Context-free and context-sensitive dynamics in recurrent neural networks

Continuous-valued recurrent neural networks can learn mechanisms for processing context-free languages. The dynamics of such networks is usually based on damped oscillation around fixed points in state space and requires that the dynamical components are arranged in certain ways. It is shown that qualitatively similar dynamics with similar constraints hold for a(n)b(n)c(n), a context-sensitive language. The additional difficulty with a(n)b(n)c(n), compared with the context-free language a(n)b(n), consists of 'counting up' and 'counting down' letters simultaneously. The network solution is to oscillate in two principal dimensions, one for counting up and one for counting down. This study focuses on the dynamics employed by the sequential cascaded network, in contrast to the simple recurrent network, and the use of backpropagation through time. Found solutions generalize well beyond training data, however, learning is not reliable. The contribution of this study lies in demonstrating how the dynamics in recurrent neural networks that process context-free languages can also be employed in processing some context-sensitive languages (traditionally thought of as requiring additional computation resources). This continuity of mechanism between language classes contributes to our understanding of neural networks in modelling language learning and processing.

Th structure of the P-II-ATP complex

P-II is a signal transduction protein that is part of the cellular machinery used by many bacteria to regulate the activity of glutamine synthetase and the transcription of its gene. The structure of P-II was solved using a hexagonal crystal form (form I). The more physiologically relevant form of P-II is a complex with small molecule effecters. We describe the structure of P-II with ATP obtained by analysis of two different crystal forms (forms II and III) that were obtained by co-crystallization of P-II with ATP. Both structures have a disordered recognition (T) loop and show differences at their C termini. Comparison of these structures with the form I protein reveals changes that occur on binding ATP. Surprisingly, the structure of the P-II/ATP complex differs with that of GlnK, a functional homologue. The two proteins bind the base and sugar of ATP in a similar manner but show differences in the way that they interact with the phosphates. The differences in structure could account for the differences in their activities, and these have been attributed to a difference in sequence at position 82. It has been demonstrated recently that P-II and GlnK form functional heterotrimers in vivo. We construct models of the heterotrimers and examine the junction between the subunits.

Structure revision and assignment of absolute stereochemistry of a marine C-21 bisfuranoterpene

The C-21 bisfuranoterpene (-)-isotetradehydrofurospongin-1 (6), previously isolated from a Western Australian Spongia sp., has been reisolated from a specimen of Spirastrella papilosa collected during scientific trawling operations in the Great Australian Eight. A 2D NMR analysis of 6 has prompted reassignment of the published structure 5, while degradation and chiral HPLC analysis have allowed determination of the absolute stereochemistry.

Crystal structure of a heptameric Sm-like protein complex from archaea: Implications for the structure and evolution of snRNPs

The Sm/Lsm proteins associate with small nuclear RNA to form the core of small nuclear ribonucleoproteins, required for processes as diverse as pre-mRNA splicing, mRNA degradation and telomere formation. The Lsm proteins from archaea are likely to represent the ancestral Sm/Lsm domain. Here, we present the crystal structure of the Lsm alpha protein from the thermophilic archaeon Methanobacterium thermoautrophicum at 2.0 Angstrom resolution. The Lsm alpha protein crystallizes as a heptameric ring comprised of seven identical subunits interacting via beta -strand pairing and hydrophobic interactions. The heptamer can be viewed as a propeller-like structure in which each blade consists of a seven-stranded antiparallel beta -sheet formed from neighbouring subunits. There are seven slots on the inner surface of the heptamer ring, each of which is lined by Asp, Asn and Arg residues that are highly conserved in the Sm/Lsm sequences. These conserved slots are likely to form the RNA-binding site. In archaea, the gene encoding Lsm alpha is located next to the L37e ribosomal protein gene in a putative operon, suggesting a role for the Lsm alpha complex in ribosome function or biogenesis. (C) 2001 Academic Press.

Comments on "Exact solution for horizontal redistribution by general similarity"

No abstract

Clarification of the Crystal Structure of [(HC(Ph2PO)3)2Cu](ClO4)2·2H2O

EXAFS spectra of [(HC(Ph2PO)(3))(2)Cu](ClO4)(2). 2H(2)O have been measured at room temperature. These show that the CuO6 unit is tetragonally elongated, rather than having the compressed tetragonal geometry previously inferred from the X-ray crystal structure determination. [GRAPHICS]

Transfer matrix eigenvalues of the anisotropic multiparametric U model

A multiparametric extension of the anisotropic U model is discussed which maintains integrability. The R-matrix solving the Yang-Baxter equation is obtained through a twisting construction applied to the underlying U-q(sl (2/1)) superalgebraic structure which introduces the additional free parameters that arise in the model. Three forms of Bethe ansatz solution for the transfer matrix eigenvalues are given which we show to be equivalent.

