Machine Learning and Computer Vision

On November 17th and December 1st I attended the Student Scientific Computing Club’s Collaborative workshops on  Artificial Neural Networks and Computer Vision. These workshops were part of a semester long series given on Fridays from 6:00-7:00.

Why would you recommend this workshop to someone else?

These workshops were good as an introduction to machine learning. They direct you to how to get the job done without talking too much about theory. The fact that it is done by peers makes it a much more laid back and fun environment.

What were the goals of the workshop?

This workshop was an overview of Keras package in Python for creating artificial neural networks. We went through an example using a convolutional neural network to read numbers from images.

Why did you choose to attend this workshop?

I have been really excited about learning about one of the greatest biological mysteries to date, the brain.  I read Ray Kurzweil's book called How to Create a Human Mind which talks about how artificial intelligence is a model that helps us understand how the brain is able to achieve high level thoughts and consciousness. One of the coolest forms of AI is artificial neural networks because they work just like our biological neurons do. Recently Google has invested 400 million dollars in studying them and have replaced a lot of their AI systems with better ones implemented with an artificial neural network. Although I will never understand artificial neural networks well enough or have enough data to apply them in a useful way I am excited to see if I can at least make one network run successfully. Although there is a beauty to their simplicity, learning to apply them in a specific way to my research is much harder than the simplest possible statistics which I use. So it is fun for me to learn about and maybe useful for my survival if AI starts displacing jobs and continues to send the U.S. into a modern civil war by addicting people to extremist media to increase ad revenues.

What's the best/coolest thing you learned during the workshop?

Learning that there was a high level easy to implement language that runs Google’s Tensorflow neural network programmjng language was the coolest thing. This makes creating world class Artificial Neural Networks accessible to anyone who knows python or R. I think accessibility to people is very important because Artificial Intelligence is most dangerous at this point in its development when it is used to manipulate people by Google and Facebook and for warfare.

What can you do now that you couldn't before?

I learned how to map the neural networks created in Keras as well as perform drop out to prevent over fitting as well as create more layers of the network.

Thanks to James Proctor and Ben Swerdlow for organizing and leading the workshops and inspiring me to try and tackle making one myself.