-
Familiarize basic python packages for deep learning such as Keras, Tensorflow etc.
-
Data pre-processing operations such as outliers and/or inconsis- tent data value management. **
-
Implement Feed forward neural network with three hidden layers for classification on CIFAR-10 dataset.**
-
Analyse the impact of optimization and weight initialization tech- niques such as Xavier initialization, Kaiming Initialization, dropout and regularization techniques and visualize the change in perform- ance. **
-
Digit classification using CNN architecture for MNIST dataset. **
-
Digit classification using pre-trained networks like VGGnet-19 for MNIST dataset and analyse and visualize performance improve- ment.**
-
Implement a simple RNN for review classification using IMDB data- set.**
-
Analyse and visualize the performance change while using LSTM and GRU instead of simple RNN.**
- Outlier detection
- Linear Classification
- Image classification on MNIST (Any pre-built dataset)
- Image classification on CIFAR-10
- Weight Initialization Optimizer analysis
- Optimizer analysis
- VGGnet-19
- RNN for IMDb
- LSTM and GRU for IMDb
- ANN based image classifier