Empirical evidence for ultrametric structure in multi-layer perceptron error surfaces

Combinatorial optimization problems share an interesting property with spin glass systems in that their state spaces can exhibit ultrametric structure. We use sampling methods to analyse the error surfaces of feedforward multi-layer perceptron neural networks learning encoder problems. The third order statistics of these points of attraction are examined and found to be arranged in a highly ultrametric way. This is a unique result for a finite, continuous parameter space. The implications of this result are discussed.

Actively shielded multi-layer gradient coil designs with improved cooling properties

In standard cylindrical gradient coils consisting of a single layer of wires, a limiting factor in achieving very large magnetic field gradients is the rapid increase in coil resistance with efficiency. This is a particular problem in small-bore scanners, such as those used for MR microscopy. By adopting a multi-layer design in which the coil wires are allowed to spread out into multiple layers wound at increasing radii, a more favourable scaling of resistance with efficiency is achieved, thus allowing the design of more powerful gradient coils with acceptable resistance values. Previously this approach has been applied to the design of unshielded, longitudinal, and transverse gradient coils. Here, the multi-layer approach has been extended to allow the design of actively shielded multi-layer gradient coils, and also to produce coils exhibiting enhanced cooling characteristics. An iterative approach to modelling the steady-state temperature distribution within the coil has also been developed. Results indicate that a good level of screening can be achieved in multi-layer coils, that small versions of such coils can yield higher efficiencies at fixed resistance than conventional two-layer (primary and screen) coils, and that performance improves as the number of layers of increases. Simulations show that by optimising multi-layer coils for cooling it is possible to achieve significantly higher gradient strengths at a fixed maximum operating temperature. A four-layer coil of 8 mm inner diameter has been constructed and used to test the steady-state temperature model. (C) 2003 Elsevier Inc. All rights reserved.

Teleportation via Multi-Qubit Channels

We investigate the problem of teleporting an unknown qubit state to a recipient via a channel of 2L qubits. In this procedure a protocol is employed whereby L Bell state measurements are made and information based on these measurements is sent via a classical channel to the recipient. Upon receiving this information the recipient determines a local gate which is used to recover the original state. We find that the 2(2L)-dimensional Hilbert space of states available for the channel admits a decomposition into four subspaces. Every state within a given subspace is a perfect channel, and each sequence of Bell measurements projects 2L qubits of the system into one of the four subspaces. As a result, only two bits of classical information need be sent to the recipient for them to determine the gate. We note some connections between these four subspaces and ground states of many-body Hamiltonian systems, and discuss the implications of these results towards understanding entanglement in multi-qubit systems.

Multi-threaded Communicating Agents in Qu-Prolog

In this tutorial paper we summarise the key features of the multi-threaded Qu-Prolog language for implementing multi-threaded communicating agent applications. Internal threads of an agent communicate using the shared dynamic database used as a generalisation of Linda tuple store. Threads in different agents, perhaps on different hosts, communicate using either a thread-to-thread store and forward communication system, or by a publish and subscribe mechanism in which messages are routed to their destinations based on content test subscriptions. We illustrate the features using an auction house application. This is fully distributed with multiple auctioneers and bidders which participate in simultaneous auctions. The application makes essential use of the three forms of inter-thread communication of Qu-Prolog. The agent bidding behaviour is specified graphically as a finite state automaton and its implementation is essentially the execution of its state transition function. The paper assumes familiarity with Prolog and the basic concepts of multi-agent systems.

Assessing flow in physical activity: The Flow State Scale-2 and Dispositional Flow Scale-2

The Flow State Scale-2 (FSS-2) and Dispositional Flow Scale-2 (DFS-2) are presented as two self-report instruments designed to assess flow experiences in physical activity. Item modifications were made to the original versions of these scales in order to improve the measurement of some of the flow dimensions. Confirmatory factor analyses of an item identification and a cross-validation sample demonstrated a good fit of the new scales. There was support for both a 9-first-order factor model and a higher order model with a global flow factor. The item identification sample yielded mean item loadings on the first-order factor of .78 for the FSS-2 and .77 for the DFS-2. Reliability estimates ranged from .80 to .90 for the FSS-2, and .81 to .90 for the DFS-2. In the cross-validation sample, mean item loadings on the first-order factor were .80 for the FSS-2, and .73 for the DFS-2. Reliability estimates ranged between .80 to .92 for the FSS-2 and .78 to .86 for the DFS-2. The scales are presented as ways of assessing flow experienced within a particular event (FSS-2) or the frequency of flow experiences in chosen physical activity in general (DFS-2).

Religious Information Poverty in Australian State Schools

This paper introduces the concept of religious information poverty in Australian state schools from an information science perspective. Information scientists have been theorising about the global information society for some time, along with its increased provision of vital information for the good of the world. Australian state schools see themselves as preparing children for effective participation in the information society, yet Australian children are currently suffering a religious illiteracy that undermines this goal. Some reasons and theories are offered to explain the existence of religious information poverty in state schools, and suggestions for professional stakeholders are offered for its alleviation.

A master equation model for bimolecular reaction via multi-well isomerization intermediates

A reversible linear master equation model is presented for pressure- and temperature-dependent bimolecular reactions proceeding via multiple long-lived intermediates. This kinetic treatment, which applies when the reactions are measured under pseudo-first-order conditions, facilitates accurate and efficient simulation of the time dependence of the populations of reactants, intermediate species and products. Detailed exploratory calculations have been carried out to demonstrate the capabilities of the approach, with applications to the bimolecular association reaction C3H6 + H reversible arrow C3H7 and the bimolecular chemical activation reaction C2H2 +(CH2)-C-1--> C3H3+H. The efficiency of the method can be dramatically enhanced through use of a diffusion approximation to the master equation, and a methodology for exploiting the sparse structure of the resulting rate matrix is established.