Sökning: "network symmetries"

Visar resultat 1 - 5 av 6 avhandlingar innehållade orden network symmetries.

1. 1. Network Parameterisation and Activation Functions in Deep Learning

   Författare :Martin Trimmel; Matematik LTH; []
   Nyckelord :NATURVETENSKAP; NATURAL SCIENCES; NATURVETENSKAP; NATURAL SCIENCES; NATURVETENSKAP; NATURAL SCIENCES; deep learning; linear region; network parameterisation; activation function; network calibration; conformal predictnio; tropical algebra; rational function; temperature scaling; network symmetries;

   Sammanfattning : Deep learning, the study of multi-layered artificial neural networks, has received tremendous attention over the course of the last few years. Neural networks are now able to outperform humans in a growing variety of tasks and increasingly have an impact on our day-to-day lives. LÄS MER

3. 2. Mathematical Foundations of Equivariant Neural Networks

   Författare :Jimmy Aronsson; Chalmers tekniska högskola; []
   Nyckelord :NATURVETENSKAP; NATURAL SCIENCES; symmetry; geometric deep learning; gauge theory; induced representations; fiber bundles; convolutional neural networks; equivariance;

   Sammanfattning : Deep learning has revolutionized industry and academic research. Over the past decade, neural networks have been used to solve a multitude of previously unsolved problems and to significantly improve the state of the art on other tasks. However, training a neural network typically requires large amounts of data and computational resources. LÄS MER

4. 3. Applications of Diversity and the Self-Attention Mechanism in Neural Networks

   Författare :Axel Berg; Matematik LTH; []
   Nyckelord :NATURVETENSKAP; NATURAL SCIENCES; NATURVETENSKAP; NATURAL SCIENCES; point cloud; self-attention; keyword spotting; deep learning; computer vision; machine learning; label diversity;

   Sammanfattning : This thesis covers three contributions in applications of neural networks. The first is related to diversity and ensemble learning, while the other two cover novel applications of the self-attention mechanism. An important aspect of training a neural network is the choice of objective function. LÄS MER

5. 4. Low Level Operations and Learning in Computer Vision

   Författare :Björn Johansson; Gösta Granlund; Rudolf Mester; Linköpings universitet; []
   Nyckelord :TECHNOLOGY; TEKNIKVETENSKAP;

   Sammanfattning : This thesis presents some concepts and methods for low level computer vision and learning, with object recognition as the primary application.An efficient method for detection of local rotational symmetries in images is presented. Rotational symmetries include circle patterns, star patterns, and certain high curvature patterns. LÄS MER

7. 5. The Multimodal Transfer Matrix Method : And its application to higher-symmetric periodic structures

   Författare :Pilar Castillo-Tapia; Oscar Quevedo-Teruel; Paolo Bacarelli; KTH; []
   Nyckelord :TEKNIK OCH TEKNOLOGIER; ENGINEERING AND TECHNOLOGY; Electrical Engineering; Elektro- och systemteknik;

   Sammanfattning : This thesis focuses on the understanding and computation of the dispersion properties of periodic structures possessing higher symmetries with the multimodal transfer matrix method (MMTMM). Periodic structures with higher symmetries are invariant after additional symmetry operations over the translation operation. LÄS MER