MIT 6.S191: Recurrent Neural Networks, Transformers, and Attention

I just can't believe how amazing the educators are and damn !! they're providing it out here for free...
Hats off to the team !!

I am a Professor and this is the best course I have found to learn about Machine learning and Deep learning....

Watching those MIT courses alongside course at my Uni in Poland, so grateful to be able to experience such a high quality education

Thank you so much MIT and instructors for making these very high quality lectures available to everyone. Students from developing countries who have aspirations to achieve something big is now possible with this type of content and information!

One of the best lectures I have seen on Sequence Models, with crystal clear explanations! :)

Over all videos on KRplus that explained about Transformer architecture (including the visual explanation) , this is the BEST EXPLANATION ever done. Simple, contextual, high level, step by step complexity progression. Thank you the educators and MIT!

Best end to the lecture: “Thank you for your attention.” ❤😂

These are some spectacular lessons. Thank you very much for making this available.

I watched and read a lot of content about Transformers and never understood what are those three Q, K, and V vectors doing so I coulnd't understand how attention works, until today when I watched this lecture doing the analogy of KRplus search and the Iron Man picture. Now it became much much clearer! Thanks for the brilliant analogies that you are making!

This is what we need in this day and age, the teaching is amazing and can be understood by people of variable intelligence. Nice work and thanks for this course.

Summary by Gemini:
The lecture is about recurrent neural networks, transformers, and attention.
The speaker, Ava, starts the lecture by introducing the concept of sequential data and how it is different from the data that we typically work with in neural networks. She then goes on to discuss the different types of sequential modeling problems, such as text generation, machine translation, and image captioning.
Next, Ava introduces the concept of recurrent neural networks (RNNs) and how they can be used to process sequential data. She explains that RNNs are able to learn from the past and use that information to make predictions about the future. However, she also points out that RNNs can suffer from vanishing and exploding gradients, which can make them difficult to train.
To address these limitations, Ava introduces the concept of transformers. Transformers are a type of neural network that does not rely on recurrence. Instead, they use attention to focus on the most important parts of the input data. Ava explains that transformers have been shown to be very effective for a variety of sequential modeling tasks, including machine translation and text generation.
In the last part of the lecture, Ava discusses the applications of transformers in various fields, such as biology, medicine, and computer vision. She concludes the lecture by summarizing the key points and encouraging the audience to ask questions.

Extremely informative, well structured and paced. A pleasure to watch and follow. Thank you.

As a CS student from University of Tehran, you guys don't have any idea how much such content could be helpful and the idea that all of this is free make it really amazing. Really appreciate it Alexander and Ava. Best hops.

Indeed commendable the way this lecture has been ordered and difficult topic like self-attention has been lucidly explained. Thanks to the instructors, really appreciated.

Your explanation of attention took me 2 revisits to this video to truly truly understand! But now when I did, my love for deep learning got stronger :)

Just watched lecture 1, looking forward to this and the lab coming after. Thanks for this great open resource!

Thank you for this amazing content! There are many concepts discussed intuitively!

These lectures are simply amazing. Thank you so much!

15:05 we have different weights matrix for generating h_t and generating y_t
h_t generated using two different weights matrix, to take contribution from previous state and current input
51:20 start of attention explanation
59:30 each attention head focus on some part similar to how each filter in cnn can learn to extract specific features like horizontal lines etc

This is incredible! Thanks a lot for this video, it’s going to help me a lot in my undergrad reasearch :)

This is my favorite subject :)
(following is self clarification of said words that feel exaggerated)
4:08 - binary classification or filtering is a sequence of steps:
- new recording
- retrieval of a constant record
- compare new and constant record
- express a property of the compare process
So, sequencing really is a property of maybe all systems.
While "wave sequencing" is built on top of a Sequencer System, that repeatedly uses the
"same actions" per sequence element.

50:30 - Attention mechnaism beautifully explained. Thank you #AvaAmini

She absolutely killed it. Amazing lecture(r)!

Came here to refresh my memory of deep learning for sequential data. I really like how Ava brings us from one algorithm to another. It makes perfect sense to me.

I am trying to step into deep learning for last couple of month. This is the best thing I have found so far. Thank you sir!.

The most intutive explanation of Self Attention I have seen!

I have used LSTM and Transformer a lot, but I can still get more insights from this lecture.

amazing lecture series, thanks for sharing this knowledge with the world. I am curious if theres a lecture on LSTM'S

This was the third video I watched in search of understanding what transformers are, and by far the best one. Thanks.

Wow, Transformers, and Attention was an absolute lifesaver! 🚀🙌 The explanations were crystal clear, and I finally have a solid grasp on these concepts. This video saved me so much time and confusion. Huge thanks to the Ava for making such an informative and engaging tutorial! Can't wait to delve deeper into the world of AI and machine learning. 🤖💡

Grateful for the efforts of MIT and its incredible professors delivering high quality free lectures. Filling every gap I have in my current classes ❤

Wonderful, easy to focus and understand :). Great quality! Grateful that this is open source!

ty to MIT for giving back a little in an impactful way

Great Presentation @8:00 minutes it really explained a circuitry I was looking forward to exploring

Amazing course! Thank you so much!

