Pycon-VQA-Theano
PyCon Workshop on training VQA models with Theano/Keras
Workshop outline :
- Recap : Basic Neural Networks with Theano
- Under the hood : Theano optimizations
- The VQA Problem
- Compile and train end-to-end networks
- Observations and results with VQA
- Experiments and Visualizations on pre-trained model
- Introduction to Keras
- Training own models on MNIST for character recognition
Prerequisites :
While most of the topics are introduced in the talk from scratch, familiarity with the following would definitely help :
- Introductory Machine Learning
- Symbolic Algebra ( focus on Symbolic Differentiation )
- Introductory Calculus, Linear Algebra
- Containers in Python
- Convolution Neural Networks ~ Recurrent Neural Networks
For hands-on :
Resources :
Workshop Resources
Neural Networks :
Theano :
- Official Theano documentation
- Getting started with Theano
- Theano : A Python framework for fast computation of mathematical expressions
Deep Learning References :
- VQA : A walkthrough blog by Avi Singh
- Recurrent Neural Networks : Effectiveness of RNNs by Andrej Karpathy
- Convolution Neural Networks : Introductory blog by Christopher Olah