MHCWeb: Converting a WWW database into a knowledge-based collaborative environment

The World Wide Web (WWW) is useful for distributing scientific data. Most existing web data resources organize their information either in structured flat files or relational databases with basic retrieval capabilities. For databases with one or a few simple relations, these approaches are successful, but they can be cumbersome when there is a data model involving multiple relations between complex data. We believe that knowledge-based resources offer a solution in these cases. Knowledge bases have explicit declarations of the concepts in the domain, along with the relations between them. They are usually organized hierarchically, and provide a global data model with a controlled vocabulary, We have created the OWEB architecture for building online scientific data resources using knowledge bases. OWEB provides a shell for structuring data, providing secure and shared access, and creating computational modules for processing and displaying data. In this paper, we describe the translation of the online immunological database MHCPEP into an OWEB system called MHCWeb. This effort involved building a conceptual model for the data, creating a controlled terminology for the legal values for different types of data, and then translating the original data into the new structure. The 0 WEB environment allows for flexible access to the data by both users and computer programs.

Stochastic structure and individual-tree growth models

The majority of past and current individual-tree growth modelling methodologies have failed to characterise and incorporate structured stochastic components. Rather, they have relied on deterministic predictions or have added an unstructured random component to predictions. In particular, spatial stochastic structure has been neglected, despite being present in most applications of individual-tree growth models. Spatial stochastic structure (also called spatial dependence or spatial autocorrelation) eventuates when spatial influences such as competition and micro-site effects are not fully captured in models. Temporal stochastic structure (also called temporal dependence or temporal autocorrelation) eventuates when a sequence of measurements is taken on an individual-tree over time, and variables explaining temporal variation in these measurements are not included in the model. Nested stochastic structure eventuates when measurements are combined across sampling units and differences among the sampling units are not fully captured in the model. This review examines spatial, temporal, and nested stochastic structure and instances where each has been characterised in the forest biometry and statistical literature. Methodologies for incorporating stochastic structure in growth model estimation and prediction are described. Benefits from incorporation of stochastic structure include valid statistical inference, improved estimation efficiency, and more realistic and theoretically sound predictions. It is proposed in this review that individual-tree modelling methodologies need to characterise and include structured stochasticity. Possibilities for future research are discussed. (C) 2001 Elsevier Science B.V. All rights reserved.

Community structure of Quaternary coral reefs compared with Recent life and death assemblages

This paper assesses the reliability with which fossil reefs record the diversity and community structure of adjacent Recent reefs. The diversity and taxonomic composition of Holocene raised fossil reefs was compared with those of modern reef coral life and death assemblages in adjacent moderate and low-energy shallow reef habitats Of Madang Lagoon, Papua New Guinea. Species richness per sample area and Shannon-Wiener diversity (H') were highest in the fossil reefs, intermediate in the life assemblages, and lowest in the death assemblages. The taxonomic composition of the fossil reefs was most similar to the combination of the life and death assemblages from the modern reefs adjacent to the two fossil reefs. Depth zonation was recorded accurately in the fossil reefs. The Madang fossil reefs represent time-averaged composites of the combined life and death assemblages as they existed at the time the reef was uplifted. Because fossil reefs include overlapping cohorts from the life and death assemblages, lagoonal facies of fossil reefs are dominated by the dominant sediment-producing taxa, which are not necessarily the most abundant in the life assemblage. Rare or slow-growing taxa accumulate more slowly than the encasing sediments and are underrepresented in fossil reef lagoons. Time-averaging dilutes the contribution of rare taxa, rather than concentrating their contribution. Consequently, fidelity indices developed for mollusks in sediments yield low values in coral reef death and fossil assemblages. Branching corals dominate lagoonal facies of fossil reefs because they are abundant, they grow and produce sediment rapidly, and most of the sediment they produce is not exported. Fossil reefs distinguished kilometer-scale variations in community structure more clearly than did the modern life assemblages. This difference implies that fossil,reefs may provide a better long-term record of community structure than modern reefs. This difference also suggests that modern kilometer-scale variation in coral reef community structure may have been reduced by anthropogenic degradation, even in the relatively unimpacted reefs of Madang Lagoon. Holocene and Pleistocene fossil reefs provide a time-integrated historical record of community composition and may be used as long-term benchmarks for comparison with modern, degraded, nearshore reefs. Comparisons between fossil reefs and degraded modern reefs display gross changes in community structure more effectively than they demonstrate local extinction of rare taxa.