Great lecture, learnt a lot. Thank you for sharing!

I was searching about RNN for my Thesis work.She solved it...Nice Miss:)

How beautifully explained. Loved it 🥰

Fully understand transformers. One of the clearest and succinct explanations out there, so intuitive. Thank you!!

I already have some knowledge on the subject, however, I like to keep myself updated and there is always something new to learn. She clearly explains how what she is teaching really works. The whole video is worth watching.

I always meant to watch these lectures since 2020, but something always comes up. Now, nothing is going to stop me. Not even nothing. Great lectures, best way to learn.

Lovely presentation!
It couldn't get more interesting!

I am an auditor and have very little to do with this subject, except for my curiosity. I feel lucky that these kind of videos are available for free

Till Now best Course,
I am doing great when I found these MIT's Lecture

I come back every year to check these lectures and to see what innovations made it into the lectures. Pleasantly surprised to see the name change, congrats!

🎯Course outline for quick navigation:
[00:09-02:02]Sequence modeling with neural networks
-[00:09-00:37]Ava introduces second lecture on sequence modeling in neural networks.
-[00:55-01:46]The lecture aims to demystify sequential modeling by starting from foundational concepts and developing intuition through step-by-step explanations.
[02:02-13:24]Sequential data processing and modeling
-[02:02-02:46]Sequential data is all around us, from sound waves to text and language.
-[03:10-03:50]Sequential modeling can be applied to classification and regression problems, with feed-forward models operating in a fixed, static setting.
-[05:02-05:26]Lecture covers building neural networks for recurrent and transformer architectures.
-[11:56-12:37]Rnn captures cyclic temporal dependency in maintaining and updating state at each time step.
[13:24-20:04]Understanding rnn computation
-[14:40-15:04]Explains rnn's prediction for next word, updating state, and processing sequential information.
-[15:05-15:47]Rnn computes hidden state update and output prediction.
-[16:17-17:05]Rnn updates hidden state and generates output in single operation.
-[18:45-19:39]The total loss for a particular input to the rnn is computed by summing individual loss terms. the rnn implementation in tensorflow involves defining an rnn as a layer operation and class, initializing weight matrices and hidden state, and passing forward through the rnn network to process a given input x.
[20:05-29:13]Rnn in tensorflow
-[20:05-20:54]Tensorflow abstracts rnn network definition for efficiency. practice rnn implementation in today's lab.
-[21:16-21:43]Today's software lab focuses on many-to-many processing and sequential modeling.
-[22:53-23:21]Sequence implies order, impacting predictions. parameter sharing is crucial for effective information processing.
-[25:04-25:29]Language must be numerically represented for processing, requiring translation into a vector.
-[28:29-28:56]Predict next word with short, long, and even longer sequences while tracking dependencies across different lengths.
[29:14-41:53]Rnn training and issues
-[30:02-30:27]Training neural network models using backpropagation algorithm for sequential information.
-[30:45-31:43]Rnns use backpropagation through time to adjust network weights and minimize overall loss through individual time steps.
-[32:03-32:57]Repeated multiplications of big weight matrices can lead to exploding gradients, making it infeasible to train the network stably.
-[35:45-37:18]Three ways to mitigate vanishing gradient problem: change activation functions, initialize parameters, use a more robust version of recurrent neural unit.
-[36:13-37:01]Relu activation function helps mitigate vanishing gradient problem by maintaining derivatives greater than one, and weight initialization with identity matrices prevents rapid shrinkage of weight updates.
-[37:54-38:25]Lstms are effective at tracking long-term dependencies by controlling information flow through gates.
-[40:18-41:13]Build rnn to predict musical notes and generate new sequences, e.g. completing schubert's unfinished symphony.
[41:53-50:11]Challenges in rnn and self-attention
-[43:58-44:40]Rnns face challenges in slow processing and limited capacity for long memory data.
-[46:37-47:00]Concatenate all time steps into one vector input for the model
-[47:21-47:45]Feed-forward network lacks scalability, loses in-order information, and hinders long-term memory.
-[48:11-48:34]Self-attention is a powerful concept in deep learning and ai, foundational in transformer architecture.
-[48:58-49:25]Exploring the power of self-attention in neural networks, focusing on attending to important parts of an input example.
[50:13-56:20]Neural network attention mechanism
-[50:13-50:43]Understanding the concept of search and its role in extracting important information from a larger data set.
-[51:52-55:24]Neural networks use self-attention to extract relevant information, like in the example of identifying a relevant video on deep learning, by computing similarity scores between queries and keys.
-[53:32-53:54]A neural network encodes positional information to process time steps all at once in singular data.
-[55:32-55:57]Comparing vectors using dot product to measure similarity.
[56:20-01:02:47]Self-attention mechanism in nlp
-[56:20-57:14]Computing attention scores to define relationships in sequential data.
-[59:11-59:39]Self-attention heads extract high attention features, forming larger network architectures.
-[01:00:32-01:00:56]Self-attention is a key operation in powerful neural networks like gpt-3.
offered by Coursnap

Thank you for this amazing and easy to understand course!
I'm a beginner of the RNN, but I can almost know all the concepts from this lecture!

