Carlos Brito

Efficient and robust query processing in dynamic environments using random walk techniques

Many existing systems for sensor networks rely on state information stored in the nodes for proper operation (e.g., pointers to parent in a spanning tree, routing information, etc). In dynamic environments, such systems must adopt failure recovery mechanisms, which significantly increase the complexity and impact the overall performance. We investigate alternative schemes for query processing based on random walk techniques. The robustness of this approach under dynamics follows from the simplicity of the process, which only requires the connectivity of the neighborhood to keep moving. In addition we show that visiting a constant fraction of sensor network, say 80%, using a random walk is efficient in number of messages and sufficient for answering many interesting queries with high quality. Finally, the natural behavior of a random walk, also provide the important properties of load-balancing and scalability.

A New Identification Condition for Recursive Models With Correlated Errors

This article establishes a new criterion for the identification of recursive linear models in which some errors are correlated. We show that identification is ensured as long as error correlation does not exist between a cause and its direct effect; no restrictions are imposed on errors associated with indirect causes.

Competitive analysis of organization networks or multicast acknowledgement: how much to wait?

We study, from the perspective of competitive analysis, the trade-off between communication cost and delay cost, or simply the send-or-wait dilemma on a hierarchical rooted tree. The problem is an abstraction of the message aggregation problem on communication networks and an organizational problem in network hierarchies.We consider the most natural variant of the problem, the distributed asynchronous regime, and give tight (within a small additive constant) upper and lower bounds on the competitive ratio of the optimization problem. We also consider the centralized version of the problem, in which there is a central entity which remains updated about any incoming messages to the network and which can control the internal delivery of messages in the network. For the centralized setting, we combine a natural rent-to-buy strategy with prediction techniques to achieve the first constant competitive ratio algorithm for any non-trivial class of network topologies.

An Information Theoretic Lower Bound for Broadcasting in Radio Networks

We consider the problem of deterministic broadcasting in undirected radio networks with limited topological information. We show that for every deterministic protocol there exists a radius 2 network which requires at least \(\Omega(n^{\frac{1}{2}})\) rounds for completing broadcast. The previous best lower bound for constant diameter networks is \(\Omega(n^{\frac{1}{4}})\) rounds, due to [23]. For networks of radius D the lower bound can be extended to \(\Omega((nD)^{\frac{1}{2}})\) rounds. This resolves the open problem posed by [23].

An Improvement of the K-SVD Algorithm with Applications on Face Recognition

Image representation is an essential issue regarding the problems related to image processing and understanding. In the last years, the sparse representation modelling for signals has been receiving a lot of attention due to its state-of-the art performance in different computer vision tasks. One of the important factors to its success is the ability to promote representations well adapted to the data which rised with the dictionary learning algorithm. The most well known of theses algorithms is the K-SVD. In this work we proposed the αK-SVD algorithm, which tries to explore the search space of possible dictionaries better than the K-SVD. Our approach is evaluated on two public face recognition databases. The results showed that our approach achieved better results than the K-SVD and LC-KSVD when the sparsity level is low.

Emergence of Autonomous Behaviors of Virtual Characters through Simulated Reproduction

This paper addresses the problem of autonomous behaviors of virtual characters. We postulate that a behavior is regarded as autonomous when the actions performed by the agent result from the interaction between its internal dynamics and the environment, rather than being externally controlled. In this work, we argue that an autonomous behavior is an agent’s solution to a given problem, which is obtained through a process of self-organization of the dynamics of a system that is composed of the agent’s controller, its body and the environment. That process allows the emergence of complex behaviors without any description of actions or objectives. We show a technique capable of adapting an artificial neural network to consistently control virtual Khepera-like robots by means of simulated reproduction, with no measure of the robots’ fitness. All the robots are either male or female, and they are capable of evolving different kinds of behaviors according to their own characteristics, guided solely by the environment’s dynamics.

Improved lower bound for deterministic broadcasting in radio networks

We consider the problem of deterministic broadcasting in radio networks when the nodes have limited knowledge about the topology of the network. We show that for every deterministic broadcasting protocol there exists a network, of radius 2, for which the protocol takes at least @W(n^1^2) rounds for completing the broadcast. Our argument can be extended to prove a lower bound of @W((nD)^1^2) rounds for broadcasting in radio networks of radius D. This resolves one of the open problems posed in Kowalski and Pelc (2004) [24], where the authors proved a lower bound of @W(n^1^4) rounds for broadcasting in constant diameter networks. We prove the new lower bound for a special family of radius 2 networks. Each network of this family consists of O(n) components which are connected to each other via only the source node. At the heart of the proof is a novel simulation argument, which essentially says that any arbitrarily complicated strategy of the source node can be simulated by the nodes of the networks, if the source node just transmits partial topological knowledge about some component instead of arbitrary complicated messages. To the best of our knowledge this type of simulation argument is novel and may be useful in further improving the lower bound or may find use in other applications.

A stochastic framework for K-SVD with applications on face recognition

In recent years, the sparse representation modeling of signals has received a lot of attention due to its state-of-the-art performance in different computer vision tasks. One important factor to its success is the ability to promote representations that are well adapted to the data. This is achieved by the use of dictionary learning algorithms. The most well known of these algorithms is K-SVD. In this paper, we propose a stochastic framework for K-SVD called

Shrinkage k-Means: A Clustering Algorithm Based on the James-Stein Estimator

\alpha