Cyclic octapeptides containing thiazole. Effect of stereochemistry and degree of flexibility on calcium binding properties

Solution conformation and calcium binding properties have been investigated for the two cyclic octapeptides cyclo(-D-Thr-D-Val(Thz)-Ile-)(2) (4) and cyclo(-Thr-Gly(Thz)-Ile-Ser-Gly(Thz)-Ile-)(5) and the results are compared to those for the cyclic octapeptides previously studied; ascidiacyclamide (1), patellamide D (2), cyclo(-Thr-D-Val(Thz)-Ile-)(2) (3), and cyclo(-Thr-D-Val-alphaAbu-Ile-)2 (6). Both 4 and 5 contain two heterocyclic thiazole ring constraints but the latter has a larger degree of flexibility as a consequence of the glycine residues within the cyclic framework. The solution conformation of 4 and 5 was determined from H-1 NMR spectra and found to be a twisted figure of eight similar to that for 2. Complexation studies using H-1 NMR and CD spectroscopy yielded 1 : 1 calcium-peptide binding constants (logK) for the two peptides (2.3 (4) and 5.7 (5)). For 5 the magnitude of the binding constant was verified by a competition titration using CD. The different calcium-binding affinities of 3 (logK = 4.0) and 4 is attributed to the stereochemistry of the threonine residue. The magnitude of the binding constant for 5 compared to 3 and 4 (all peptides containing two thiazole ring constrains) demonstrates that the increase in flexibility of the cyclic peptide has a dramatic effect on the Ca2+ binding ability. The affinity for Ca2+ thus decreases in the order (6 similar to 5 > 3 > 2 similar to 1 > 4). The number of carbonyl donors available on each peptide has only a limited effect on calcium binding. The most important factor is the flexibility, which allows for a conformation of the peptide capable of binding calcium efficiently.

Analytical solution to the zero-inertia problem for surge flow phenomena in nonprismatic channels

Surge flow phenomena. e.g.. as a consequence of a dam failure or a flash flood, represent free boundary problems. ne extending computational domain together with the discontinuities involved renders their numerical solution a cumbersome procedure. This contribution proposes an analytical solution to the problem, It is based on the slightly modified zero-inertia (ZI) differential equations for nonprismatic channels and uses exclusively physical parameters. Employing the concept of a momentum-representative cross section of the moving water body together with a specific relationship for describing the cross sectional geometry leads, after considerable mathematical calculus. to the analytical solution. The hydrodynamic analytical model is free of numerical troubles, easy to run, computationally efficient. and fully satisfies the law of volume conservation. In a first test series, the hydrodynamic analytical ZI model compares very favorably with a full hydrodynamic numerical model in respect to published results of surge flow simulations in different types of prismatic channels. In order to extend these considerations to natural rivers, the accuracy of the analytical model in describing an irregular cross section is investigated and tested successfully. A sensitivity and error analysis reveals the important impact of the hydraulic radius on the velocity of the surge, and this underlines the importance of an adequate description of the topography, The new approach is finally applied to simulate a surge propagating down the irregularly shaped Isar Valley in the Bavarian Alps after a hypothetical dam failure. The straightforward and fully stable computation of the flood hydrograph along the Isar Valley clearly reflects the impact of the strongly varying topographic characteristics on the How phenomenon. Apart from treating surge flow phenomena as a whole, the analytical solution also offers a rigorous alternative to both (a) the approximate Whitham solution, for generating initial values, and (b) the rough volume balance techniques used to model the wave tip in numerical surge flow computations.

Analytical solution for tidal propagation in a coupled semi-confined/phreatic coastal aquifer

An analytical solution is derived for tidal fluctuations in a coupled coastal aquifer system consisting of a semi-confined aquifer, a thin semi-permeable layer and a phreatic aquifer. Based on the solution, we study the interactions (via leakage) between the confined and unconfined aquifers in response to tides. The results show that, under certain conditions, leakage from the confined aquifer can affect considerably the tidal water table fluctuation in the phreatic aquifer and vice versa. Ignoring these effects could lead to errors in estimating aquifer properties based on tidal signals. (C) 2002 Elsevier Science Ltd. All rights reserved.

Genetic algorithms for design of liquid retaining structure

In this paper, genetic algorithm (GA) is applied to the optimum design of reinforced concrete liquid retaining structures, which comprise three discrete design variables, including slab thickness, reinforcement diameter and reinforcement spacing. GA, being a search technique based on the mechanics of natural genetics, couples a Darwinian survival-of-the-fittest principle with a random yet structured information exchange amongst a population of artificial chromosomes. As a first step, a penalty-based strategy is entailed to transform the constrained design problem into an unconstrained problem, which is appropriate for GA application. A numerical example is then used to demonstrate strength and capability of the GA in this domain problem. It is shown that, only after the exploration of a minute portion of the search space, near-optimal solutions are obtained at an extremely converging speed. The method can be extended to application of even more complex optimization problems in other domains.