Bounds on the number of samples needed for neural learning.

The relationship between the number of hidden nodes in a neural network, the complexity of a multiclass discrimination problem, and the number of samples needed for effect learning are discussed. Bounds for the number of samples needed for effect learning are given. It is shown that Omega(min (d,n) M) boundary samples are required for successful classification of M clusters of samples using a two-hidden-layer neural network with d-dimensional inputs and n nodes in the first hidden layer.     

A VLSI architecture for a half-edge-based corner detector

Corner detection is a low-level feature detection operator that is of great use in image processing applications, for example, optical flow and structure from motion by image correspondence. The detection of corners is a computationally intensive operation. Past implementations of corner detection techniques have been restricted to software. In this paper we propose an efficient very large-scale integration (VLSI) architecture for detection of corners in images. The corner detection technique is based on the half-edge concept and the first directional derivative of Gaussian. Apart from the location of the corner points, the algorithm also computes the corner orientation and the corner angle and outputs the edge map of the image. The symmetrical properties of the masks are utilized to reduce the number of convolutions effectively, from eight to two. Therefore, the number of multiplications required per pixel is reduced from 1800 to 392. Thus, the proposed architecture yields a speed-up factor of 4.6 over conventional convolution architectures. The architecture uses the principles of pipelining and parallelism and can be implemented in VLSI.     

Performance of hashed cache data migration schemes on multicomputers

Many researchers approach the problem of programming distributed memory machines by assuming a global shared name space. Thus the user views the distributed memory of the machine as though it were shared. A major issue that arises at this point is how to manage the memory. When a processor accesses data stored on another processor's memory, data must be moved between the two processors. Once these data are retrieved from another processor's memory, several interesting issues are raised. Where should these data be stored locally? What transformations must be performed to the code to guarantee that the nonlocal accesses reference the correct memory location? What optimizations can be performed to reduce the time spent in accessing the nonlocal data? In this paper we examine various data migration mechanisms that allow an explicit and controlled mapping of data to memory. We describe, experimentally evaluate, and model a set of schemes for storing and retrieving off-processor array elements. The schemes are all based on using hash tables for efficient access of nonlocal data. The three different techniques evaluated are the basic hashed cache, partial enumeration, and full enumeration, the details of which are described in the paper. In all three schemes, nonlocal data are stored in hash tables—the difference is in the amount of memory used by the schemes and the retrieval mechanisms for nonlocal data.     

An inverse solution methodology for estimating the diffusion coefficient of gases in Athabasca bitumen from pressure-decay data

An inverse solution methodology is developed for the estimation of diffusion coefficient of gases in highly viscous, oil-sands bitumens from isothermal, pressure-decay measurements. The approach involves modeling the rate of change in pressure using the diffusion equation for the liquid phase coupled with a mass balance equation for the gas phase. The inverse solution framework is utilized to arrive at two graphical techniques for estimating the diffusion coefficient. Both techniques involve the determination of the slope of a straight line resulting from plotting the experimental data in accordance with the developed model. An advantage of the proposed techniques is that the diffusion coefficient is estimated directly, i.e. without making it an adjustable parameter. The novelty of the proposed method is in its simplicity as well as its ability to isolate portions of the pressure-decay data affected by experimental fluctuations. The effect of the initial pressure on the predicted diffusion coefficient and pressure-decay profile was also investigated. The diffusion coefficients of CO₂, CH₄, C₂H₆ and N₂ in Athabasca bitumen at 50–90 °C and about 8 MPa were estimated and compared with literature values.     

Collection of biological materials in biodiversity prospecting in India: problems and solutions

Forests are the chief resource for the collection and exploration of biological materials. The past few decades have witnessed a large scale deforestation in India due to substantial pressures generated by population growth, leading to demand for more land for agriculture, urbanization and industrial activities, in addition to increased demand for fuel wood and timber. This has resulted in the loss of soil cover, habitat destruction, environmental degradation and ecological imbalance. This scenario has created a progressive awareness for the conservation and restoration of habitats and, thus, the declaration of many forest areas into protected zones, such as national parks, biosphere reserves, etc., including the protection of some marine areas, by both the National and State Governments. Normally, permission for biological collecting is not granted in these protected areas. In India, forests are a State subject and grant for collection permission is vested with the State Forest Departments. In the absence of any rules, regulations and guidelines, either from National or State Governments, forest authorities impose their terms and conditions, which are arbitrary and even contradictory at times, in the process of granting collecting permits. A set of new rules to be applied throughout the country is needed.     

Mathematical modeling of mixing phenomena in a gas stirred liquid bath

A macroscopic, steady state energy balance model has been formulated to describe mixing phenom-ena in a liquid bath stirred by injecting gas through a straight nozzle fitted axially at the bottom of the vessel. This, along with experimental data on a water model previously reported, was employed to make predictions. Input energy terms considered in the model consist of buoyancy energy and empirically determined fraction of gas kinetic energy. Dissipation of energy was attributed to liquid circulation and bubble slip. The two-phase plume was assumed to be a truncated cone whose dimen-sions depended upon operating conditions. Numerical solution of model equations gave liquid velocity and gas hold-up inside the plume as well as liquid circulation rate and liquid velocity in the region outside the plume. Influence of process variables, e.g., gas flow rate, bath height, and nozzle diameter, have been predicted. Validity of the model has been established by comparing some pre-dicted entrainment ratios with those experimentally measured by other investigators. Empirical cor-relations to predict circulation time and circulation number have been proposed. Circulation number was found to vary between 2 and 12 in contrast to the existing assumption in the literature of a con-stant value of 3. Usefulness of these correlations in predicting mixing time for industrial vessels has been demonstrated. Formerly a Graduate Student in the De-partment of Metallurgical Engineering at the Indian Institute of Technol-ogy, Kanpur     

Viscosity of compressed cold lake bitumen

Viscosity data over a temperature range of 37-115°C and a pressure range of 0-10 MPa gauge are presented for a Cold Lake bitumen sample. These data show that the compression of Cold Lake bitumen results in a significant increase in its viscosity. Results are compared with those from a similar study on an Athabasca bitumen sample. Also presented are two correlations for the effects of pressure and temperature on the viscosity of Cold Lake bitumen.     

Modelling a circulating fluidized bed riser reactor with gas—solids downflow at the wall

A predictive model was developed for the fully developed zone of a circulating fluidized bed (CFB) riser reactor operating in the fast fluidization regime that overcomes limitations of existing models. The model accounts for the upward flow of gas and solids in the core and downward flow of the two phases in the annulus. Additionally, a numerical solution methodology for the simulation of a riser reactor employing the hydrodynamic model was devised. A simulation was performed using the fast, main Claus reaction to demonstrate the effects of backmixing in the fast fluidization regime. It was found that the molar flow rates of the reactants leaving a fast fluidized CFB riser reactor were significantly higher than those leaving an identical reactor operating in the pneumatic transport regime.