Learning the dynamics of embedded clauses

Recent work by Siegelmann has shown that the computational power of recurrent neural networks matches that of Turing Machines. One important implication is that complex language classes (infinite languages with embedded clauses) can be represented in neural networks. Proofs are based on a fractal encoding of states to simulate the memory and operations of stacks. In the present work, it is shown that similar stack-like dynamics can be learned in recurrent neural networks from simple sequence prediction tasks. Two main types of network solutions are found and described qualitatively as dynamical systems: damped oscillation and entangled spiraling around fixed points. The potential and limitations of each solution type are established in terms of generalization on two different context-free languages. Both solution types constitute novel stack implementations - generally in line with Siegelmann's theoretical work - which supply insights into how embedded structures of languages can be handled in analog hardware.

Solution structure of CnErg1 (Ergtoxin), a HERG specific scorpion toxin

The three-dimensional structure of chemically synthesized CnErg1 (Ergtoxin), which specifically blocks HERG (human ether-a-go-go-related gene) K+ channels, was determined by nuclear magnetic resonance spectroscopy. CnErg1 consists of a triple-stranded beta-sheet and an a-helix, as is typical of K+ channel scorpion toxins. The peptide structure differs from the canonical structures in that the first beta-strand is shorter and is nearer to the second beta-strand rather than to the third beta-strand on the C-terminus. There is also a large hydrophobic patch on the surface of the toxin, surrounding a central lysine residue, Lys13. We postulate that this hydrophobic patch is likely to form part of the binding surface of the toxin. (C) 2003 Published by Elsevier Science B.V. on behalf of the Federation of European Biochemical Societies.

Water table waves in an unconfined aquifer: Experiments and modeling

[1] Comprehensive measurements are presented of the piezometric head in an unconfined aquifer during steady, simple harmonic oscillations driven by a hydrostatic clear water reservoir through a vertical interface. The results are analyzed and used to test existing hydrostatic and nonhydrostatic, small-amplitude theories along with capillary fringe effects. As expected, the amplitude of the water table wave decays exponentially. However, the decay rates and phase lags indicate the influence of both vertical flow and capillary effects. The capillary effects are reconciled with observations of water table oscillations in a sand column with the same sand. The effects of vertical flows and the corresponding nonhydrostatic pressure are reasonably well described by small-amplitude theory for water table waves in finite depth aquifers. That includes the oscillation amplitudes being greater at the bottom than at the top and the phase lead of the bottom compared with the top. The main problems with respect to interpreting the measurements through existing theory relate to the complicated boundary condition at the interface between the driving head reservoir and the aquifer. That is, the small-amplitude, finite depth expansion solution, which matches a hydrostatic boundary condition between the bottom and the mean driving head level, is unrealistic with respect to the pressure variation above this level. Hence it cannot describe the finer details of the multiple mode behavior close to the driving head boundary. The mean water table height initially increases with distance from the forcing boundary but then decreases again, and its asymptotic value is considerably smaller than that previously predicted for finite depth aquifers without capillary effects. Just as the mean water table over-height is smaller than predicted by capillarity-free shallow aquifer models, so is the amplitude of the second harmonic. In fact, there is no indication of extra second harmonics ( in addition to that contained in the driving head) being generated at the interface or in the interior.

A review of drainage and spontaneous rupture in free standing thin films with tangentially immobile interfaces

A review of spontaneous rupture in thin films with tangentially immobile interfaces is presented that emphasizes the theoretical developments of film drainage and corrugation growth through the linearization of lubrication theory in a cylindrical geometry. Spontaneous rupture occurs when corrugations from adjacent interfaces become unstable and grow to a critical thickness. A corrugated interface is composed of a number of waveforms and each waveform becomes unstable at a unique transition thickness. The onset of instability occurs at the maximum transition thickness, and it is shown that only upper and lower bounds of this thickness can be predicted from linear stability analysis. The upper bound is equivalent to the Freakel criterion and is obtained from the zeroth order approximation of the H-3 term in the evolution equation. This criterion is determined solely by the film radius, interfacial tension and Hamaker constant. The lower bound is obtained from the first order approximation of the H-3 term in the evolution equation and is dependent on the film thinning velocity A semi-empirical equation, referred to as the MTR equation, is obtained by combining the drainage theory of Manev et al. [J. Dispersion Sci. Technol., 18 (1997) 769] and the experimental measurements of Radoev et al. [J. Colloid Interface Sci. 95 (1983) 254] and is shown to provide accurate predictions of film thinning velocity near the critical thickness of rupture. The MTR equation permits the prediction of the lower bound of the maximum transition thickness based entirely on film radius, Plateau border radius, interfacial tension, temperature and Hamaker constant. The MTR equation extrapolates to Reynolds equation under conditions when the Plateau border pressure is small, which provides a lower bound for the maximum transition thickness that is equivalent to the criterion of Gumerman and Homsy [Chem. Eng. Commun. 2 (1975) 27]. The relative accuracy of either bound is thought to be dependent on the amplitude of the hydrodynamic corrugations, and a semiempirical correlation is also obtained that permits the amplitude to be calculated as a function of the upper and lower bound of the maximum transition thickness. The relationship between the evolving theoretical developments is demonstrated by three film thickness master curves, which reduce to simple analytical expressions under limiting conditions when the drainage pressure drop is controlled by either the Plateau border capillary pressure or the van der Waals disjoining pressure. The master curves simplify solution of the various theoretical predictions enormously over the entire range of the linear approximation. Finally, it is shown that when the Frenkel criterion is used to assess film stability, recent studies reach conclusions that are contrary to the relevance of spontaneous rupture as a cell-opening mechanism in foams. (C) 2003 Elsevier Science B.V. All rights reserved.

