*Neural Comput ; 31(3): 538-554, 2019 03.*

##### RESUMO

This letter deals with neural networks as dynamical systems governed by finite difference equations. It shows that the introduction of

##### Assuntos

Redes Neurais de Computação , Simulação por Computador*IEEE Trans Neural Netw Learn Syst ; 29(10): 5166-5172, 2018 10.*

##### RESUMO

This paper presents a sequential learning framework for sensors in a network, where a few sensors assume the role of an instructor to train other sensors in the network. The instructors provide estimated labels for measurements of new sensors. These labels are possibly noisy, because a classifier of the instructor may not be perfect. A recursive density estimator is proposed to obtain the true measurement model (i.e., the observation density conditioned on the label) in spite of the training with noisy labels. Specifically, this paper answers the question "Can a sensor train other sensors?", provides necessary conditions for sensors to act as instructors, presents a sequential learning framework using recursive nonparametric kernel density estimation, and provides a convergence rate for the expected error in an observation density. The underlying concepts are illustrated and validated with simulation results.

*Neural Comput ; 30(9): 2500-2529, 2018 09.*

##### RESUMO

Estimation of a generating partition is critical for symbolization of measurements from discrete-time dynamical systems, where a sequence of symbols from a (finite-cardinality) alphabet may uniquely specify the underlying time series. Such symbolization is useful for computing measures (e.g., Kolmogorov-Sinai entropy) to identify or characterize the (possibly unknown) dynamical system. It is also useful for time series classification and anomaly detection. The seminal work of Hirata, Judd, and Kilminster ( 2004 ) derives a novel objective function, akin to a clustering objective, that measures the discrepancy between a set of reconstruction values and the points from the time series. They cast estimation of a generating partition via the minimization of their objective function. Unfortunately, their proposed algorithm is nonconvergent, with no guarantee of finding even locally optimal solutions with respect to their objective. The difficulty is a heuristic nearest neighbor symbol assignment step. Alternatively, we develop a novel, locally optimal algorithm for their objective. We apply iterative nearest-neighbor symbol assignments with guaranteed discrepancy descent, by which joint, locally optimal symbolization of the entire time series is achieved. While most previous approaches frame generating partition estimation as a state-space partitioning problem, we recognize that minimizing the Hirata et al. ( 2004 ) objective function does not induce an explicit partitioning of the state space, but rather the space consisting of the entire time series (effectively, clustering in a (countably) infinite-dimensional space). Our approach also amounts to a novel type of sliding block lossy source coding. Improvement, with respect to several measures, is demonstrated over popular methods for symbolizing chaotic maps. We also apply our approach to time-series anomaly detection, considering both chaotic maps and failure application in a polycrystalline alloy material.

*IEEE Trans Cybern ; 48(6): 1898-1909, 2018 Jun.*

##### RESUMO

This paper presents information-theoretic performance analysis of passive sensor networks for detection of moving targets. The proposed method falls largely under the category of data-level information fusion in sensor networks. To this end, a measure of information contribution for sensors is formulated in a symbolic dynamics framework. The network information state is approximately represented as the largest principal component of the time series collected across the network. To quantify each sensor's contribution for generation of the information content, Markov machine models as well as x-Markov (pronounced as cross-Markov) machine models, conditioned on the network information state, are constructed; the difference between the conditional entropies of these machines is then treated as an approximate measure of information contribution by the respective sensors. The x-Markov models represent the conditional temporal statistics given the network information state. The proposed method has been validated on experimental data collected from a local area network of passive sensors for target detection, where the statistical characteristics of environmental disturbances are similar to those of the target signal in the sense of time scale and texture. A distinctive feature of the proposed algorithm is that the network decisions are independent of the behavior and identity of the individual sensors, which is desirable from computational perspectives. Results are presented to demonstrate the proposed method's efficacy to correctly identify the presence of a target with very low false-alarm rates. The performance of the underlying algorithm is compared with that of a recent data-driven, feature-level information fusion algorithm. It is shown that the proposed algorithm outperforms the other algorithm.

