[Deep Learning Specialization] Neural Networks and Deep Learning - week 1:: seoftware

📜 강의 정리 

* Cousera 강의 중 Andrew Ng 교수님의 Deep Learning Specialization 강의를 공부하고 정리한 내용입니다.

* 영어 공부를 하려고 영어로 강의를 정리하고 있습니다. 혹시 틀린 부분이나 어색한 부분이 있다면 댓글로 알려주시거나 넘어가주시면 감사하겠습니다

---

What you'll learn

1. Neural Networks and Deep Learning
2. Improving Deep Neural Networks : Hyperparameter tuning, Regularization and Optimization
3. Structuring your Machine Learning project
4. Convolutional Neural Networks
5. Natural Language Processing: Building sequence models

 

In specialization,

there is a process to build a cat recognizer in first course(1).

Then in the second course(2), I'll learn about the practical aspects of deep learning - how to actually get my network I've built to perform well. 

The strategy for building a machine learning system has changed in the era of deep learning. I'll learn about the new best practices learning and end to end deep learning in the third course(3).

In the fourth course(4), I'll learn about CNN.

In the fifth course(5), I'll learn about RNN or LSTM for NLP

---

What is Neural Network?

Neural Network example (y is price)

Neural Networks figure out functions that accurately map from x to y. The way you manage neural network is that when you implement it, you need to give it just the input x and the output y for a number of examples in your training set and all this things in the middle(density layer), they will figure by itself.

So, in the example of "Housing Price Prediction", the input features might be the size, number of bedrooms, the zip code or postal code, and the wealth of the neightborhood. And so given these input features, the job of the neural network will be to predict the price y.

---

Supervised learning with Neural Networks

In supervised learning, people give datasets which consist of input x and output y to network.

There are some examples of neural networks(NN) - Standard NN, Convolutional NN, Recurrent NN, Hybrid NN

 

Datasets divided into two types. One is "Structured", the other is "Unstructured".

• Structured data has a very well defined meaning such as size of the house, the number of bedrooms.
• Unstructured data refers to things like audio, raw audio, or images where you might want to recognize what's in the image or text.

---

Why is deep learning taking off?

 

• Main drivers behind the rise of deep learning
  • 1. Data - Huge amount of data thanks to the digitalization of society
  • 2. Computation, Hardware(GPU) - Environment to be able to train a big enough NN
  • 3. Algorithms - ReLU activation function makes neural networks run much faster