Discriminative learning of Bayesian networks via factorized conditional log-likelihood

We propose an efficient and parameter-free scoring criterion, the factorized conditional log-likelihood (ˆfCLL), for learning Bayesian network classifiers. The proposed score is an approximation of the conditional log-likelihood criterion. The approximation is devised in order to guarantee decomposability over the network structure, as well as efficient estimation of the optimal parameters, achieving the same time and space complexity as the traditional log-likelihood scoring criterion. The resulting criterion has an information-theoretic interpretation based on interaction information, which exhibits its discriminative nature. To evaluate the performance of the proposed criterion, we present an empirical comparison with state-of-the-art classifiers. Results on a large suite of benchmark data sets from the UCI repository show that ˆfCLL-trained classifiers achieve at least as good accuracy as the best compared classifiers, using significantly less computational resources.

Algorithms for Exact Structure Discovery in Bayesian Networks

Bayesian networks are compact, flexible, and interpretable representations of a joint distribution. When the network structure is unknown but there are observational data at hand, one can try to learn the network structure. This is called structure discovery. This thesis contributes to two areas of structure discovery in Bayesian networks: space--time tradeoffs and learning ancestor relations. The fastest exact algorithms for structure discovery in Bayesian networks are based on dynamic programming and use excessive amounts of space. Motivated by the space usage, several schemes for trading space against time are presented. These schemes are presented in a general setting for a class of computational problems called permutation problems; structure discovery in Bayesian networks is seen as a challenging variant of the permutation problems. The main contribution in the area of the space--time tradeoffs is the partial order approach, in which the standard dynamic programming algorithm is extended to run over partial orders. In particular, a certain family of partial orders called parallel bucket orders is considered. A partial order scheme that provably yields an optimal space--time tradeoff within parallel bucket orders is presented. Also practical issues concerning parallel bucket orders are discussed. Learning ancestor relations, that is, directed paths between nodes, is motivated by the need for robust summaries of the network structures when there are unobserved nodes at work. Ancestor relations are nonmodular features and hence learning them is more difficult than modular features. A dynamic programming algorithm is presented for computing posterior probabilities of ancestor relations exactly. Empirical tests suggest that ancestor relations can be learned from observational data almost as accurately as arcs even in the presence of unobserved nodes.

Ion probing with Luminescent PbS Quantum dots in PVA

The PbS quantum dots synthesized in PVA are used to investigate their photoluminescence (PL) response to various ions such as Zn, Cd, Hg, Ag, Cu, Fe, Mn, Co, Cr and Ni ions. The enhancement in the photoluminescence intensity is observed with specific ions namely Zn, Cd, Hg and Ag. Among these four ions, the PL response to Hg and Ag even at sub-micro-molar concentrations is quite high, approximately an order of magnitude higher than Zn and Cd. It is interesting to observe that the change in Pb and S molar ratio has profound effect on the selectivity of these ions. The samples prepared under excess of S are quite effective compared to Pb. Indeed, the later one has hardly any effect on the photoluminescence response. These results also indicate that the sensitivity of these QDs could be fine-tuned by controlling the S concentration at the surface. Contrary to the above, Cu, Fe and Co quenches the photoluminescence. Another interesting property of PbS in PVA observed is photo-brightening mechanism due to the curing of the polymer with laser. However, the presence of excess ions at the surface changes its property to photo-darkening/brightening that depends on the direction of carrier transfer mechanism (from QDs to the surface adsorbed metal ions or vice-versa), which is an interesting feature for metal ion detectivity.

Bayesian Decentralized Spectrum Sensing in Cognitive Radio Networks

This paper considers the problem of spectrum sensing, i.e., the detection of whether or not a primary user is transmitting data by a cognitive radio. The Bayesian framework is adopted, with the performance measure being the probability of detection error. A decentralized setup, where N sensors use M observations each to arrive at individual decisions that are combined at a fusion center to form the overall decision is considered. The unknown fading channel between the primary sensor and the cognitive radios makes the individual decision rule computationally complex, hence, a generalized likelihood ratio test (GLRT)-based approach is adopted. Analysis of the probabilities of false alarm and miss detection of the proposed method reveals that the error exponent with respect to M is zero. Also, the fusion of N individual decisions offers a diversity advantage, similar to diversity reception in communication systems, and a tight bound on the error exponent is presented. Through an analysis in the low power regime, the number of observations needed as a function of received power, to achieve a given probability of error is determined. Monte-Carlo simulations confirm the accuracy of the analysis.

Probing Single and Bilayer Graphene Field Effect Transistors By Raman Spectroscopy

This article is a review of our work related to Raman studies of single layer and bilayer graphenes as a function Fermi level shift achieved by electrochemically top gating a field effect transistor. Combining the transport and in situ Raman studies of the field effect devices, a quantitative understanding is obtained of the phonon renormalization due to doping of graphene. Results are discussed in the light of time dependent perturbation theory, with electron phonon coupling parameter as an input from the density functional theory. It is seen that phonons near and Gamma and K points of the Brillouin zone are renormalized very differently by doping. Further, Gamma-phonon renormalization is different in bilayer graphene as compared to single layer, originating from their different electronic band structures near the zone boundary K-point. Thus Raman spectroscopy is not only a powerful probe to characterize the number of layers and their quality in a graphene sample, but also to quantitatively evaluate electron phonon coupling required to understand the performance of graphene devices.