*IEEE Trans Neural Netw Learn Syst ; 29(9): 4128-4139, 2018 09.*

##### RESUMO

This paper proposes a Bayesian nonparametric regression model of panel data for sequential pattern classification. The proposed method provides a flexible and parsimonious model that allows both time-independent spatial variables and time-dependent exogenous variables to be predictors. Not only this method improves the accuracy of parameter estimation for limited data, but also it facilitates model interpretation by identifying statistically significant predictors with hypothesis testing. Moreover, as the data length approaches infinity, posterior consistency of the model is guaranteed for general data-generating processes under regular conditions. The resulting model of panel data can also be used for sequential classification. The proposed method has been tested by numerical simulation, then validated on an econometric public data set, and subsequently validated for detection of combustion instabilities with experimental data that have been generated in a laboratory environment.

*Entropy (Basel) ; 20(6)2018 May 23.*

##### RESUMO

This paper presents a nonparametric regression model of categorical time series in the setting of conditional tensor factorization and Bayes network. The underlying algorithms are developed to provide a flexible and parsimonious representation for fusion of correlated information from heterogeneous sources, which can be used to improve the performance of prediction tasks and infer the causal relationship between key variables. The proposed method is first illustrated by numerical simulation and then validated with two real-world datasets: (1) experimental data, collected from a swirl-stabilized lean-premixed laboratory-scale combustor, for detection of thermoacoustic instabilities and (2) publicly available economics data for causal inference-making.

*IEEE Trans Cybern ; 47(1): 93-104, 2017 Jan.*

##### RESUMO

This paper addresses the problem of target detection in dynamic environments in a semi-supervised data-driven setting with low-cost passive sensors. A key challenge here is to simultaneously achieve high probabilities of correct detection with low probabilities of false alarm under the constraints of limited computation and communication resources. In general, the changes in a dynamic environment may significantly affect the performance of target detection due to limited training scenarios and the assumptions made on signal behavior under a static environment. To this end, an algorithm of binary hypothesis testing is proposed based on clustering of features extracted from multiple sensors that may observe the target. First, the features are extracted individually from time-series signals of different sensors by using a recently reported feature extraction tool, called symbolic dynamic filtering. Then, these features are grouped as clusters in the feature space to evaluate homogeneity of the sensor responses. Finally, a decision for target detection is made based on the distance measurements between pairs of sensor clusters. The proposed procedure has been experimentally validated in a laboratory setting for mobile target detection. In the experiments, multiple homogeneous infrared sensors have been used with different orientations in the presence of changing ambient illumination intensities. The experimental results show that the proposed target detection procedure with feature-level sensor fusion is robust and that it outperforms those with decision-level and data-level sensor fusion.

*J Nanosci Nanotechnol ; 16(4): 3447-56, 2016 Apr.*

##### RESUMO

A systematic, hybrid density functional theory study of interaction between SiGe nanotubes (SiGeNTs) and X (X = H, O, H2 and 02) have been performed using the hybrid functional B3LYP and an all electron 3-21G* basis set implemented in GAUSSIAN 09 suite of software. All possible internal. and external adsorption sites were considered, and it was found that H prefers to move onto top of an atom site while O prefers to incorporate into NT wall by breaking the bonds. Adsorption energies for H is ~2.0 eV and for O it is ~5.0 eV. Controlled adsorption of atomic H and several molecular O give rises to defect density states in the frontier orbital region. H rich adsorptions predict the difference between highest occupied molecular orbital (HOMO) energy and the lowest unoccupied molecular orbital (LUMO) energy increase while O rich adsorptions predict the decrease in HOMO-LUMO energy gap. O and O2 adsorptions predict definite ionic bonding character while H atomic adsorptions predict covalent bonding. H2 is very neutral towards the adsorption into SiGeNTs and clealy shows the physisorption adsorption. Considering the all adsorptions, the adsorptions happened within the Si vicinity of the SiGeNT shows the most stable and preferred adsorption region.

