/ai-practice

Some practices on AI so I can remember what I am doing through the learnings

Primary LanguagePython

AI practice

I will be doing some learning and personal projects so for keeping those learnings I created and shared this repo.

Pactice 0: Implemented from scratch

  • simplest-ann.py: Here I implemented the simplest ANN I am aware of, that learns how to multiply by two
  • simple-neuron.py: Here a neuron implemented, in this case, using several weights and doing the np.dot product, etc. in a more generic way

Practice 1: Supervised models

  • recognizing-letters.py: Implements a simple digit recognizer based on one database already available in keras called MNIST
  • recognizing-letters-hidden.py: Implements a simple digit recognizer based on one database already available in keras called MNIST (adding some hidden layers to see how performance improves)
  • sentiment-analysis.py: Recognizes how a movie is being perceived by reading IMDB comments and can predict based on comments

Practice 2: Word embedding

  • word-embedding.py: For understanding how to use the concept of embedding words into vectors and how those vectors are modified to show how close words are related learned from a corpus