ARE ARTIFICIAL NEURAL NETS AS SMART AS A RAT

The origins of artificial neural networks are related to animal conditioning theory: both are forms of connectionist theory, which in turn derives from the empiricist philosophers' principle of association. The parallel between animal learning and neural nets suggests that interaction between them should benefit both sides.

The Computational Complexity of Dominance and Consistency in CP-Nets

We investigate the computational complexity of testing dominance and consistency in CP-nets. Previously, the complexity of dominance has been determined for restricted classes in which the dependency graph of the CP-net is acyclic. However, there are preferences of interest that define cyclic dependency graphs; these are modeled with general CP-nets. In our main results, we show here that both dominance and consistency for general CP-nets are PSPACE-complete. We then consider the concept of strong dominance, dominance equivalence and dominance incomparability, and several notions of optimality, and identify the complexity of the corresponding decision problems. The reductions used in the proofs are from STRIPS planning, and thus reinforce the earlier established connections between both areas.

Identification of CO adsorbed at Ru and Pt sites on a polycrystalline Pt/Ru electrode and the observation of their oxidation and free interchange under open circuit conditions

The spontaneous oxidation of CO adsorbates on a Pt electrode modified by Ru under open circuit (OC) conditions in perchloric acid solution has been followed, for the first time, using in situ FTIR spectroscopy, and the dynamics of the surface processes taking place have been elucidated. The IR data show that adsorbed CO present on both the Ru and Pt domains and can be oxidized by the oxygen-containing adlayer on the Ru in a chemical process to produce CO under OC conditions. There is a free exchange of CO is between the Ru and Pt sites. Oxidation of CO may take place at the edges of the Ru islands, but CO is transfer, at least on the time scale of these experiments, allows the two different populations to maintain equilibrium. Oxidation is limited in this region by the rate of supply of oxygen to die surface of the catalyst. A mechanism is postulated to explain the observed behavior.

Decision Making by Credal Nets

Credal nets are probabilistic graphical models which extend Bayesian nets to cope with sets of distributions. This feature makes the model particularly suited for the implementation of classifiers and knowledge-based systems. When working with sets of (instead of single) probability distributions, the identification of the optimal option can be based on different criteria, some of them eventually leading to multiple choices. Yet, most of the inference algorithms for credal nets are designed to compute only the bounds of the posterior probabilities. This prevents some of the existing criteria from being used. To overcome this limitation, we present two simple transformations for credal nets which make it possible to compute decisions based on the maximality and E-admissibility criteria without any modification in the inference algorithms. We also prove that these decision problems have the same complexity of standard inference, being NP^PP-hard for general credal nets and NP-hard for polytrees.

Improving parameter learning of Bayesian nets from incomplete data

This paper addresses the estimation of parameters of a Bayesian network from incomplete data. The task is usually tackled by running the Expectation-Maximization (EM) algorithm several times in order to obtain a high log-likelihood estimate. We argue that choosing the maximum log-likelihood estimate (as well as the maximum penalized log-likelihood and the maximum a posteriori estimate) has severe drawbacks, being affected both by overfitting and model uncertainty. Two ideas are discussed to overcome these issues: a maximum entropy approach and a Bayesian model averaging approach. Both ideas can be easily applied on top of EM, while the entropy idea can be also implemented in a more sophisticated way, through a dedicated non-linear solver. A vast set of experiments shows that these ideas produce significantly better estimates and inferences than the traditional and widely used maximum (penalized) log-likelihood and maximum a posteriori estimates. In particular, if EM is adopted as optimization engine, the model averaging approach is the best performing one; its performance is matched by the entropy approach when implemented using the non-linear solver. The results suggest that the applicability of these ideas is immediate (they are easy to implement and to integrate in currently available inference engines) and that they constitute a better way to learn Bayesian network parameters.

Interchange of entire communities: microbial community coalescence

Microbial communities are enigmatically diverse. We propose a novel view of processes likely affecting microbial assemblages, which could be viewed as the Great American Interchange en miniature: the wholesale exchange among microbial communities resulting from moving pieces of the environment containing entire assemblages. Incidental evidence for such ‘community coalescence’ is accumulating, but such processes are rarely studied, likely because of the absence of suitable terminology or a conceptual framework. We provide the nucleus for such a conceptual foundation for the study of community coalescence, examining factors shaping these events, links to bodies of ecological theory, and we suggest modeling approaches for understanding coalescent communities. We argue for the systematic study of community coalescence because of important functional and applied consequences.

