Kognitive Systeme, SS2017, Vorlesung

06 |

0:00:00 Starten

0:00:22 Neural Networks

0:00:29 The brain is:

0:03:19 Von Neumann Computer - Neural Computation

0:14:34 Coding of Speach: ""Bottleneck Features""

0:15:34 Using Neural Nets

0:16:56 Neural Models

0:17:30 Apllications

0:19:15 Advanced Neural Models

0:19:29 Time Varying Patterns

0:24:02 Classical Human - Computer Interaction

0:25:16 Better Human-Machine Interaction

0:26:29 Interpreting Human Communication

0:27:46 Humans and Computers

0:29:38 Speech

0:30:05 A Few Related Sciences

0:33:01 Anatomy of Speech Production

0:37:17 Speech Production

0:45:04 Convolution

0:45:32 Trandfer Functions of the Diffrent Components of Speech Production

0:50:36 Diffrent Vocal Tract Shapes

0:52:39 Formants

0:57:43 Spektogramme

1:04:00 Vokale im Zeitbereich

1:05:11 Consonants

1:07:01 Vocal Tract Shapes of Consonants Fricatives

1:07:48 Vocal Tract Model of Speech

1:12:31 Dimensions of Difficulty

1:19:38 Speech: State-of-the Art

05 |

0:00:00 Starten

0:00:06 Decision Function g(x)

0:03:26 Perceptron - Nonlinearities

0:04:11 The Perceptron

0:05:44 Linear Discriminant Functions

0:06:56 Perceptron

0:08:36 Perceptron Criterion Function

0:15:43 Linearly Separable Samples and the Solution Region in Weight Space

0:20:04 Finding a Solution Region by Gradient Search

0:24:40 Perceptron Learning

0:34:10 Variations

0:35:07 Effect of the Margin on the Solution Region

0:35:35 Nonseparable behavior

0:42:32 Network of Neurons/ Multi-Layer Perceptron

0:45:21 Connectionist Units

0:48:53 Training the MLP by Error Back-Propagation

0:53:23 Backpropagation of Error

0:56:19 Derivative dy/dx

1:05:41 Neural Network Demo

04 |

0:00:00 Starten

0:05:02 Non-Parametric Techniques: Parzen Windows

0:10:45 K-Nearest Neighbors (KNN)

0:19:28 KNN-Classifier: Problem

0:20:50 Decision Function g(x)

0:24:35 Linear Discriminant Functions

0:35:49 Unsupervised Learning

0:38:42 Unsupervised Classification

0:58:31 Clustering

1:19:18 Linear Discriminant Functions

1:21:08 Perceptron Criterion Function

03 |

0:00:00 Starten

0:00:11 Classification Problem

0:07:56 Bayes Decision Theory

0:20:49 Classifier Design in Practice

0:24:59 GaussianClassifier

0:51:12 Principal Component Analysis (PCA)

1:07:28 Minimum Error Rate Classification

1:11:38 Non-Parametric Techniques: Parzen Windows

02 |

0:00:00 Starten

0:01:23 Abtasten

0:01:52 Digitalisierung von Signalen

0:02:17 Quantisierung

0:02:31 Faltung

0:03:04 Idee der Fourier Transformation und Fourier-Reihen

0:03:41 Fourier Reihenzerlegung

0:04:54 Fourierreihen für Rechteckfunktion

0:05:18 Fouriertransformation

0:08:01 Eigenschaften von Fouriertransformierten

0:10:53 Faltung

0:18:49 Zusammenhänge

0:21:45 Typische Fouriertransformationen 1

0:25:11 Fourier Reihen und Transformationen

0:32:38 Typische Fouriertransformationen 2

0:35:15 Von kontinuierlich nach diskret

0:37:59 Aliasing

0:41:17 Abtast/Sampling Theorem

0:42:23 Abtasten in der Praxis

0:43:11 Beheben von Aliasing

0:45:06 Korrelation

0:50:41 Schablonenanpassung (Template Matching)

0:56:00 Klassifikation 1+2

1:09:45 Classification Problem

1:14:42 Bayes Decision Theory

1:19:37 Two classes case:

1:20:54 Hypothetical class-conditional probability density function

1:21:22 A posteriori probabilities

1:22:13 Examples Decision boundraries

01 |

0:00:00 Starten

0:00:10 Digitalisierung von Signalen

0:03:34 Abtasten / Sampling und Rekonstruktion

0:04:26 Sampling - Aliasing

0:05:55 Beheben von Aliasing

0:11:10 Quantisierung

0:15:58 Quantisierung - Bilder

0:17:28 Diracfunktion

0:21:25 Faltung

0:30:32 Demo

0:33:51 Convolution Demo

0:43:36 Beispiel Faltung

0:45:19 Idee der Fourier Transformation und Fourier-Reihen

0:55:26 Fourier Reihenzerlegung

1:02:25 Fourier series demo

1:07:14 Fourierreihe

1:08:19 Fouriertransformation