*IEEE Trans Image Process ; 25(1): 24-38, 2016 Jan.*

##### RESUMO

Dictionary learning algorithms have been successfully used for both reconstructive and discriminative tasks, where an input signal is represented with a sparse linear combination of dictionary atoms. While these methods are mostly developed for single-modality scenarios, recent studies have demonstrated the advantages of feature-level fusion based on the joint sparse representation of the multimodal inputs. In this paper, we propose a multimodal task-driven dictionary learning algorithm under the joint sparsity constraint (prior) to enforce collaborations among multiple homogeneous/heterogeneous sources of information. In this task-driven formulation, the multimodal dictionaries are learned simultaneously with their corresponding classifiers. The resulting multimodal dictionaries can generate discriminative latent features (sparse codes) from the data that are optimized for a given task such as binary or multiclass classification. Moreover, we present an extension of the proposed formulation using a mixed joint and independent sparsity prior, which facilitates more flexible fusion of the modalities at feature level. The efficacy of the proposed algorithms for multimodal classification is illustrated on four different applications--multimodal face recognition, multi-view face recognition, multi-view action recognition, and multimodal biometric recognition. It is also shown that, compared with the counterpart reconstructive-based dictionary learning algorithms, the task-driven formulations are more computationally efficient in the sense that they can be equipped with more compact dictionaries and still achieve superior performance.

##### Assuntos

Processamento de Imagem Assistida por Computador/métodos , Aprendizado de Máquina , Reconhecimento Automatizado de Padrão/métodos , Algoritmos , Identificação Biométrica , Face/anatomia & histologia , Humanos*J Phys Condens Matter ; 26(23): 235502, 2014 Jun 11.*

##### RESUMO

All-electron density functional theory was used to investigate Î´-plutonium (Î´-Pu) alloyed with gallium (Ga) impurities at 3.125, 6.25, 9.375 atomic (at)% Ga concentrations. The results indicated that the lowest energy structure is anti-ferromagnetic, independent of the Ga concentration. At higher Ga concentrations (>3.125 at%), the position of the Ga atoms are separated by four nearest neighbor Pu-Pu shells. The results also showed that the lattice constant contracts with increasing Ga concentration, which is in agreement with experimental data. Furthermore with increasing Ga concentration, the face-centered-cubic structure becomes more stably coupled with increasing short-range disorder. The formation energies show that the alloying process is exothermic, with an energy range of -0.028 to -0.099 eV/atom. The analyses of the partial density of states indicated that the Pu-Ga interactions are dominated by Pu 6d and Ga 4p hybridizations, as well as Ga 4s-4p hybridizations. Finally, the computed formation energies for vacancy and hydrogen-vacancy complexes within the 3.125 at% Ga cell were 1.12 eV (endothermic) and -3.88 eV (exothermic), respectively. In addition, the hydrogen atom prefers to interact much more strongly to the Pu atom than the Ga atom in the hydrogen-vacancy complex.

##### Assuntos

Ligas/química , Gálio/química , Hidrogênio/química , Imãs/química , Plutônio/química , Teoria Quântica , Modelos Químicos , Simulação de Dinâmica Molecular , Estrutura Molecular*J Nanosci Nanotechnol ; 14(2): 1710-33, 2014 Feb.*

##### RESUMO

In this paper, we have reviewed some of the recent theoretical studies on the electronic and structural properties of silicon nanotubes from single-walled to double-walled nanostructures, primarily focusing on the studies performed by the present authors. The studies so far have not indicated any metallic behavior in both single-walled and double-walled silicon nanotubes. Atomic and molecular adsorptions of elements including hydrogen, oxygen and alkali metals on single-walled silicon nanotubes are also reviewed and new results presented in detail. A systematic study of molecular adsorption and co-adsorptions of hydrogen and oxygen molecules in zigzag silicon nanotube (SiNT) has been performed using hybrid density functional theory. For adsorption of two hydrogen molecules in SiNT (10, 0), the original diatomic molecular structure was maintained after adsorption. The most preferred final site for hydrogen molecules is the on-top site. For adsorption of two oxygen molecules, the most preferred sites are bridge sites, which are the parallel or zigzag bridge sites. Complete dissociation, partial dissociation and non-dissociation were observed for adsorption of two oxygen molecules. Peroxide structure and Si-O-O structures have also been observed in adsorption of two oxygen molecules with smaller adsorption energies rather than complete dissociation. For the co-adsorption of one hydrogen molecule and one oxygen molecule, the hydrogen molecule is slightly polarized, and a suppression effect on HOMO-LUMO gap was observed.