Probing the inter-tube transport in aligned and random multiwall carbon nanotubes

The temperature and magnetic field dependence of conductivity has been used to probe the inter-tube transport in multiwall carbon nanotubes (MWNTs). The scanning electron microscopy images show highly aligned and random distribution of MWNTs. The conductivity in aligned carbon nanotube (ACNT) and random carbon nanotube (RCNT) samples at low temperature follows T-1/2 (at T < 8 K) and T-3/4 (at T > 8 K) dependence in accordance with the weak localization and electron-electron (e-e) interaction model. The values of diffusion coefficient in ACNT and RCNT are 0.25 x 10(-2) and 0.71 x 10(-2) cm(2) s(-1), respectively, indicating that larger number of inter-tube junctions in later enhances the bulk transport. The positive magnetoconductance (MC) data in both samples show that the weak localization contribution is dominant. However, the saturation of MC at higher fields and lower temperatures indicate that e-e interaction is quite significant in RCNT. The T-3/4 and T-1/2 dependence of inelastic scattering length (l(in)) in ACNT and RCNT samples show that the inelastic e-e scattering is more important in aligned tubes. (C) 2011 American Institute of Physics. doi:10.1063/1.3552911]

Bayesian Learning for Joint Sparse OFDM Channel Estimation and Data Detection

The impulse response of a typical wireless multipath channel can be modeled as a tapped delay line filter whose non-zero components are sparse relative to the channel delay spread. In this paper, a novel method of estimating such sparse multipath fading channels for OFDM systems is explored. In particular, Sparse Bayesian Learning (SBL) techniques are applied to jointly estimate the sparse channel and its second order statistics, and a new Bayesian Cramer-Rao bound is derived for the SBL algorithm. Further, in the context of OFDM channel estimation, an enhancement to the SBL algorithm is proposed, which uses an Expectation Maximization (EM) framework to jointly estimate the sparse channel, unknown data symbols and the second order statistics of the channel. The EM-SBL algorithm is able to recover the support as well as the channel taps more efficiently, and/or using fewer pilot symbols, than the SBL algorithm. To further improve the performance of the EM-SBL, a threshold-based pruning of the estimated second order statistics that are input to the algorithm is proposed, and its mean square error and symbol error rate performance is illustrated through Monte-Carlo simulations. Thus, the algorithms proposed in this paper are capable of obtaining efficient sparse channel estimates even in the presence of a small number of pilots.

Probing potential energy surfaces in confined systems: behavior of mean-square displacement in zeolites

Time evolution of mean-squared displacement based on molecular dynamics for a variety of adsorbate-zeolite systems is reported. Transition from ballistic to diffusive behavior is observed for all the systems. The transition times are found to be system dependent and show different types of dependence on temperature. Model calculations on a one-dimensional system are carried out which show that the characteristic length and transition times are dependent on the distance between the barriers, their heights, and temperature. In light of these findings, it is shown that it is possible to obtain valuable information about the average potential energy surface sampled under specific external conditions.

Probing the structural changes in the phase transitions of a Bi2MoO6 catalyst: the nature of the intermediate-temperature phase

The nature of the phase transitions of Bi2MoO6 has been investigated by the combined use of X-ray diffraction and Xray absorption spectroscopy. The distorted MoO6 octahedra in the low-temperature form are shown to undergo further distortion in the intermediate-temperature form before transforming to MoO4 tetrahedra in the high-temperature phase. (C) 1995 Academic Press, Inc.

Probing the role of the fully conserved Cys126 in triosephosphate isomerase by site-specific mutagenesis - distal effects on dimer stability

Cys126 is a completely conserved residue in triosephosphate isomerase that is proximal to the active site but has been ascribed no specific role in catalysis. A previous study of the C126S and C126A mutants of yeast TIM reported substantial catalytic activity for the mutant enzymes, leading to the suggestion that this residue is implicated in folding and stability [Gonzalez-Mondragon E et al. (2004) Biochemistry43, 3255–3263]. We re-examined the role of Cys126 with the Plasmodium falciparum enzyme as a model. Five mutants, C126S, C126A, C126V, C126M, and C126T, were characterized. Crystal structures of the 3-phosphoglycolate-bound C126S mutant and the unliganded forms of the C126S and C126A mutants were determined at a resolution of 1.7–2.1 Å. Kinetic studies revealed an approximately five-fold drop in kcat for the C126S and C126A mutants, whereas an approximately 10-fold drop was observed for the other three mutants. At ambient temperature, the wild-type enzyme and all five mutants showed no concentration dependence of activity. At higher temperatures (> 40 °C), the mutants showed a significant concentration dependence, with a dramatic loss in activity below 15 μm. The mutants also had diminished thermal stability at low concentration, as monitored by far-UV CD. These results suggest that Cys126 contributes to the stability of the dimer interface through a network of interactions involving His95, Glu97, and Arg98, which form direct contacts across the dimer interface.

