Global ETD Search

Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.

Results

Showing 1 to 20 of 51 for “"forward propagation"”.

  1. Enhanced Neural Network Training Using Selective Backpropagation and Forward Propagation

    … measures to determine when to backpropagate and forward propagate in order to reduce training time while maintaining classification performance. This thesis tests these new algorithms on the MNIST and CASIA datasets, and achieves successful results with both algorithms on the two datasets. The …

    vt Repository record for Enhanced Neural Network Training Using Selective Backpropagation and Forward Propagation (opens in a new tab)

  2. Image processing and forward propagation using binary representations, and robust audio analysis using deep learning

    … computation can also be used for efficient forward propagation in deep neural networks. Quantified deep neural networks have recently been proposed with the goal of improving computational time performance and memory requirements while maintaining as much as possible classification …

    uvic Repository record for Image processing and forward propagation using binary representations, and robust audio analysis using deep learning (opens in a new tab)

  3. Bottom backscatter mapping: the effect of secondary multipath interactions

    … of this type of analysis would be the removal of propagation effects resulting in a large-scale mapping of scattering strengths. Furthermore, analysis with a quasi-CW propagation model suggests a strong correlation between the small-scale fluctuations in the reverberation signal and the bottom …

    nps Repository record for Bottom backscatter mapping: the effect of secondary multipath interactions (opens in a new tab)

  4. Forward scattering of a pulsed continuous wave signal through laminar and turbulent thermal plumes

    The results of an experiment examining the forward propagation of an acoustic signal through a buoyant plume are discussed. Two distinct testing sights were used. One made use of a small fresh water tank in NUWC to provide a controlled plume. The other us

    nps Repository record for Forward scattering of a pulsed continuous wave signal through laminar and turbulent thermal plumes (opens in a new tab)

  5. Knowledge Distillation Circumvents Nonlinearity for Optical Convolutional Neural Networks

    … As such, CNNs require fast runtime (forward propagation) to process high-resolution visual streams in real-time. This is still a challenging task even with state-of-the-art graphics and tensor processing units. The bottleneck in computational efficiency primarily occurs in the …

    washington Repository record for Knowledge Distillation Circumvents Nonlinearity for Optical Convolutional Neural Networks (opens in a new tab)

  6. Analysis of online learning algorithms in machine learning

    … estimate is simultaneously updated using forward propagation of the SDE state derivatives, which asymptotically converges to the direction of steepest descent. We rigorously prove convergence of the online forward propagation algorithm for linear SDE models and present its numerical …

    oxford Repository record for Analysis of online learning algorithms in machine learning (opens in a new tab)

  7. Vortex burst behavior of a dynamically pitched delta wing under the influence of a von Kàrmàn vortex street and unsteady freestream

    … negative effect (i.e., an acceleration of the forward propagation velocity) on the burst location. This negative effect was almost independent of the actual acceleration or range of a over which it occurred. Deceleration, on the other hand, was found to delay the forward burst movement along …

    wichita-thes Repository record for Vortex burst behavior of a dynamically pitched delta wing under the influence of a von Kàrmàn vortex street and unsteady freestream (opens in a new tab)

  8. Prediction of the beamformed acoustic ray arrival structure for the 1992 Barents Sea coastal tomography test

    In an effort to solve the forward propagation problem associated with the 1992 Barents Sea Polar Front Experiment Tomography Test, the transmission of a 224 Hz pulse signal from a near bottom sound source to a vertical hydrophone array was simulated based on three-dimensional ray theory. based on …

    nps Repository record for Prediction of the beamformed acoustic ray arrival structure for the 1992 Barents Sea coastal tomography test (opens in a new tab)

  9. Forward and Back-Propagation with an Analog Neural Network

    … The methodology for the circuits consists of forward propagation, calculating new weights, and backpropagating the weights until trained. It is observed that with the trained weights for a pair of inputs, a new pair of similar inputs can predict desired outputs. As a specific use case, the …

    carleton Repository record for Forward and Back-Propagation with an Analog Neural Network (opens in a new tab)

  10. Learning Hyperparameters for Inverse Problems by Deep Neural Networks

    … parameters for newly-obtained data by forward propagation through the DNNs. The network-obtained regularization parameters can be computed more efficiently and may even lead to more accurate solutions compared to existing regularization parameter selection methods. Numerical results for …

    vt Repository record for Learning Hyperparameters for Inverse Problems by Deep Neural Networks (opens in a new tab)

