Abstract
In this work we present XMuSer, a multi-relational framework suitable to explore temporal patterns available in multi-relational databases. XMuSer's main idea consists of exploiting frequent sequence mining, using an efficient and direct method to learn temporal patterns in the form of sequences. Grounded on a coding methodology and on the efficiency of sequential miners, we find the most interesting sequential patterns available and then map these findings into a new table, which encodes the multi-relational timed data using sequential patterns. In the last step of our framework, we use an ILP algorithm to learn a theory on the enlarged relational database that consists on the original multi-relational database and the new sequence relation. We evaluate our framework by addressing three classification problems. Moreover, we map each one of three different types of sequential patterns: frequent sequences, closed sequences or maximal sequences.

Abstract
Relational reinforcement learning is a promising branch of reinforcement learning research that deals with structured environments. In these environments, states and actions are differentiated by the presence of certain types of objects and the relations between them and the objects that are involved in the actions. This makes it ultimately suited for tasks that require the manipulation of multiple, interacting objects, such as tasks that a future house-holding robot can be expected to perform like cleaning up a dinner table or storing away done dishes. However, the application of relational reinforcement learning to robotics has been hindered by assumptions such as discrete and atomic state observations. Typical robotic observation systems work in a streaming setup, where objects are discovered and recognized and their placement within their surroundings is determined in a quasi continuous manner instead of a state based one. The resulting information stream can be compared to a set of multiple inter-related data streams. In this paper, we propose an adaptive windowing strategy for generating a stream of learning examples and enabling relational learning from this kind of data. Our approach is independent from the learning algorithm and is based on a gradient search over the space of parameter values, i.e., window sizes, guided by the estimation of the testing error. The proposed algorithm performs online and is data driven and flexible. To the best of our knowledge, this is the first work addressing this problem. Our ideas are empirically supported by an extensive experimental evaluation in a controlled setup using artificial data. © 2011 IEEE.

Abstract
The tutorial presents the state-of-the-art in mobile and ubiquitous data stream mining and discusses open research problems, issues, and challenges in this area. © 2011 Authors.

Abstract
As in batch learning, one may identify a class of streaming real-world problems which require the modeling of several targets simultaneously. Due to the dependencies among the targets, simultaneous modeling can be more successful and informative than creating independent models for each target. As a result one may obtain a smaller model able to simultaneously explain the relations between the input attributes and the targets. This problem has not been addressed previously in the streaming setting. We propose an algorithm for inducing multi-target model trees with low computational complexity, based on the principles of predictive clustering trees and probability bounds for supporting splitting decisions. Linear models are computed for each target separately, by incremental training of perceptrons in the leaves of the tree. Experiments are performed on synthetic and real-world datasets. The multi-target regression tree algorithm produces equally accurate and smaller models for simultaneous prediction of all the target attributes, as compared to a set of independent regression trees built separately for each target attribute. When the regression surface is smooth, the linear models computed in the leaves significantly improve the accuracy for all of the targets. © 2011 ACM.

Abstract
Computer games are attracting increasing interest in the Artificial Intelligence (AI) research community, mainly because games involve reasoning, planning and learning [1]. One area of particular interest in the last years is the creation of adaptive game AI. Adaptive game AI is the implementation of AI in computer games that holds the ability to adapt to changing circumstances, i.e., to exhibit adaptive behavior during the play. This kind of adaptation can be created using Machine Learning techniques, such as neural networks, reinforcement learning and bioinspired methods. In order to learn online, a system needs to overcome the main difficulties imposed by games: processing time and memory requirements. Learning in a game needs to be fast and the memory available is usually not enough to store a large number of training examples to a traditional Machine Learning technique. In this context, methods for mining data streams seem to be a natural approach. Data streams are, by definition, sequences of training examples that arrive over time [2]. In the data stream scenario, algorithms are usually incremental and capable of adapting the decision model when a change in the distribution of the training examples is detected. One particularly interesting algorithm for mining data streams is the Very Fast Decision Tree (VFDT) [3]. VFDTs are incremental decision trees designed specifically to meet the data stream problem requirements. In this paper, we analyse the use of VFDTs in the task of learning in a Computer RolePlaying Game context. First, we simulate data from manually designed tactics for a Computer RolePlaying Game, based on Spronck's static tactics [4], and test the suitability of VFDT to rapid learn these tactics. Afterwards, we conduct an experiment in order to simulate a data stream of examples where changes of tactics occur over time, and analyse how VFDT and some of its variations respond to these changes in the target concept. © 2010 IEEE.

Abstract
The problem of real-time extraction of meaningful patterns from time-changing data streams is of increasing importance for the machine learning and data mining communities. Regression in time-changing data streams is a relatively unexplored topic, despite the apparent applications. This paper proposes an efficient and incremental stream mining algorithm which is able to learn regression and model trees from possibly unbounded, high-speed and time-changing data streams. The algorithm is evaluated extensively in a variety of settings involving artificial and real data. To the best of our knowledge there is no other general purpose algorithm for incremental learning regression/model trees able to perform explicit change detection and informed adaptation. The algorithm performs online and in real-time, observes each example only once at the speed of arrival, and maintains at any-time a ready-to-use model tree. The tree leaves contain linear models induced online from the examples assigned to them, a process with low complexity. The algorithm has mechanisms for drift detection and model adaptation, which enable it to maintain accurate and updated regression models at any time. The drift detection mechanism exploits the structure of the tree in the process of local change detection. As a response to local drift, the algorithm is able to update the tree structure only locally. This approach improves the any-time performance and greatly reduces the costs of adaptation.