Wow just amazing, no words left. Really Thanks 🙏

I just started learning about RNN and LSTM especially for NLP and found this video very helpful to me. It would be really exciting if you provided a video about transformers in more depth :)

What an amazing content! Thank you! ❤️

Finally I understand the transformer concept now. Great lecture series👍!

query key value pairs always put me off whener I start to learn about transformers, this time I actually finished the video. Thanks MIT

Amazing . thank you MIT.

Thanks for this amazing course

The fact that these videos now have millions of views.... the world is evolving so fast scientifically or at least scientific culture.

Very intuitive explanation, thanks!

I am 6 years old, and I have been able to follow everything said, after watching 3 times.

I worked in spatial statistics during my graduation. And now, I think your classes will push me more and more towards the machine learning. Looking forward to apply my learning in my upcoming level of study. Thanks for your efforts 💝

Thank you Ava Soleimany and MIT ☺😊🤗💜

Code showed at RNN Intuition chapter at 14:00 makes thing clear af. I literally said "Wow"

Awesome Course, Very easy to understand+++, Thx all MIT instructors 😊😊😊

Thank you so much for the free course. Benifit and appreciate

Simply brilliant!

best Friday after-work fun thanks!

legendary lecture, thank you for sharing

00:16 Building neural networks for handling sequential data
03:19 Sequential data introduces new problem definitions for neural networks
10:03 Recurrent Neural Networks link computation and information via recurrent relation.
13:37 RNN processes temporal information and generates predictions.
20:22 Key criteria for designing effective RNNs
23:33 Recurrent neural networks design criteria and need for more powerful architectures.
30:08 Back propagation through time in RNN involves back propagating loss through individual time steps and handling sequential information.
33:23 Vanishing gradient problem in recurrent neural networks
40:03 RNNs used for music generation and sentiment classification
43:32 RNNs have encoding bottlenecks and processing limitations
49:45 Self-attention involves identifying important parts and extracting relevant information.
52:51 Transformers eliminate recurrence and capture positional order information through positional encoding and attention mechanism.
59:35 Self-attention heads extract salient features from data.
1:02:49 Starting work on the labs

Great Teamwork of Alex Amini and Ava Amini.

This is some really deep learning. MIT is the height of institutional education. 👏👏. Thanks for sharing.

3:00 Sequencial Data

Thank you for the awesome lecture

Thanks for sharing!

Mr Amini thanks for your channel

Whoever is listening to this awesome lecture I just want to say, Attention is all you need!!

Thank you very much for this great oppurtunity to watch MIT lectures. always dreamt of a world class education and finally im doing a degree in AI and such videos are supporting my learning process very much

This is shockingly good. Thank you.

Thanks for sharing such high quality content! 👌

Great material and the best educator!. Thank you for the fantastic video! The material was not only informative but also engaging, and the quality of the presentation was top-notch. Your depth of knowledge truly shines through, making the learning experience both enriching and enjoyable. Presented such complex material with such ease. You've done an exceptional job in communicating the concepts clearly. Great work!" and everything is free! Great job MIT team!!

Best explanation ever!!!! thank you

Thank you@MIT

She is fantastic at teaching. I love how easily understandable she makes it. Thank you, Prof Amini.

Really helpful! ⭐️

Salutes hopr to come back MIT Deep learning. I feel you peple need to look deep inro life

Pretty straight forward lecture.

ThNks mit

The way forward is dynamic quantim computing, possible throug blackhole nets

Thank you so much

Would like to see the coming lectures and the interesting student projects!

That's the reason why people wanted to go to the top universities such as MIT!! The explanation is so clear!!!

I've always wanted to study deep learning, but I never really knew where to start. This MIT course was my answer

It is striving to bring back our memory of interrelationship and oneness

This is the best lecture on KRplus! Thank you for the clear explanation. I wish you could delve deeper into the transformer architecture, though, as it was only covered in the last 15 minutes. Nevertheless, this is the most understandable video on the topic. I've watched nearly all of them, but this one stands out as the best! It would be great if you provided a more detailed explanation of transformers.

Awesome Course, Very easy to understand+++

This is amazing. Studying from Kenya, and this absolutely is quality lectures.

I can't wait to watch

She is so good!!!!🎉🎉❤❤

Great lecture

Great job simplifying very complex understanding the functions of neural networks!
Avi MD MBA, MS, MHA

Great 👍 presentation 👏

que increible! esto es genial!

This video is so good, i love it. But I was wondering, where can I find the coding part of RNN and transformers?