The creation of novel recommendation algorithms for social networks is currently struggling with the volume of available data originating in such environments. Given that social networks can be modeled as graphs, a distributed graph-oriented support to exploit the computing capabilities of clusters arises as a necessity. In this thesis, a platform for graph storage and processing named Graphly is proposed along with GraphRec, an API for easy specification of recommendation algorithms. Graphly and GraphRec hide distributed programming concerns from the user while still allowing fine-tuning of the remote execution. For example, users may customize an algorithm execution using job distribution strategies, without modifying the original code. GraphRec also simplifies the design of graph-based recommender systems by implementing well-known algorithms as “primitives” that can be reused.

Semi-Supervised Learning (SSL) techniques have become very relevant since they require a small set of labeled data. In this scenario, graph-based SSL algorithms provide a powerful framework for modeling manifold structures in high-dimensional spaces and are effective for the propagation of the few initial labels present in training data through the graph. An important step in graph-based SSL methods is the conversion of tabular data into a weighted graph. The graph construction has a key role in the quality of the classification in graph-based methods. Nevertheless, most of the SSL literature focuses on developing label inference algorithms without studying graph construction methods and its effect on the base algorithm performance. This PhD project aims to study this issue and proposes new methods for graph construction from flat data and improves the performance of the graph-based algorithms.

In this work, we deal with a relatively new statistical tool in machine learning: the estimation of the ratio of two probability densities, or density ratio estimation for short. As a side piece of research that gained its own traction, we also tackle the task of parameter selection in learning algorithms based on kernel methods.

The classical theory assumes that each concept is represented by a set of features In this thesis, I investigate a hybrid knowledge representation that are shared by all the instances that are abstracted by approach that combines classic knowledge the concept. In this way, concepts can be viewed as rules representations, such as rules and ontologies, for classifying objects based on features. The prototype theory, with other cognitively plausible representations, on the other hand, states that concepts are represented such as prototypes and exemplars. The resulting through a typical instance, which has the typical features of framework can combine the strengths of the instances of the concept. Finally, the exemplar theory assumes each approach of knowledge representation, avoiding that each concept is represented by a set of exemplars their weaknesses. It can be used for developing of it. These exemplars are real entities that were previously knowledge-based systems that combine logicbased experienced by the agent. In theories based on prototypes or reasoning and similarity-based reasoning in exemplars, the categorization of a given entity is performed problem-solving processes.

Feature Selection plays an important role in machine learning and data mining, and it is often applied as a data pre-processing step. This task can speed up learning algorithms and sometimes improve their performance. In multi-label learning, label dependence is considered another aspect that can contribute to improve learning performance. A replicable and wide systematic review performed by us corroborates this idea. Based on this information, it is believed that considering label dependence during feature selection can lead to better learning performance. The hypothesis of this work is that multi-label feature selection algorithms that consider label dependence will perform better than the ones that disregard it. To this end, we propose multi-label feature selection algorithms that take into account label relations. These algorithms were experimentally compared to the standard approach for feature selection, showing good performance in terms of feature reduction and predictability of the classifiers built using the selected features.

Automatic detection of relevant terms in medical reports is useful for educational purposes and for clinical research. Natural language processing techniques can be applied in order to identify them. The main goal of this research is to develop a method to identify whether medical reports of imaging studies (usually called radiology reports) written in Spanish are important (in the sense that they have non-negated pathological findings) or not. We also try to identify which finding is present and if possible its relationship with anatomical entities.

To overcome this problem, Blei [Blei et al., 2003] proposed Latent Dirichlet Allocation (LDA), a fully Bayesian This article presents a bipartite graph propagation Model with a consistent generative model. LDA has influenced method to be applied to different tasks in the machine a huge amount of work and have become a mainstay learning unsupervised domain, such as topic in modern statistical machine learning. LDA based models extraction and clustering. We introduce the objectives have a rigorous mathematical treatment of decomposed operations and hypothesis that motivate the use of graph that discover the latent groups (topics). From the based method, and we give the intuition of the proposed practitioner's perspective, creating a new model and deriving Bipartite Graph Propagation Algorithm. The it to an effective and implementable inference algorithm are contribution of this study is the development of new hard and tiresome tasks [Rajesh et al., 2014].

Smart grids enhance power grids by integrating electronic equipment, communication systems and computational tools. In a smart grid, consumers can insert energy into the power grid. We propose a new energy management system (called RLbEMS) that autonomously defines a policy for selling or storing energy surplus in smart homes. This policy is achieved through Batch Reinforcement Learning with historical data about energy prices, energy generation, consumer demand and characteristics of storage systems. In practical problems, RLbEMS has learned good energy selling policies quickly and effectively. We obtained maximum gains of 20.78% and 10.64%, when compared to a Naive-greedy policy, for smart homes located in Brazil and in the USA, respectively. Another important result achieved by RLbEMS was the reduction of about 30% of peak demand, a central desideratum for smart grids.

List of committee members on the program committees of the IJCAI 2015 conference.

This is an exciting time to be an artificial intelligence researcher. AI technologies and applications have truly entered our everyday lives, with AI systems in use throughout society. Against this backdrop of AI’s remarkable success, the Twenty-Fourth International Joint Conference on Artificial Intelligence (IJCAI-2015), to be held in Buenos Aires, Argentina between 25 and 31 July 2015, is poised to break several records. This is the first time the flagship international AI conference has been held in South America, and the number of submissions to the technical program has reached an historical high. These proceedings collect some of the most exciting research taking place in AI today and offer a window into the future. The theme of this year’s conference is Artificial Intelligence and Arts. Being held in Argentina, the home of Tango, the conference will feature invited talks, performances, demos and a technical track dedicated to the exploration and celebration of AI’s growing role in the Arts, both in enriching and producing Arts and in injecting art into AI to make it an elegant and more accessible scientific discipline.