##### Assuntos

Modelos Químicos , Modelos Moleculares , Nanotubos/química , Nanotubos/ultraestrutura , Adsorção , Simulação por Computador , Condutividade Elétrica , Transporte de Elétrons , Conformação Molecular , Tamanho da Partícula , Propriedades de Superfície*J Phys Condens Matter ; 25(26): 265001, 2013 Jul 03.*

##### RESUMO

Density functional theory calculations have been performed to provide details of the structural and charge-transfer details related to the solid solution of hydrogen in (Î´)-plutonium. We follow the Flanagan model that outlines the process by which hydrogen interacts with a metal to produce hydride phases, via a sequence of surface, interstitial and defect-bound (trapped) states. Due to the complexities of the electronic structure in plutonium solid-state systems, we take the pragmatic approach of adopting the 'special quasirandom structure' to disperse the atomic magnetic moments. We find that this approach produces sound structural and thermodynamic properties in agreement with the available experimental data. In Î´-Pu, hydrogen has an exothermic binding energy to all of the states relevant in the Flanagan model, and, furthermore, is anionic in all these states. The charge transfer is maximized (i.e. most negative for hydrogen) in the hydride phase. The pathway from surface to hydride is sequentially exothermic, in the order surface < interstitial < grain boundary < vacancy < hydride (hydride being the most exothermic state). Thus, we find that there is no intermediate state that involves an endothermic increase in energy, consistent with the general experimental observations that the hydriding reaction in plutonium metal can proceed with zero apparent activation barrier.

##### Assuntos

Elétrons , Hidrogênio/química , Plutônio/química , Catálise , Modelos Moleculares , Teoria Quântica , Termodinâmica*IEEE Trans Syst Man Cybern B Cybern ; 42(3): 647-59, 2012 Jun.*

##### RESUMO

This paper presents a procedure for behavior identification of mobile robots, which requires limited or no domain knowledge of the underlying process. While the features of robot behavior are extracted by symbolic dynamic filtering of the observed time series, the behavior patterns are classified based on language measure theory. The behavior identification procedure has been experimentally validated on a networked robotic test bed by comparison with commonly used tools, namely, principal component analysis for feature extraction and Bayesian risk analysis for pattern classification.

##### Assuntos

Algoritmos , Inteligência Artificial , Teorema de Bayes , Técnicas de Apoio para a Decisão , Modelos Teóricos , Reconhecimento Automatizado de Padrão/métodos , Robótica/métodos , Simulação por Computador , Movimento (Física)*IEEE Trans Syst Man Cybern B Cybern ; 41(3): 783-91, 2011 Jun.*

##### RESUMO

This paper presents a statistical-mechanics-inspired procedure for optimization of the sensor field configuration to detect mobile targets. The key idea is to capture the low-dimensional behavior of the sensor field configurations across the Pareto front in a multiobjective scenario for optimal sensor deployment, where the nondominated points are concentrated within a small region of the large-dimensional decision space. The sensor distribution is constructed using location-dependent energy-like functions and intensive temperature-like parameters in the sense of statistical mechanics. This low-dimensional representation is shown to permit rapid optimization of the sensor field distribution on a high-fidelity simulation test bed of distributed sensor networks.

##### Assuntos

Inteligência Artificial , Técnicas de Apoio para a Decisão , Modelos Teóricos , Movimento (Física) , Reconhecimento Automatizado de Padrão/métodos , Transdutores , Algoritmos , Simulação por Computador*IEEE Trans Syst Man Cybern B Cybern ; 39(6): 1505-15, 2009 Dec.*

##### RESUMO

This paper formulates a self-organization algorithm to address the problem of global behavior supervision in engineered swarms of arbitrarily large population sizes. The swarms considered in this paper are assumed to be homogeneous collections of independent identical finite-state agents, each of which is modeled by an irreducible finite Markov chain. The proposed algorithm computes the necessary perturbations in the local agents' behavior, which guarantees convergence to the desired observed state of the swarm. The ergodicity property of the swarm, which is induced as a result of the irreducibility of the agent models, implies that while the local behavior of the agents converges to the desired behavior only in the time average, the overall swarm behavior converges to the specification and stays there at all times. A simulation example illustrates the underlying concept.

