NLP

Lesson-01

• Review-01
  1. sentence generator (Data driven)
  2. Language model
• Assignment-01
  1. design sentences generator
  2. n-gram language model
  3. Pattern Match (optional)
• Assignment-01-optional
  1. Pattern Match & Segment Match (EN)
  2. chat-robort (CN) -> unfinished

Lesson-02

• Review-02
  1. Transportation route choice (BFS,DFS)
  2. Boston House Prices (sklearn, linear regression)
• Assignment-02
  1. The best subway route in Beijing

Lesson-03

• Review-03

  1. Linear Regression problem using sklearn.linear_model (random data)

  2. Linear Regression problem using KNN Model

  3. Decision Tree (Loss - Entropy&Gini, choosing BestFeature)

  4. A simple example using K-Means

• Assignment-03

  1. Decision Tree (input:data, output:decision)

  2. Linear Regression( change loss function)

Lesson-04

• Review-04
  1. build a FCN from scratch
  2. predict Boston House Prices (i.from scratch ii.keras)
• Assignment-04
  1. binary classification of digits
  2. Challenge: multiclass classification of digits

Lesson-06

• Review-06

  1. Basics of Tensorflow

  2. CNN based on MNIST(i.keras, ii.Tensorflow)

• Assignment-06

  1. CNN based on Cifar-10

  2. build a CNN from scratch (Conv layer, Pooling layer)

Lesson-07

• Assignment-07
  1. Generating names with a character-level RNN (see Pytorch tutorials)
  2. Add more layers in one RNN
  3. Use nn.LSTM and nn.GRU
  4. Connect multiple RNNs to form a more complex RNN

Lesson-08

• Review-08 -> unfinished
  1. Transformer
• Assignment-08
  1. Only questions, no code exercise

Lesson-09

• Assignment-09
  1. Kaggle: What Causes Heart Disease? Explaining the Model

Lesson-10

• Review-10

  1. SVM
  2. Random Forest
  3. XGboost

• Assignment-10 -> unfinished

  1. Build a text classifier if a piece of news is published by the Xinhua news agency
  2. (Option) Try different machine learning algorithms

Lesson-11

• Assignment-11
  1. Cut Rod Problem
  2. Edit distance
  3. Pinyin Auto Correction Problem