Deep Learning Essentials

7 - Convolutional Networks
1. Handling
2. Layer
  1. Activation
  2. Padding
    1. k = (d-m+2p)/s+1
      Annotations:
      - Wird noch korrigiert. Falsche Formel
  3. Stride
    1. Employes a vertical and horizonal axis
    2. The number of steps a kernel is moved over the input activation matrix is called stride s.
3. Pooling
  1. reduce the dimensionlity
4. Convolution
5. AlexNET - 2012
6 - Model Assessment
5 - Model Training
4 - Weight Adaptation
3 - Multilayer Perceptron
2 - Linear Models
1. Supervised Learning
  1. Learning with labels
    1. each pattern x has a label information y
    2. pair (x_i, y_i)
      1. training set
      2. ground truth
    3. pair (x'_i, y'_i)
      1. predict set
    4. If the label is discrete, e.g., {0, 1} or {muffin, chihuahua}, the learning problem is called classification
      1. classificatoin
    5. if hoices is explored ol detection. continuous First, (y ∈ R) it is called regression.
      1. regression
2. Linear Regression
  1. found in natural and technical processes
  2. The basic linear model (1)
    1. x € IR
    2. weight factor
      1. w € IR called slope
    3. parameter
      1. b € IR called inter
    4. linear relationship
  3. Least Squares
    1. With First, the least squares formulation, the coefficients can be derived. (2)
    2. means squared error (MSE)
  4. Linear Regression Coefficients
    1. Weight and intercept can be mathematically derived as (3)
      1. x Strich mens x_1 ... x_n and the same for the label y Strich
  5. Example Fit, Illustration of linear model that is fiied to the patterns minimizing the MSE
3. Nearest Neighbors
  1. K-nearest neighbors (kNN) searches for labels based on nieghborhoods in data space.
1 - Introduction
1. A.I.
  1. Intelligence is
    1. learn from observations
      1. others experiences
      2. own experiences
  2. related to
    1. Data Science
    2. Big Data
2. Technologies
  1. deep learning
    1. TFlearn
    2. Keras
    3. Tensorflow
  2. mashine learning
    1. scikit-learn
  3. share on github
  4. devlopment of scripts
    1. Jupyter
  5. Research
    1. archivx
  6. Powerfull Hardware
    1. AWS
    2. NVIDIA GRPUs
8 - Neuroevolution
1. Genetic Alogrithm
  1. mimicking biological evolution
    1. Crossover
    2. Mutation
    3. Selection
  2. Optimization
  3. examples
9 - Auto-encoder
10 - Generative Adversarial Networks

Next up

Deep Learning Essentials

Description

Resource summary

Media attachments

Similar

	Created by Mark Otten over 6 years ago