*Nanotechnology ; 19(33): 335706, 2008 Aug 20.*

##### RESUMO

Ab initio calculations within the framework of hybrid density functional theory and the finite cluster approximation have been performed for the electronic and geometric structures of three different types of armchair germanium carbide nanotube, from (3, 3) to (11, 11). Full geometry and spin optimizations with unrestricted symmetry have been performed. Physically pertinent quantities of interest such as the cohesive energies, band gaps, radial buckling, density of states, dipole moments, and Mulliken charge distributions have been investigated in detail for all nanotubes. For type I nanotubes, the largest cohesive energy obtained is 4.092 eV/atom, whereas for type II and type III nanotubes, the values are 3.987 eV/atom and 3.968 eV/atom, respectively. For optimized type I nanotubes, Ge atoms moved toward the tube axis and C atoms moved in the opposite direction after relaxation, opposite to the trends observed in types II and III. The band gaps for type I nanotubes are larger than the bulk 3C-GeC gap, varying between 2.666 and 3.016 eV, while type II and type III nanotubes have significantly lower band gaps, with all nanotubes being semiconducting in nature. Mulliken charge analysis indicates primarily ionic behavior for type I GeC nanotubes and a mixed ionic with covalent behavior for the other two types. None of the tubes appear to be magnetic. Applications in the field of nano-optoelectronic devices, molecular electronics, and band gap engineering are envisioned for GeC nanotubes.

*Nanotechnology ; 18(49): 495706, 2007 Dec 12.*

##### RESUMO

Using ab initio hybrid density functional theory based calculations, we report here the electronic and geometric structure properties of three different types of single-walled zigzag silicon carbide nanotube simulated by finite clusters with dangling bonds saturated by hydrogen atoms. These three types differ in the spatial arrangements of Si and C atoms. Full geometry and spin optimizations have been performed without any symmetry constraints. A detailed comparison of the structures and stabilities of the three types of nanotube is presented. Our calculations show type 1 structures to be more stable than type 2 structures, consistent with another result found in the literature. The cohesive energies/atom of the newly proposed type 3 nanotubes lie in between type 1 and type 2. The dependence of the electronic band gaps on the respective tube diameters, energy density of states and dipole moments as well as Mulliken charge distributions have been investigated. For all types of nanotube, Si atoms moved outward of the tube axis making two concentric cylinders of Si and C atoms after relaxation, contrary to some published results in the literature for type 1 zigzag nanotubes. The band gaps for type 1 and type 2 nanotubes show an oscillatory pattern as the diameter increases. Unlike the other two types, the band gap for type 3 nanotubes decreases monotonically with increasing tube diameter. All the tubes studied here appear to have triplet ground states except for type 1 (3, 0). It is expected that these tubes with significant surface reconstructions, varieties of band gaps, and magnetic properties would have interesting and important applications in the field of band gap engineering and molecular electronics.

*ISA Trans ; 45(4): 477-90, 2006 Oct.*

##### RESUMO

Identification of statistical patterns from observed time series of spatially distributed sensor data is critical for performance monitoring and decision making in human-engineered complex systems, such as electric power generation, petrochemical, and networked transportation. This paper presents an information-theoretic approach to identification of statistical patterns in such systems, where the main objective is to enhance structural integrity and operation reliability. The core concept of pattern identification is built upon the principles of Symbolic Dynamics, Automata Theory, and Information Theory. To this end, a symbolic time series analysis method has been formulated and experimentally validated on a special-purpose test apparatus that is designed for data acquisition and real-time analysis of fatigue damage in polycrystalline alloys.

##### Assuntos

Inteligência Artificial , Interpretação Estatística de Dados , Modelos Biológicos , Modelos Estatísticos , Reconhecimento Automatizado de Padrão/métodos , Simulação por Computador , Fatores de Tempo*J Nanosci Nanotechnol ; 6(1): 43-53, 2006 Jan.*

##### RESUMO

Silicon fullerene like nanostructures with six carbon atoms on the surface of Si60 cages by substitution, as well as inside the cage at various symmetry orientations have been studied within the generalized gradient approximation to density functional theory. Full geometry optimizations have been performed without any symmetry constraints using the Gaussian 03 suite of programs and the LANL2DZ basis set. Thus, for the silicon atom, the Hay-Wadt pseudopotential with the associated basis set are used for the core electrons and the valence electrons, respectively. For the carbon atom, the Dunning/Huzinaga double zeta basis set is employed. Electronic and geometric properties of the nanostructures are presented and discussed in detail. It was found that optimized silicon-carbon fullerene like nanostructures have increased stability compared to bare Si60 cage and the stability depends on the orientation of carbon atoms, as well as on the nature of bonding between silicon and carbon atoms and also on the carbon-carbon bonding.