The lexical nature of syntactic ambiguity resolution.

Ambiguity resolution is a central problem in language comprehension. Lexical and syntactic ambiguities are standardly assumed to involve different types of knowledge representations and be resolved by different mechanisms. An alternative account is provided in which both types of ambiguity derive from aspects of lexical representation and are resolved by the same processing mechanisms. Reinterpreting syntactic ambiguity resolution as a form of lexical ambiguity resolution obviates the need for special parsing principles to account for syntactic interpretation preferences, reconciles a number of apparently conflicting results concerning the roles of lexical and contextual information in sentence processing, explains differences among ambiguities in terms of ease of resolution, and provides a more unified account of language comprehension than was previously available. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Gradient calculations for dynamic recurrent neural networks: asurvey

Surveys learning algorithms for recurrent neural networks with hidden units and puts the various techniques into a common framework. The authors discuss fixed point learning algorithms, namely recurrent backpropagation and deterministic Boltzmann machines, and nonfixed point algorithms, namely backpropagation through time, Elman's history cutoff, and Jordan's output feedback architecture. Forward propagation, an on-line technique that uses adjoint equations, and variations thereof, are also discussed. In many cases, the unified presentation leads to generalizations of various sorts. The author discusses advantages and disadvantages of temporally continuous neural networks in contrast to clocked ones continues with some "tricks of the trade" for training, using, and simulating continuous time and recurrent neural networks. The author presents some simulations, and at the end, addresses issues of computational complexity and learning speed.     

Independent components of magnetoencephalography: localization

We applied second-order blind identification (SOBI), an independent component analysis method, to MEG data collected during cognitive tasks. We explored SOBI's ability to help isolate underlying neuronal sources with relatively poor signal-to-noise ratios, allowing their identification and localization. We compare localization of the SOBI-separated components to localization from unprocessed sensor signals, using an equivalent current dipole modeling method. For visual and somatosensory modalities, SOBI preprocessing resulted in components that can be localized to physiologically and anatomically meaningful locations. Furthermore, this preprocessing allowed the detection of neuronal source activations that were otherwise undetectable. This increased probability of neuronal source detection and localization can be particularly beneficial for MEG studies of higher-level cognitive functions, which often have greater signal variability and degraded signal-to-noise ratios than sensory activation tasks.     

A multi-stage down-draft evaporative cool tower for semi-enclosed spaces: Experiments with a water spraying system

A novel down-draft evaporative cool tower (DECT) was developed which incorporates a secondary air inlet and a complex longitudinal section that comprises two partly overlapping cones. The objective of this design is to conserve water while providing equal or superior cooling to conventional DECTs, in which dry ambient air is drawn in through a single inlet at the top and cooler moist air is delivered at the bottom.The complex section and the addition of a secondary air intake near the middle required a sophisticated water spraying system. In addition to maximizing the cooling potential, the spraying system was designed to limit spray drift beyond the base of the tower, to reduce maintenance costs (especially due to clogging of the sprayers) and to minimize pumping energy. Analysis shows that maximum cooling may be obtained either by employing a very fine spray, requiring the introduction of a relatively small volume of water, or by spraying a larger volume of coarser drops. However, spraying fine drops requires more pumping power, finer nozzles are more likely to clog and small drops of water aggravate the problem of drift near the tower base. If full evaporation of the water spray is not required and excess water is collected for reuse, the second option is thus preferable. In addition to the theoretical analysis, the paper presents experimental findings on temperature reduction, water consumption and cooling output of an 8-m high prototype tower constructed at Sede-Boqer, Israel.     

Green building standards as socio-technical actors in municipal environmental policy

Methods for the environmental assessment and certification of ‘green’ buildings are increasingly being adopted by local governments and other entities as a means of meeting environmental, and often energy-related, goals. Literature on building environmental assessment has examined how these tools are interpreted by stakeholders in design and construction, but less attention has been paid to their interpretation and usage as objects of policy – despite this being a channel through which they have a potentially huge impact. Based on a case study of the emergence of green building in Israel, and drawing on socio-technical literatures, this paper explores the meanings attributed to building environmental assessment in the policy context. It finds that these meanings include a platform for divergent environmental goals and a proxy measure for greenhouse gas abatement. The analysis suggests that for policy-makers, the significance of green building lies not in its constituting a set of environmental benchmarks but as a standard that gains currency as a black-boxed policy object. An open discussion is needed on the various logics driving the use of green building tools in public policy, and what they hope to achieve.     

Performance analysis of a simple roof cooling system with irrigated soil and two shading alternatives

The use of passive cooling on roofs holds a significant unfulfilled potential in hot-arid regions. In this study, the contribution of a watered soil with two types of shading for roof cooling was assessed. Two test cells of approximately 4 m² in area and 2.5 m high were monitored during the summer season. Both cells were covered with a 16-cm layer of soil. One was untreated while the other was watered and shaded consecutively by means of an overhead shading mesh and a layer of lightweight gravel. Temperature profiles were measured across the section of each roof, from the top surface of the soil to the ceiling inside the chamber, and embedded heat flux plates were also used to evaluate the cooling effect. A comparison of the two shading strategies demonstrated that while the mesh provided more cooling over a daily cycle, the daytime cooling potential, which is crucial in a desert climate, was higher with lightweight gravel.     

MEG source localization using an MLP with a distributed output representation

We present a system that takes realistic magnetoencephalographic (MEG) signals and localizes a single dipole to reasonable accuracy in real time. At its heart is a multilayer perceptron (MLP) which takes the sensor measurements as inputs, uses one hidden layer, and generates as outputs the amplitudes of receptive fields holding a distributed representation of the dipole location. We trained this Soft-MLP on dipolar sources with real brain noise and converted the network's output into an explicit Cartesian coordinate representation of the dipole location using two different decoding strategies. The proposed Soft-MLPs are much more accurate than previous networks which output source locations in Cartesian coordinates. Hybrid Soft-MLP-start-LM systems, in which the Soft-MLP output initializes Levenberg-Marquardt, retained their accuracy of 0.28 cm with a decrease in computation time from 36 ms to 30 ms. We apply the Soft-MLP localizer to real MEG data separated by a blind source separation algorithm, and compare the Soft-MLP dipole locations to those of a conventional system.     

Comments on `Dynamic programming approach to optimal weightselection in multilayer neural networks' [with reply]

The commenter claims that in the above-titled paper (ibid., vol.2, p.465-467, July 1991), which presents an efficient algorithm using dynamic programming to find weights which load a set of examples into a feedforward neural network with minimal error, a contradiction lies buried in the paper's notation. In reply, the author maintains that the comments are due to some misunderstandings about the implementation of dynamic-programming-based algorithms and clarifies the work     

Recency in verb phrase attachment.

Four experiments investigated attachment preferences in constructions involving 3 verb phrases (VPs) followed by an attaching modifier. Readers preferred attachment to the most recent (lowest) VP site overall and preferred to attach the modifier to the middle VP over the highest VP, demonstrating a monotonic recency-based preference ordering. This pattern could not be attributed to lexical or plausibility-based preferences. The results contrast with the pattern for relative clause attachment into 3 potential noun phrase sites, where the preference ordering is nonmonotonic, and support the multiple-constraint theory described by E. Gibson and N. J. Pearlmutter (1998), which proposes that recency/locality and a secondary factor, predicate proximity, combine with lexical, grammatical, prosodic, and contextual constraints to determine attachment preferences. (PsycINFO Database Record (c) 2010 APA, all rights reserved)