[Tier 1, Lecture 04c] This video provides a primer on neural networks for machine learning and artificial intelligence. Neural networks are biologically inspired and provide the backbone of many modern ML/AI frameworks.
This video was produced at the University of Washington, and we acknowledge funding support from the Boeing Company
%%% CHAPTERS %%%
0:00 Overview
2:15 What is a Neural Network?
5:17 The Perceptron (History of Neural Networks)
6:39 Deep Learning
8:50 A Diversity of Architectures: the Neural Network Zoo
11:30 CNN: Convolutional Neural Networks
13:11 RNN: Recurrent Neural Networks
14:01 Autoencoder Networks
16:20 Outro
excited for Transformers lecture
The lectures of Professor Brunton are outstanding from all points of view: fachlich, pädagogisch, organisatorisch and, why not, sprachlich (my first language is not English). For me, as a 78 old control engineer, your lectures are really a pleasure... Thank You very much for your knowledge, time and energy
Thank you professor, Best recap for beginners
Glad to hear it!
Very good and informative video as always. I Would really love to see more videos on this and if possible after this a series on CFD and/or FEA.
what do you think are the interesting things in computational fluid dynamics at the moment?
such a great analogy with the periodic table to our current list of models and what kinds of problems they are good for solving. Look forward to the day that we have a nice lookup table, or even better, a NN that looks at our dataset and the problem at hand and gives us a list of potential models and how probable that they are the "best" model to choose for this problem.
Steve thank you very much I follow all of your videos and books, big fan of you! I really enjoy how you explain, I’ve learned a lot.
So ready to dive into this series. Using the biological system analogy, what makes a learning model ‘smart’?Thank you Steve.
Finally a good channel for learning ai! KZhead is filled with opportunists and I'm glad to find this channel thank you so much
Thank u steve for continuing to make wonderful and relevant content
Excellent summary and explanation 👏🏻 Keep up the great work!
Thank you!
Impressive explanation for such a hot topic
Crystal. And needed. Suggests what the math might look like -- enough so to want to go on to the next installment. Thanks so much.
Great lecture!
Lol, I just typed 'convolutional neural network' into KZhead, and then, 3 seconds later, I received the notification about this video :D
awesome video
Great one! I would also be interested in the thought of RNNs for CTR estimations for seasonality considerations.
This serie is gold! Thabk you guus
I am eager to learn more about deep autoenconder !
Great vid, tx.
Thanks for the excellent explanation. Can you share the information about your book that you mentioned in the video?
Really good Got the jist of Neural Nets
If lets say we succeeded in pinning the behavior of neural networks rigorously, what do you think the "physical laws" of neural networks would look like? how can we write them down?
Thank you very much
You're welcome!
@eigensteve. How about creating a new playlist for this Machine Learning Primer ? Thank You for your consideration.
For people who build neural networks, where do they get the data from? are there special repositories that provide datasets?
Hi Steve what level of math do I need to read your engineering mathematics book. Seems like calc 1-3 and lin alg?
Please which logiciel do use to do your presentation like that
Can the model parameters be the weights themselves?
Sir, please also try to make videos on neural operators.
Is it new tutorial and video or it’s the earlier version?
Where can we get these slides?
basically nested giant 'if-else'
Not really… more like routing tables based on computations.
I often read social media comments about the evil things AI will do, and I think, "Other simpler methods can do that now. AI would just get in the way." Of course, telling them so is a waste of time. My recent interest is all the writers suing OpenAI over copyright. Again I think, "If the system is not trained with your intellectual property, it does not take your intellect into account, leading to possible bias." Telling them that is also a waste of time. (:
Huh it seems like people with science and engineering training can use their skills to make neural networks more systematic... like "making a science" out of it