Course: MIT 6.S191: Introduction to Deep Learning

Video Lecture: https://youtu.be/QvkQ1B3FBqA

Lecturers: Alexander Amini and Ava Soleimany

Topics:

• Deep sequence modeling
• Intuition and Applications of RNNs
• Drawbacks of RNNs
• Self-Attention and Transformers

Sequences in the Wild

Sequential data is all around us. There is audio (sequence of sound waves), text (sequence of characters/words), stock prices, DNA sequences, and more.

Applications of Sequential Modeling in the Real-World

In contrast to the typical FFNN problems we learned about in the previous lecture (e.g., binary classification), sequential models allows for a variety of different types of problems (e.g., sentiment classification, image captioning, machine translation, etc.).