Tetiana will kick off with an 1 hour theoretical presentation on Neural Networks, upon which we will build our understanding on what an RNN is and how it is different from a standard NN. We will also try to understand what an LSTM is and outline its advantages over a standard RNN. Following theory, Tetiana will give a practical walkthrough of building an RNN and will analyse a sample data set.
In this meetup, we will only have enough time to cover Recurrent Neural Networks. If you’d like to learn Convolutional neural networks, join the next meetup in april
You’ll get more benefit from this talk, if you have a prior understanding of how standard feedforward neural networks work, how back propagation works etc.
Don’t have that understanding yet? Not to worry! Here’s a reading list, you can have a look through. We particularly recommend the background material on NNs:
For those of you who haven’t implemented a basic Neural Network before, I would highly recommend going through the materials of this course. This is a resource for background knowledge, it doesn’t cover RNNs or CNNs.
YOU MAY ALSO LIKE:
Recurrent and Convolutional Neural Networks - Part 1
Tetiana is a mathematician turned data scientist currently working with NanoTechGalaxy on developing machine learning algorithms for medical image processing. She is also working on AI risk research as part of the Pareto Fellowship awarded by the Centre of Effective Altruism.