Cytotoxic iron chelators: Characterization of the structure, solution chemistry and redox activity of ligands and iron complexes of the di-2-pyridyl ketone isonicotinoyl hydrazone (HPKIH) analogues

Di-2-pyridyl ketone isonicotinoyl hydrazone (HPKIH) and a range of its analogues comprise a series of monobasic acids that are capable of binding iron (Fe) as tridentate (N,N,O) ligands. Recently, we have shown that these chelators are highly cytotoxic, but show selective activity against cancer cells. Particularly interesting was the fact that cytotoxicity of the HPKIH analogues is maintained even after complexation with Fe. To understand the potent anti-tumor activity of these compounds, we have fully characterized their chemical properties. This included examination of the solution chemistry and X-ray crystal structures of both the ligands and Fe complexes from this class and the ability of these complexes to mediate redox reactions. Potentiometric titrations demonstrated that all chelators are present predominantly in their charge-neutral form at physiological pH (7.4), allowing access across biological membranes. Keto-enol tautomerism of the ligands was identified, with the tautomers exhibiting distinctly different protonation constants. Interestingly, the chelators form low-spin (diamagnetic) divalent Fe complexes in solution. The chelators form distorted octahedral complexes with Fe-II, with two tridentate ligands arranged in a meridional fashion. Electrochemistry of the Fe complexes in both aqueous and non-aqueous solutions revealed that the complexes are oxidized to their ferric form at relatively high potentials, but this oxidation is coupled to a rapid reaction with water to form a hydrated (carbinolamine) derivative, leading to irreversible electrochemistry. The Fe complexes of the HPKIH analogues caused marked DNA degradation in the presence of hydrogen peroxide. This observation confirms that Fe complexes from the HPKIH series mediate Fenton chemistry and do not repel DNA. Collectively, studies on the solution chemistry and structure of these HPKIH analogues indicate that they can bind cellular Fe and enhance its redox activity, resulting in oxidative damage to vital biomolecules.

Bridgehead arylation: A direct route to advanced intermediates for the synthesis of C-20 diterpene alkaloids

[GRAPHICS] Rapid access to the ABCE ring system of the C-20 diterpene alkaloids was achieved by silver(I)-promoted intramolecular Friedel-Crafts arylation of a functional group-specific 5-bromo-3-azabicyclo[3.3.1]nonane derivative.

Effects of mercury on the arterial blood pressure of anesthetized rats

The available data suggests that hypotension caused by Hg2+ administration may be produced by a reduction of cardiac contractility or by cholinergic mechanisms. The hemodynamic effects of an intravenous injection of HgCl2 (5 mg/kg) were studied in anesthetized rats (N = 12) by monitoring left and right ventricular (LV and RV) systolic and diastolic pressures for 120 min. After HgCl2 administration the LV systolic pressure decreased only after 40 min (99 ± 3.3 to 85 ± 8.8 mmHg at 80 min). However, RV systolic pressure increased, initially slowly but faster after 30 min (25 ± 1.8 to 42 ± 1.6 mmHg at 80 min). Both right and left diastolic pressures increased after HgCl2 treatment, suggesting the development of diastolic ventricular dysfunction. Since HgCl2 could be increasing pulmonary vascular resistance, isolated lungs (N = 10) were perfused for 80 min with Krebs solution (continuous flow of 10 ml/min) containing or not 5 µM HgCl2. A continuous increase in pulmonary vascular resistance was observed, suggesting the direct effect of Hg2+ on the pulmonary vessels (12 ± 0.4 to 29 ± 3.2 mmHg at 30 min). To examine the interactions of Hg2+ and changes in cholinergic activity we analyzed the effects of acetylcholine (Ach) on mean arterial blood pressure (ABP) in anesthetized rats (N = 9) before and after Hg2+ treatment (5 mg/kg). Using the same amount and route used to study the hemodynamic effects we also examined the effects of Hg2+ administration on heart and plasma cholinesterase activity (N = 10). The in vivo hypotensive response to Ach (0.035 to 10.5 µg) was reduced after Hg2+ treatment. Cholinesterase activity (µM h-1 mg protein-1) increased in heart and plasma (32 and 65%, respectively) after Hg2+ treatment. In conclusion, the reduction in ABP produced by Hg2+ is not dependent on a putative increase in cholinergic activity. HgCl2 mainly affects cardiac function. The increased pulmonary vascular resistance and cardiac failure due to diastolic dysfunction of both ventricles are factors that might contribute to the reduction of cardiac output and the fall in arterial pressure.

