Why we need LSTM networks, how they work step by step, and full explanations: visual and mathematical!
0:00 Problem with Simple RNNs
11:45 Goal of LSTM
12:55 Introducing the Cell State
14:27 Step 1: The Candidate Cell State
15:52 Step 2: The Forget Gate
17:48 Step 3: The Input Gate
18:18 Step 4: The New Cell State
21:15 Step 5: The Output Gate
22:31 Step 6: The New Output State
22:57 Visual Diagram
27:14 Recap all Variables
29:49 Why does this work?
My Patreon : https://www.patreon.com/user?u=49277905
---
RNN Video : https://www.youtube.com/watch?v=DFZ1UA7-fxY
Vanishing Gradient Video : https://www.youtube.com/watch?v=ncTHBi8a9uA
Intro to Neural Networks : https://www.youtube.com/watch?v=xx1hS1EQLNw
Backpropagation : https://www.youtube.com/watch?v=kbGu60QBx2o
---
Icon References:
Dog icons created by Freepik - Flaticon
https://www.flaticon.com/free-icons/dog
Gate icons created by Freepik - Flaticon
https://www.flaticon.com/free-icons/gate
Memory icons created by Freepik - Flaticon
https://www.flaticon.com/free-icons/memory
Crystal ball icons created by Freepik - Flaticon
https://www.flaticon.com/free-icons/crystal-ball
Khao manee cat icons created by Pixel perfect - Flaticon
https://www.flaticon.com/free-icons/khao-manee-cat