Welcome to the Machine Learning Noob's Heaven

The repository contains some of many projects which I had done while I was a kid in Machine Learning. Currently the python notebooks are oriented more towards Deep Learning focussing on some old school DL projects which I found very interesting. For few on them it was tough for me to understand there working and to run them in Google Colab, so I wrote the notebooks trainable in Google Colab to help others. And that's why I invite each of the viewer to contribute to this repository with projects or implementations which you think are important part of Machine Learning pipeline and help ML newcomers to learn more efficiently.

1. CNN

The folder contains Python project Notebook using Convolutional Neural Networks for different purposes. Currently it has following:

• Autoencoder
• Image Classification
• Denoising Autoencoder
• Layerwise output Visualization

Folder can the accessed from here

      1.1 AutoEncoders

     

      AutoEncoders were one of the famous image Compression algorithms. The model consists of an encoder CNN following with a decoder CNN. Folder can the accessed from here

      1.2 Classification

      CNNs are widely used for Image Classification tasks such as Multi-Label Classification, Multi-Class Classification and Single Label Classfication. For now this folder has MNIST Digit Classification. Folder can the accessed from here

      1.3 Denoising Autoencoder

     

      These Deep Learning model architechure is similar to AutoEncoders but are used to serve other purpose. When noisy image is passed through this autoencoder the resultant output is the denoised clear image. Folder can the accessed from here

      1.4 Layerwise output Visualization

     

      While studying CNN, their's always a question in my mind that "What does a node of CNN looks at?". After figuring it out I made this notebook to help others to visualize the CNN node's Vision. Folder can the accessed from here

2. GAN

The folder contains Python project Notebook using Generative Adversarial Networks for different purposes. Currently it has following:

• DCGAN
• Face Generator(GAN)

      2.1 DCGAN

      Deep Convolutional Generative Adversarial Networks use CNN for both of its component i.e Discriminator and Generator. Folder can the accessed from here

      2.2 Face Generator

     

      GANs Application to generate Faces of those people whom neither did you or the neural network has seen before. Folder can the accessed from here

3. MLP

This folder will contain all the Multilayer perceptron implementations. Till Date it has Sentiment Analysis and Self-Implemented MLP model. Folder can the accessed from here

4. Music Generator

An implementation of RNN and LSTMs to listen initial MIDI tones and then continue it to complete whole tune by generating next tones automatically. Folder can the accessed from here

5. RL

This folder will contain old school project implementations using OpenAI gym and Reinforcement Learning. Till Date this contains:

• Dynamic Programming
• Monte Carlo
• Q-Learning

Folder can the accessed from here

6. RNN

This folder will contain project implementations of RNNs and LSTMs. Till Date this contains:

• Sentiment Analysis (RNN)
• Automatic Text Generator

Folder can the accessed from here

7. Word Embedding

Word Embedding consist of a technique which is word to vector (word2vec). This converts each word into a vector and helps in forming relations between words and their meaning.

Folder can the accessed from here

I invite each of the viewer to contribute to this repository with projects or implementations which you think are important part of Machine Learning pipeline and help ML newcomers to learn more efficiently. Thanks for contributing to collective intelligence!!