Differential lifetime success and performance of native and non-native Atlantic salmon examined under communal natural conditions.

The lifetime success and performance characteristics of communally reared offspring of wild native Burrishoole (native), ranched native (ranched) and non-native (non-native) Atlantic salmon Salmo salar from the adjacent Owenmore River were compared. Non-native year parr showed a substantial downstream migration, which was not shown by native and ranched parr. This appears to have been an active migration rather than competitive displacement and may reflect an adaptation to environmental or physiographic conditions within the Owenmore River catchment where the main nursery habitat is downstream of the spawning area. There were no differences between native and ranched in smolt output or adult return. Both of these measures, however, were significantly lower for the non-native group. A greater proportion of the non-native Atlantic salmon was taken in the coastal drift nets compared to the return to the Burrishoole system, probably as a result of the greater size of the non-native fish. The overall lifetime success of the non-native group, from fertilized egg to returning adult, was some 35% of native and ranched. The ranched group showed a significantly greater male parr maturity, a greater proportion of 1+ year smolts, and differences in sex ratio and timing of freshwater entry of returning adults compared to native, which may have fitness implications under specific conditions.

‘Care just changes your life’: Factors impacting on the mental health of children and young people in care in Northern Ireland.

This paper represents one element of a research project carried out into the mental health needs of children and young people with experiences of care in Northern Ireland. Focusing exclusively on qualitative data collected from 51 young people in care and aftercare, it discusses in the first instance how the challenges and difficulties faced by young people can manifest themselves in feelings and behaviours that may exemplify poor mental well-being. In doing so it provides an understanding of mental health in the context of these young people’s lives. Through offering a more detailed account of some of the specific issues that put these young people at increased risk, it highlights areas for further work and consideration as a means of protecting them against these risks. These include: dealing with experiences prior to care; easing and ‘‘normalising’’ the experience of living in care; and enhancing ‘‘safety nets’’ after care. A key objective of the research is to inform policy and practice through the accounts of children and young people. It is argued that more work needs to be done to find creative ways of enhancing the day-to-day experiences of young people while in care and when leaving care.

Melting of a tetrahedral network model of silica

Thermal properties of an idealised tetrahedral network model of silica are investigated by Monte Carlo simulations. The interatomic potential consists of anharmonic stretching and bending terms, plus a short range repulsion. The model includes a bond interchange rule similar to the well known Wooten, Winer and Weaire (WWW) algorithm (see Phys. Rev. Lett., 1985, 54, 1392). Simulations reveal an apparent first order melting transition at T = 2200 K. The computed changes in the local coordination upon melting are consistent with experimental and ab initio data.

Thermodynamic properties and atomistic structure of the dry amorphous silica surface from a reactive force field model

A force field model of the Keating type supplemented by rules to break, form, and interchange bonds is applied to investigate thermodynamic and structural properties of the amorphous SiO2 surface. A simulated quench from the liquid phase has been carried out for a silica sample made of 3888 silicon and 7776 oxygen atoms arranged on a slab similar to 40 angstrom thick, periodically repeated along two directions. The quench results into an amorphous sample, exposing two parallel square surfaces of similar to 42 nm(2) area each. Thermal averages computed during the quench allow us to determine the surface thermodynamic properties as a function of temperature. The surface tension turns out to be gamma=310 +/- 20 erg/cm(2) at room temperature and gamma=270 +/- 30 at T=2000 K, in fair agreement with available experimental estimates. The entropy contribution Ts-s to the surface tension is relatively low at all temperatures, representing at most similar to 20% of the surface energy. Almost without exceptions, Si atoms are fourfold coordinated and oxygen atoms are twofold coordinated. Twofold and threefold rings appear only at low concentration and are preferentially found in proximity of the surface. Above the glass temperature T-g=1660 +/- 50 K, the mobility of surface atoms is, as expected, slightly higher than that of bulk atoms. The computation of the height-height correlation function shows that the silica surface is rough in the equilibrium and undercooled liquid phase, becoming smooth below the glass temperature T-g.