Proposta de um algoritmo GPC adaptativo com baixo custo computacional

Esta dissertação propõe um algoritmo do Controlador Preditivo Generalizado (GPC) com horizonte de controle igual a um para ser aplicado em plantas industriais com modelos variantes no tempo, simples o su ficiente para ser implementado em Controlador Lógico Programável (PLC). A solução explícita do controlador é obtida em função dos parâmetros do modelo e dos parâmetros de sintonia do GPC (horizonte nal de predição hp e o fator de supressão do sinal de controle ), além das entradas e saídas presentes e passadas. A sintonia do fator de supressão e do horizonte de previsão GPC é feita através do lugar das raízes da equação característica do sistema em malha fechada, sempre que os parâmetros do modelo da planta industrial (estável ou instável em malha aberta) forem modificados.

Radial Distortion Self-Calibration

In cameras with radial distortion, straight lines in space are in general mapped to curves in the image. Although epipolar geometry also gets distorted, there is a set of special epipolar lines that remain straight, namely those that go through the distortion center. By finding these straight epipolar lines in camera pairs we can obtain constraints on the distortion center(s) without any calibration object or plumbline assumptions in the scene. Although this holds for all radial distortion models we conceptually prove this idea using the division distortion model and the radial fundamental matrix which allow for a very simple closed form solution of the distortion center from two views (same distortion) or three views (different distortions). The non-iterative nature of our approach makes it immune to local minima and allows finding the distortion center also for cropped images or those where no good prior exists. Besides this, we give comprehensive relations between different undistortion models and discuss advantages and drawbacks.

Automated traffic route identification through the shared nearest neighbour algorithm

Many organisations need to extract useful information from huge amounts of movement data. One example is found in maritime transportation, where the automated identification of a diverse range of traffic routes is a key management issue for improving the maintenance of ports and ocean routes, and accelerating ship traffic. This paper addresses, in a first stage, the research challenge of developing an approach for the automated identification of traffic routes based on clustering motion vectors rather than reconstructed trajectories. The immediate benefit of the proposed approach is to avoid the reconstruction of trajectories in terms of their geometric shape of the path, their position in space, their life span, and changes of speed, direction and other attributes over time. For clustering the moving objects, an adapted version of the Shared Nearest Neighbour algorithm is used. The motion vectors, with a position and a direction, are analysed in order to identify clusters of vectors that are moving towards the same direction. These clusters represent traffic routes and the preliminary results have shown to be promising for the automated identification of traffic routes with different shapes and densities, as well as for handling noise data.

A creative solution for rapid and inexpensive model-making in product design processes

Innovative Developments in Virtual and Physical Prototyping

Growth and nutrient uptake of coffee seedlings cultivated in nutrient solution with and without silicon addition

In recent years, the application of silicon (Si) in crops, including coffee, has become a common practice. The objective of this study was to assess the silicon uptake by coffee seedlings and its effects on plant growth, water and macro and micronutrient uptake. The research was conducted using nutrient solution in a greenhouse at the Departamento de Fitotecnia da Universidade Federal de Viçosa, in a completely randomized design with two treatments (with and without silicon) and three replications. Each plot consisted of three plants grown in a 800 mL vessel containing the treatment solutions. At every three days, water consumption, the concentration of OH - and the depletion of Si and K were assessed in the nutrient solutions. After 33 days, the plants were assessed with regard to their fresh and dry weight of leaves, roots and stem, shoot height and total length of the plant (shoot and root). Number of leaves and internodes, and the content and accumulation of silicon, macro, and micronutrients were also determined. The consumption of water, the amount of potassium uptake and, biomass accumulation were greater in plants grown in solution without silicon addition. However, the concentration of OH- in the solution and the amount of silicon uptake were greater in plants grown in solution with added silicon. Silicon accumulation was greater in leaves than in stem and roots. Silicon decreased coffee plant accumulation of phosphorus, potassium, calcium, zinc, copper and iron.