design multiagent systems (MASs). Based on that, this article presents an extensive but not exhaustive review about the most significant trust and reputation models published over the past two decades, and hundreds of models were analyzed using two perspectives. The first one is a combination of trust dimensions and principles proposed by some relevant authors in the field, and the models are discussed using an MAS perspective. The second one is the discussion of these dimensions taking into account some types of interaction found in MASs, such as coalition, argumentation, negotiation, and recommendation. By these analyses, we aim to find significant relations between trust dimensions and types of interaction so it would be possible to construct MASs using the most relevant dimensions according to the types of interaction, which may help developers in the design of MASs.

A learning agent to help drive vehicles

This paper presents the development of an intelligent agent used to assist vehicle drivers. The agent has a set of resources to generate its action policy: road and vehicle features and a knowledge base containing conduct rules. The perception of the agent is ensured by a set of sensors, which provide the agent with data such as speed, position and conditions of the brakes. The main agent behaviour is to carry out action plans involving: increase, maintain or reduce speed. The main effort of this research was the induction of conduct rules from data of previous trips. These rules form a classifier used for the selection of actions forming the conduction plan. Results observed with the experiments have showed that the proposed classifier increases the efficiency throughout the conduction of vehicles.

Agents for Collaborative Filtering

This paper describes a new generic agent-based framework for collaborative filtering. Usually, collaborative filtering tools use large collaborative document databases to model users’ preferences. Nevertheless, we believe that collaborative filtering can be accomplished with decentralized systems in which user’s preferences are learned from small individual databases. Such a distributed approach gives more flexibility in adding and removing users, and distributes computational effort naturally, economizing computing resources. In addition, it facilitates a continuous evolution of the users’ preferences. At the same time, agents are commonly used to decentralize generic systems, which leads us to adopt a Multi-Agent System for representing the collaborative environment. In particular, we propose Personal Assistants to communicate with the user and to acquire users’ models. Then, a Decision Agent uses such models to decide who must receive the incoming documents.

A survey on feature drift adaptation

This paper provides insights into a nearly neglected type of drift: feature drifts.Existing works on feature drift detection and adaptation are surveyed.Existing works are empirically analyzed showing how challenging this problem is.This paper provides insights into future directions for research into feature drift detection and adaptation. Data stream mining is a fast growing research topic due to the ubiquity of data in several real-world problems. Given their ephemeral nature, data stream sources are expected to undergo changes in data distribution, a phenomenon called concept drift. This paper focuses on one specific type of drift that has not yet been thoroughly studied, namely feature drift. Feature drift occurs whenever a subset of features becomes, or ceases to be, relevant to the learning task; thus, learners must detect and adapt to these changes accordingly. We survey existing work on feature drift adaptation with both explicit and implicit approaches. Additionally, we benchmark several algorithms and a naive feature drift detection approach using synthetic and real-world datasets. The results from our experiments indicate the need for future research in this area as even naive approaches produced gains in accuracy while reducing resources usage. Finally, we state current research topics, challenges and future directions for feature drift adaptation.

WEB Image Classification Based on the Fusion of Image and Text Classifiers

This paper presents a novel method for the classification of images that combines information extracted from the images and contextual information. The main hypothesis is that contextual information related to an image can contribute in the image classification process. First, independent classifiers are designed to deal with images and text. From the images color, shape and texture features are extracted. These features are used with a neural network (NN) classifier to carry out image classification. On the other hand, contextual information is processed and used with a Naive Bayes (NB) classifier. At the end, the outputs of both classifiers are combined through heuristic rules. Experimental results on a database of more than 5,000 HTML documents have shown that the combination of classifiers provides a meaningful improvement (about 16%) in the correct image classification rate relative to the results provided by the NN classifier alone.

Railroad Driving Model Based on Distributed Constraint Optimization

Distributed Constraint Optimization Problems (DCOP) has emerged as one of the most important multiagent coordination techniques, due to its ability to optimize the global function modeled as a set of constraints. This paper presents a study that proposes to use of distributed constraint optimization techniques to achieve energetic efficiency on railroad driving. The objectives are to present a real world problem with dynamic behavior as a DCOP, compare the main algorithms for this formalism and evaluate the quality of the solutions obtained.