Hi, I'm Bevan Smith. Welcome to my data science channel.
The aim is to present a relatively thorough overview of introductory machine learning and data science topics.
I hope to present topics on what is machine learning, supervised learning, regression and classification, cross-validation and also to show the viewer actual examples using Python and Scikit-learn.
Eventually I want to present more advanced topics describing linear and logistic regression, decision trees, random forest, boosting and neural networks.
I also plan to present topics on feature selection using various model-agnostic methods such as LIME and SHAP.
I hope you learn something from this channel. Please give me feedback, I would love to hear from you.
Пікірлер
I am watching Machine Learning playlists. Your teaching is full of enthusiasm enabling me to be engaged in your instructions. I really appreciate your wholehearted instruction.
thank you
You are a lengend, Thanks man.
One of the best videos to start with..thanks
Thank you for your videos. Very well explained. However, how the mini batch looks like in a practice? How to put multiple rows in the network in the same time? Do we need bigger layer? Could you provide some details how to do that? Thank you.
Excellent video with neat and clear explanation for a beginner to learn and motivate towards neural networks.
Why does weight updating use a minus sign, instead of a plus sign? 24:34
In gradient descent we want to tweak the weights/biases until we obtain a minimum error in our cost function. So for that we need to compute the negative of the gradient of the cost function, multiply it by a learning rate and add it to the previous value. This negative means we are moving downhill in the cost function (so to speak)
The best i have seen till date .. superb
Great video!
This is the very best video on explaining Back Propagation! It is very clear and well-designed for anyone needing to learn more about AI. I look forward to seeing other videos from Bevan.
Thank youuu the pace is for meee haha
Thank you! I was looking for those exact regression model examples.
After 2 years since publishing, your video is still a gem 💥
THE BEST VIDEO FOR UNDERSTANDING BACK PROPAGATION!!!! Thank you sir <3
Is each mini batch using the average of the losses batch to update its weights and biases? this part is unclear.
however large the batch size is, it calculates a mean squared error (if regression) of those samples
I watch half of the video and I already liked it.
haha South African accent. baie dankie Bevan!
lekker bru
This channel is so underrated...
Hello Bevan Thank you for your excellent videos on neural networks. I have a question pertaining to this video covering Back Propagation. At about 14:30 you present the equation for determining the updated weight, W7. You are subtracting the product of η and the partial derivative of the Cost (Error) Function with respect to W7. However, this product does not yield a delta W7, i.e., a change in W7. It would seem that the result of this product is more like a delta of the Cost Function, not W7, and it is not mathematically consistent to adjust W7 by a change in the Cost Function. Rather we should adjust W7 by a small change in W7. Put another way, if these quantities had physical units, the equation would not be consistent in units. From this perspective, It would be more consistent to use the reciprocal of the partial derivative shown. I’m unsure if this would yield the same results. Can you explain how using the derivative as shown to get the change in W7 (or indeed in any of the weights) is mathematically consistent?
This is amazing, so clear and easy to understand!
So what is the clever part of back prop? Why does it have a special name and it isn't just called "gradient estimation"? How does it save time? It looks like it just calculates all derivatives one by one
it is the main reason why we can train neural nets. The idea in training neural nets is to obtain the weights and biases throughout the network that will give us good predictions. The gradients you speak of get propagated back through the network in order to update the weights to be more accurate each time we add in more training data
Thank you so much , Sir. Best explanation I have seen on this platform .
i just wondering in the last part when i try to calculate partial derivative of w4 the result i got is -3711 but in the video it is -4947. then i make sure so i changed the last equation part to x1 (60) and it gives me the same result like in the video which -2783, so im not sure if i miss something since he didnt write the calculation from w4
Thank you Bevan, nice video. In 5-fold CV, we end up with 5 models. Which model should we use for deployment?
You don't use any of them. I know it sounds confusing. Ok, say now you want to compare three models, a random forest, linear regression and neural net. For each model you perform k-fold CV and average the performance. Then you take the model that gave the best average CV performance and train that model on the entire dataset for deployment. The point about having good CV performance is that it shows that that model will perform best on unseen data. The whole point of cross validation is to see how well a model performs on unseen data. I suggest you have a good long chat with chatGPT to get more detailed answers
Best video I have found for BP! Thanks for all your efforts.
Perfectly explain 😌 finally the best explanation ❤ thankyou sir
just outstanding - re watched it and it made it so clear!
thanks so much for the clarity. helps tremendously! lvoe this.
I like how well you simplify the concepts
May I ask how you can determine the bias to be added in z1 calculation?
Thank you so much. it was so simple
22:53 Why Z1 is equal to -0,5?
Finally i understood something on this topic
very happy to see you back
Really like your teaching style...Any chance you can do a matrix vector calculus version of neural nets when you have finished with reinforcement learning?
It's a very clearly elaborated Q-Learning, I really enjoyed it!
I am just beginner in machine learning but have found it more beneficial your class.
could you please send me the ppt file
By far this is the best explanation. Clear, precise, detailed instructions. Well good and thank you so much 🙏
Fantastic tutorial Bevan, thank you very much!!
You are the KZreadr I have met who can explain all the specific calculation processes clearly and patiently. I appreciate you creating this video. It helps a lot. I wonder if you can make a video about Collaborative filtering?
Man, this is at least the 15th video on the topic I watch, including several books, related to the back propagation, and this is the best one. All previous videos just skip a lot of explanation, focusing on how this backpropagation is important and crucial and what it allows to do, instead of doing the step-by-step overview. This video contains zero bs, and only has clear explanations, thank you
The whole video: "Mini-batch" is like a batch but smaller. No calculus
illegally underrated
Looking forward to this series!
Excellent! was hoping you would be back with some new material.
More to come!
looking forward to it kind sir!
Please do a coding example soon ?
im getting there
I was wondering why didn't you write the active fucnton in your output neuron? is there a specific reason?
Most likely because it is just a linear summation of the inputs. There was no need to pass it through a non linear activation function
Midterm tomorrow, this was the only video in several I watched that finally made me understand linear regression.
Mister, you have saved my life lol, thank you!!!