Neural Networks Future Directions Video Lecture Transcript
This transcript was automatically generated, so there may be discrepancies between the video and the text.

Hi, everybody. Welcome back in this video. We're gonna talk about some future directions you could take while trying to learn more about neural networks. ... So in this neural network, in these uh series of videos on neural networks, we talked about everything from the uh foundation of neural networks with the Perceptron and the theory of the Perceptron moving into feed forward networks, uh both the theory and how to implement them in S K R and or in car.
Uh Then we, we gave you a brief introduction into convolutional neural networks, both the theory and in Caris as well as recurrent neural networks in the theory. And in Caris, uh we also showed you how to load a model once you've trained it and saved it. And uh so you can come back to your model when you're done training it, which can take a long time.
So before we wrap up 100% on neural networks, I wanted to touch on some future directions in case you're looking to learn more. Uh So what have we learned? We talked about this? What is left to learn? Uh There's a lot to learn left in neural networks. This has been a very active area in the past 5 to 10 years. So theoretically, there are things known as auto encoders and transformers.
Uh There are things called generative adversarial networks transfer learning. And more uh practically speaking, you may want to learn more about Caris, you may wanna learn more about tensor flow, which is what Caris is built on top of. And you may also want to learn more about Pytorch, which is another Python package used for building neural networks.
Um We have used primarily what is known as our, our computers C P U central processing unit. Uh But you may want to learn how you can use uh GP us in order to uh more quickly train uh neural networks. Um And you may also want to know how can I train or implement neural networks on a server or a cloud computing environment. So this is a lot left to learn both theoretically and practically.
And I'm gonna do my best to give you uh where you might want to start looking to continue learning. Uh So theoretically, I really do, I've, I think I've linked to it in every book. I really like this book or every notebook. Uh I really like this book Neural Networks and Deep learning. I think it gives a good foundation. Uh It is a little bit math heavy.
Uh But if you're looking for a good theoretical foundation on um neural networks, it covers deep learning, convolutional networks and recurrent neural networks, as well as some other topics uh for applied. Um There's deep learning with Python. So we did sort of go over some of the most introductory materials of this book with our implementation of things in cars.
But there's also like a lot of additional stuff, like some nice tips and tricks for how to build uh convolutional neural networks with not a lot of data. Uh More examples with recurrent neural networks. Um There's also a newer edition of the book which I'm sure has um additional material in it that we did not cover. Uh This is also just a good general purpose Python machine learning book that does also uh talk about uh implementing neural networks uh using Caris and tensor flow.
Um I suggest that you check this out just because it's one of my favorite books on machine learning and data science. Uh And there's also a book Deep Learning with Pie Toch, which I think is sort of like the P to P toch version of the Deep Learning with Python. It's written by somebody else. I haven't looked at it yet, but it's uh comes from the same people who publish Deep Learning with Python.
Uh Once you get through those or if maybe you want to dive straight into the documentation, I've also included links to all of the documentation for car tensor flow and Pie Torch here. Uh So if you're interested in neural networks. I hope that you, these are good avenues for you to start learning more. Uh You can also just do web searches and see if you can pop up on your own.
Um But I hope that we've provided you with a good foundation to learn uh more about neural networks. Um uh Yeah, so that I hope you enjoyed learning about this and I hope you're able to continue learning at a, at a good pace. All right, have a great rest of your day. I hope you enjoyed watching all of our neural network content. Bye.