  11. Efficient multidimensional uncertainty quantification of high speed circuits using advanced polynomial chaos approaches

    … integrated microsystems. Hence, modeling the forward propagation of uncertainty from the device-level parameters to the system-level response of high-speed circuits and systems forms a crucial requirement of modern computer-aided design (CAD) tools. This thesis presents novel approaches based …

    colostate Repository record for Efficient multidimensional uncertainty quantification of high speed circuits using advanced polynomial chaos approaches (opens in a new tab)

  12. THE INFLUENCE OF A THERMALLY BLOOMED ATMOSPHERE ON TARGET IMAGE QUALITY

    … are determined using a physical optics propagation code. The wavelength of the return radiation used to sense the distortions is shown to be an important consideration in the stability of the return image. The return targets images for several realistic scenarios are reconstructed based …

    arizona-thes Repository record for THE INFLUENCE OF A THERMALLY BLOOMED ATMOSPHERE ON TARGET IMAGE QUALITY (opens in a new tab)

  13. Rough Surface Scattering and Propagation over Rough Terrain in Ducting Environments

    The problem of rough surface scattering and propagation over rough terrain in ducting environments has been receiving considerable attention in the literature. One popular method of modeling this problem is the parabolic wave equation (PWE) method. In this method, the Helmholtz wave equation is …

    vt Repository record for Rough Surface Scattering and Propagation over Rough Terrain in Ducting Environments (opens in a new tab)

  14. Fast algorithms for frequency domain wave propagation

    … fields, numerical simulation and modeling of the forward propagation problem is important to the design and analysis of many systems; a few examples which rely on these computations are the development of metamaterial technologies and geophysical prospecting for natural resources. There are two …

    texas Repository record for Fast algorithms for frequency domain wave propagation (opens in a new tab)

  15. Performance assessment of multi-walled carbon nanotube interconnects using advanced polynomial chaos schemes

    … such issues, simulation solvers to model forward propagation of uncertainties or variations in random processes at the device level to the network response are required. Polynomial Chaos Expansion (PCE) of the random variables is the most common technique to model the unpredictability in …

    colostate Repository record for Performance assessment of multi-walled carbon nanotube interconnects using advanced polynomial chaos schemes (opens in a new tab)

  16. Studying CNN representations through activation dimensionality reduction and visualization

    … methods for studying DNN representations. During forward propagation, each DNN layer non-linearly transforms the input space in some way that is useful for minimizing a loss function. To understand how DNNs represent their inputs, this work develops methods to examine each DNN layer’s activation …

    waterloo Repository record for Studying CNN representations through activation dimensionality reduction and visualization (opens in a new tab)

  17. Superscalar RISC-V Processor with SIMD Vector Extension

    … on the Intel Arria 10 FPGA board. The forward propagation of a convolution neural network model is evaluated by the standalone superscalar processor and the integration of the vector co-processor. The vector program with software-level optimizations achieves 9.53× improvement on …

    sask Repository record for Superscalar RISC-V Processor with SIMD Vector Extension (opens in a new tab)

  18. Privacy-aware distributed machine learning

    … ``frozen" clients, capable only of forward propagation, can contribute their data to server-side training, demonstrably improving overall model performance. Finally, we evaluate the Muon optimizer for use as an outer-loop, or federated optimizer. Muon orthogonalizes momentum updates, …

    manitoba Repository record for Privacy-aware distributed machine learning (opens in a new tab)

  19. The Influence of Lifting Condensation Level on Low-Level Outflow and Rotation in Simulated Supercell Thunderstorms

    … becomes more negatively buoyant, there is more forward propagation of near-surface circulation relative to the mesocyclone aloft. It is shown that there is a negative, statistically significant lag correlation between the separation of these mid- and low-level circulations and subsequent …

    tamu Repository record for The Influence of Lifting Condensation Level on Low-Level Outflow and Rotation in Simulated Supercell Thunderstorms (opens in a new tab)

  20. All-optical ultrafast switches based on thin film coated-tilted fiber Bragg grating

    … tilted fiber Bragg gratings (TFBG), coupling the forward propagation core mode to hundreds of discrete backward propagation cladding modes provide a novel platform for light-matter interaction by modulation of material permittivity via the evanescent field of cladding modes which in turn modulate …

    carleton Repository record for All-optical ultrafast switches based on thin film coated-tilted fiber Bragg grating (opens in a new tab)

Page 1 of 3