M13.2 CONNECT THE MECHANISM
Carry a state from one step to the next
A network that reads the café's day one hour at a time and keeps a running hunch about it. Then watch why its memory of 7 a.m. fades by noon.
LESSON OVERVIEW14 min lesson
Lesson overview
A network that reads the café's day one hour at a time and keeps a running hunch about it. Then watch why its memory of 7 a.m. fades by noon.
What you’ll explore
- Recurrent networks update a hidden state using new input and prior state; shared parameters process sequences, while backpropagation through time handles their repeated use.
GO TO THE SOURCE
Original explanations, connected to the research.
Dive into Deep Learning — recurrent neural networksDive into Deep Learning — backpropagation through timeFinding Structure in Time (Elman, 1990)Learning Long-Term Dependencies with Gradient Descent is Difficult (Bengio, Simard & Frasconi, 1994)On the Difficulty of Training Recurrent Neural Networks (Pascanu, Mikolov & Bengio, 2013)DeepAR: Probabilistic Forecasting with Autoregressive Recurrent Networks (Salinas et al., 2017)Suggest a correction
A precise note can make an explanation better.
Choose the scene and describe what needs attention. Download a feedback file to share through a channel you already use. This page does not send feedback or connect you with a reviewer.