Probing the role of the fully conserved Cys126 in triosephosphate isomerase by site-specific mutagenesis - distal effects on dimer stability

DatabaseStructural data are available in the Protein Data Bank under the accession numbers 3PVF, 3PY2, and 3PWA. Structured digital abstract Tim binds to Tim by x-ray crystallography (View interaction).

Probing the limits of steric constraints in the oligomerization of copper(I) complexes: structures of two sterically congested oligomers

Oligomeric copper(I) clusters are formed by the insertion reaction of copper(I) aryloxides into heterocumulenes. The effect of varying the steric demands of the heterocumulene and the aryloxy group on the nuclearity of the oligomers formed has been probed. Reactions with copper(I)2-methoxyphenoxide and copper(I)2-methylphenoxide with PhNCS result in the formation of hexameric complexes hexakis[N-phenylimino(aryloxy)methanethiolato copper(I)] 3 and 4 respectively. Single crystal X-ray data confirmed the structure of 3. Similar insertion reactions of CS2 with the copper(I) aryloxides formed by 2,6-di-tert-butyl-4-methylphenol and 2,6-dimethylphenol result in oligomeric copper(I) complexes 7 and 8 having the (aryloxy)thioxanthate ligand. Complex 7 was confirmed to be a tetramer from single crystal X-ray crystallography. Reactions carried out with 2-mercaptopyrimidine, which has ligating properties similar to N-alkylimino(aryloxy)methanethiolate, result in the formation of an insoluble polymeric complex 11. The fluorescence spectra of oligomeric complexes are helpful in determining their nuclearity. Ir has been shown that a decrease in the steric requirements of either the heterocumulene or aryloxy parts of the ligand can compensate for steric constraints acid facilitate oligomerization. (C) 1999 Elsevier Science Ltd. All rights reserved.

Probing the Allosteric Mechanism in Pyrrolysyl-tRNA Synthetase Using Energy-Weighted Network Formalism

Pyrrolysyl-tRNA synthetase (PyIRS) is an atypical enzyme responsible for charging tRNA(Pyl) with pyrrolysine, despite lacking precise tRNA anticodon recognition. This dimeric protein exhibits allosteric regulation of function, like any other tRNA synthetases. In this study we examine the paths of allosteric communication at the atomic level, through energy-weighted networks of Desulfitobacterium hafniense PyIRS (DhPyIRS) and its complexes with tRNA(Pyl) and activated pyrrolysine. We performed molecular dynamics simulations of the structures of these complexes to obtain an ensemble conformation-population perspective. Weighted graph parameters relevant to identifying key players and ties in the context of social networks such as edge/node betweenness, closeness index, and the concept of funneling are explored in identifying key residues and interactions leading to shortest paths of communication in the structure networks of DhPylRS. Further, the changes in the status of important residues and connections and the costs of communication due to ligand induced perturbations are evaluated. The optimal, suboptimal, and preexisting paths are also investigated. Many of these parameters have exhibited an enhanced asymmetry between the two subunits of the dimeric protein, especially in the pretransfer complex, leading us to conclude that encoding of function goes beyond the sequence/structure of proteins. The local and global perturbations mediated by appropriate ligands and their influence on the equilibrium ensemble of conformations also have a significant role to play in the functioning of proteins. Taking a comprehensive view of these observations, we propose that the origin of many functional aspects (allostery rand half-sites reactivity in the case of DhPyIRS) lies in subtle rearrangements of interactions and dynamics at a global level.

Probing the mobility of lithium in LISICON: Li+/H+ exchange studies in Li2ZnGeO4 and Li2+2xZn1-xGeO4

We investigated Li+/H+ exchange in the lithium ion conductors (LISICONS) [ Li2+2xZn1-xGeO4; x = 0.5 ( I) and x = 0.75 (II)] and their parent, gamma-Li2ZnGeO4. Facile exchange of approximately 2x lithium ions per formula unit occurs with both the LISICONS in dilute acetic acid, while the parent material does not exhibit an obvious Li+/H+ exchange under the same conditions. The results can be understood in terms of lithium ion distribution in the crystal structures: the parent Li2ZnGeO4, where all the lithium ions form part of the tetrahedral framework structure, does not exhibit a ready Li+/H+ exchange; LISICONS, where lithium ions are distributed between framework ( tetrahedral) and nonframework sites, undergo a facile Li+/H+ exchange of the nonframework site lithium ions. Accordingly, Li+/H+ exchange in dilute aqueous acetic acid provides a convenient probe to distinguish between the mobile and the immobile lithium ions in lithium ion conductors.