Pattern Recognition @ Cgu.Csie.2014

In this semester, the CGU/CSIE Pattern Recognition course will adopt a Mooc course named

Neural Networks for Machine Learning

https://www.coursera.org/course/neuralnets

as the major course to follow up.

We will dictate part of the video lecture on the class and you view and study them in detail at home/dorm. In each class we will have a quiz about the assigned lecture and discuss it later on.
You can find the list of topics covered in this course.

https://class.coursera.org/neuralnets-2012-001/wiki/syllabus

Lecture 1: Introduction

Lecture 2: The Perceptron learning procedure

Lecture 3: The backpropagation learning proccedure

Lecture 4: Learning feature vectors for words

Lecture 5: Object recognition with neural nets

Lecture 6: Optimization: How to make the learning go faster

Lecture 7: Recurrent neural networks

Lecture 8: More recurrent neural networks

Lecture 9: Ways to make neural networks generalize better

Lecture 10: Combining multiple neural networks to improve generalization

Lecture 11: Hopfield nets and Boltzmann machines

Lecture 12: Restricted Boltzmann machines (RBMs)

Lecture 13: Stacking RBMs to make Deep Belief Nets

Lecture 14: Deep neural nets with generative pre-training

Lecture 15: Modeling hierarchical structure with neural nets

Lecture 16: Recent applications of deep neural nets

Video LecturesHelp

Having trouble viewing lectures? Try changing players. Your current player format is html5. Change to flash.

  Lecture1

(expanded, click to collapse)

• CompletedWhy do we need machine learning? [13 min]
  
  

  Slides for Why do we need machine learning? [13 min]

   

  Slides (pdf) for Why do we need machine learning? [13 min]

   

  Subtitles (text) for Why do we need machine learning? [13 min]

   

  Subtitles (srt) for Why do we need machine learning? [13 min]

  Video (MP4) for Why do we need machine learning? [13 min]
• CompletedWhat are neural networks? [8 min]
  
  

  Subtitles (text) for What are neural networks? [8 min]

   

  Subtitles (srt) for What are neural networks? [8 min]

  Video (MP4) for What are neural networks? [8 min]
• CompletedSome simple models of neurons [8 min]
  
  

  Subtitles (text) for Some simple models of neurons [8 min]

   

  Subtitles (srt) for Some simple models of neurons [8 min]

  Video (MP4) for Some simple models of neurons [8 min]
• CompletedA simple example of learning [6 min]
  
  

  Subtitles (text) for A simple example of learning [6 min]

   

  Subtitles (srt) for A simple example of learning [6 min]

  Video (MP4) for A simple example of learning [6 min]
• CompletedThree types of learning [8 min]
  
  

  Subtitles (text) for Three types of learning [8 min]

   

  Subtitles (srt) for Three types of learning [8 min]

  Video (MP4) for Three types of learning [8 min]

  Lecture2

(collapsed, click to expand)

  Lecture3

(collapsed, click to expand)

  Lecture4

(collapsed, click to expand)

  Lecture5

(collapsed, click to expand)

  Lecture6

(collapsed, click to expand)

  Lecture7

(collapsed, click to expand)

  Lecture8

(collapsed, click to expand)

  Lecture9

(collapsed, click to expand)

  Lecture10

(collapsed, click to expand)

  Lecture11

(collapsed, click to expand)

  Lecture12

(collapsed, click to expand)

  Lecture13

(collapsed, click to expand)

  Lecture14

(collapsed, click to expand)

  Lecture15

(collapsed, click to expand)

  Lecture16

--------------------------------------------------------------------

QuizzesHelp

  Lecture1

(expanded, click to collapse)

• Lecture 1 QuizHelp

  Attempt Quiz

  Due Date	Tue 16 Oct 2012 11:59 AM CST If you submit after the due date (but before the hard deadline), your submission score will be penalized 25%.
  Hard Deadline	Tue 30 Oct 2012 11:59 AM CST If you submit any time after the hard deadline, you will not receive credit.
  Effective Score	N/A Your effective score will be the score of your first attempt made before the hard deadline.
  # of Attempts	0 / 1

  Lecture2

(collapsed, click to expand)

  Lecture3

(collapsed, click to expand)

  Lecture4

(collapsed, click to expand)

  Lecture5

(collapsed, click to expand)

  Lecture6

(collapsed, click to expand)

  Lecture7

(collapsed, click to expand)

  Lecture8

(collapsed, click to expand)

  Lecture9

(collapsed, click to expand)

  Lecture10

(collapsed, click to expand)

  Lecture11

(collapsed, click to expand)

  Lecture12

(collapsed, click to expand)

  Lecture13

(collapsed, click to expand)

  Lecture14

(collapsed, click to expand)

  Lecture15

(collapsed, click to expand)

  Final Exam

--------------------------------------------------------------------

Lecture 1 QuizHelp

Warning: The hard deadline has passed. You can attempt it, but you will not get credit for it. You are welcome to try it as a learning exercise.

In accordance with the Coursera Honor Code, I (Renyuan Lyu) certify that the answers here are my own work.

Question 1

We often don't know how much data we will need in order for a learning system to generalize well from training data to test data on a given task.
True or false: when choosing how much data to give to a learning system in order to make it generalize well, we need to make sure that we don't give it too much data.

True

False

Question 2

Data can change over time, in particular we might observe different input/output relationships. In order to account for this we can adapt our learning system to the new data by, for example, training on new examples.

If the relationship between inputs and outputs for old examples has not changed, how can we prevent a neural network from forgetting about the old data?

Ignore the issue and hope that everything will be OK.

Train two networks, one for old data and one for new data.

Prevent the system from changing the weights too much.

Train on a mix of old and new data.

Question 3

Which of the following are good reasons for why we are interested in unsupervised learning?

It allows us to learn from vast amounts of unlabelled data.

It allows academic researchers to publish more papers.

It lets us avoid supervised learning entirely.

It can be used to learn features that may help with supervised tasks.

Question 4

Which of the following tasks are neural networks good at?

Storing lists of names and birth dates.

logical reasoning

Recognizing fragments of words in a pre-processed sound wave.

Recognizing badly written characters.

Question 5

Which number is biggest?

The number of synapes in a human brain.

The number of bits of Random Access Memory (usually just called memory) in a modern laptop.

The number of milleseconds in a human lifetime.

The Greek national debt in euros

Question 6

Which of the following facts provides support for the theory that the local neural circuits in most parts of the cortex all use the same general purpose learning algorithm?

Brain scans show that different functions (like object recognition and language understanding) are located in different parts of the cortex.

The fine-scale anatomy of the cortex looks pretty much the same all over.

If the visual input is sent to the auditory cortex of a newborn ferret, the "auditory" cells learn to do vision.

If part of the cortex is removed early in life, the function that it would have served often gets relocated to another part of cortex.

In accordance with the Coursera Honor Code, I (Renyuan Lyu) certify that the answers here are my own work.

--------------------------------------------------------------------

ryLec001

http://goo.gl/Ir76sX