MIT 6.S191 (2020): Introduction to Deep Learning 

Skatījumi 988 tūkst.
99% 9 807 74

MIT Introduction to Deep Learning 6.S191: Lecture 1
Foundations of Deep Learning
Lecturer: Alexander Amini
January 2020
For all lectures, slides, and lab materials: introtodeeplearning.com
Lecture Outline
0:00 - Introduction
4:14 - Course information
8:10 - Why deep learning?
11:01 - The perceptron
13:07 - Activation functions
15:32 - Perceptron example
18:54 - From perceptrons to neural networks
25:23 - Applying neural networks
28:16 - Loss functions
31:14 - Training and gradient descent
35:13 - Backpropagation
39:25 - Setting the learning rate
43:43 - Batched gradient descent
46:46 - Regularization: dropout and early stopping
51:58 - Summary
Subscribe to @stay up to date with new deep learning lectures at MIT, or follow us on @MITDeepLearning on Twitter and Instagram to stay fully-connected!!

Zinātne un tehnoloģija

Publicēšanas datums


8 feb 2020






Mans atskaņošanas saraksts
Skatīties vēlāk
komentāri 100   
Wang Doje
Wang Doje Pirms Gada
Everyone watching this video lectures knows how the quality of the lecture is.. It's speechless and crystal clear I can say. Thanks MIT and prof Alexander Amini 👌
Shurricane Pirms Mēneša
I agree
swaraj dhami
swaraj dhami Pirms 2 Mēnešiem
@Aman Jain how is this lecture compared to Andrew NG deep learning.ai? Any review
Sebastian Jonson Ögren
Sebastian Jonson Ögren Pirms 3 Mēnešiem
Wow påwqqwe
Cutting_Edge_Tech Pirms 4 Mēnešiem
Exactly what I would like to say that the lecture is amazing and speechless. Thank you, Alexander and MIT.
Aman Jain
Aman Jain Pirms 5 Mēnešiem
@Alexander Amini sir your lecture is very fine can I get the lab work questions also it will be very good for understanding and learning
REPALA SUDHAMSH Pirms 3 dienām
Very much helpful
Sơn Đặng Thái
Sơn Đặng Thái Pirms 4 dienām
I would like to say thank to Mr Alexander Amini and his team for sharing comprehensive knowledge about Deep learning.
Fabio Souza
Fabio Souza Pirms 6 dienām
Wonderful explanation about the process of convolution, the steps of a CNN, and the most common applications! Thank you very much!
Ayoub Massoudi
Ayoub Massoudi Pirms 12 dienām
At minute 23, there is a little mistake I think, the final output uses g(zj) not zj!
Mark Reynolds
Mark Reynolds Pirms 15 dienām
The hilarious child rahilly cough because day biomechanically remove atop a righteous bagel. godly, brief geometry
Alexandre P. dos Santos
Alexandre P. dos Santos Pirms 15 dienām
Just a question, why not look for the global minimum, when looking for minimum loss, instead of taking a random position?
HAENARA SHIN Pirms 18 dienām
The best lecture. Period.
Alex Poma
Alex Poma Pirms 22 dienām
You are a great teacher, is easy understand you, all was clear and your examples was greats, thanks for put it free for alls
Kalai Anand Ratnam
Kalai Anand Ratnam Pirms 27 dienām
Thank You So Much For The Awesome Content. Top Class Explanation.
Kerstin Brachtendorf
Kerstin Brachtendorf Pirms Mēneša
It's great that you always provide us with the best update as regards the market space. For a very good reason, the market is still favourable and generally exhibiting bullish traits. although i have made so much gains ever since i got into *bitcoin trading* and so far the leading asset has been able to hold a very good support at 30,000 region against the US dollars with all indications showing that we are still very much in a bullish wave. At this point it is very smart of us to keep growing the little we hold by finding a working pattern and accurate daily signals, for me have made huge profits trading with Daron Wilber's daily signals which turns out to be very accurate and well predictable. He is so strategic in his methods and I am surprised at how accurate his signals are. Reach him on (Telegram *(@Daron_signals)* and WhatsApp *(+44-7862-071-076 )* ) provides some of the best insights I have been exposed to in crypto trading.
Cường Hoàng
Cường Hoàng Pirms Mēneša
The best cut architecturally rescue because train timely desert throughout a jumbled desert. unnatural, lewd engineer
yusra suliman
yusra suliman Pirms Mēneša
Very good refresh. Much appreciated, thank you
Ashik T M
Ashik T M Pirms Mēneša
Introduction was awesome
Prathamesh Joshi
Prathamesh Joshi Pirms Mēneša
Where can I find the PPTs ?
Ramkripal Yadav
Ramkripal Yadav Pirms Mēneša
What,s your recommended book sir
Ramkripal Yadav
Ramkripal Yadav Pirms Mēneša
For deep learning and algorithm
tamerius1 Pirms Mēneša
At 32:17 it says remember W contains the bias W^0 but it didn't because the size (1 to m) had to match that of the input! See W being defined at 12:43
Nasrul Islam
Nasrul Islam Pirms Mēneša
This is so interesting that I knew almost all of these theories, yet it helped to sharpen my understanding. Are the other lectures in the course available online?
gaoeln laojehc
gaoeln laojehc Pirms Mēneša
MK Pau-Student
MK Pau-Student Pirms Mēneša
WOW I am getting more and more interested in Deep learning, thanks for such a great course
where can i find the course as it is not there on website
aditya shirsath
aditya shirsath Pirms Mēneša
Best explanation I totally loved it
Shyne Sukumaran
Shyne Sukumaran Pirms Mēneša
What a lecture! Everything directly went to my brain permanently! Thank you so much for sharing your lecture!
Chalak Programmer
Chalak Programmer Pirms Mēneša
Awesome 4:00 lvlocal.info/must/HkWcby8wPfNPnFtL5cLPgQplaylists
Virat Bhatt
Virat Bhatt Pirms Mēneša
How if we let's imagine lets take 3 single neuron and take their output and the predicted model and then make a 1 single neuron with help of the previous outputs and model and predict the best of best possible answer. Is is possible can anyone help me with this?
haneen nabil
haneen nabil Pirms Mēneša
hi what previously knowledge i need to have?
Omar Morales
Omar Morales Pirms Mēneša
it's sad to see that videos like this with actual learning and potential to wonderful discoveries and applications, have 5X less views than stupid prank videos. Success is not for all, that's for sure.
Omar Morales
Omar Morales Pirms Mēneša
Thanks @Alexander Amini
Eva Peggy Best
Eva Peggy Best Pirms Mēneša
Thanks for sharing this incredible course online! Very clear to understand. Thanks to Alexander Amini and MIT!
Krishna Meghwal
Krishna Meghwal Pirms Mēneša
What are the prerequisites for taking this course?
Jim Greggas
Jim Greggas Pirms Mēneša
This is seriously excellent: Such clear & concise explanations to a complicated topic. The slides were also very useful. I honestly can't imagine it being better delivered. After continuously coming back and forth to this subject I'd say this is the best foundational material I have seen on the topic and ultimately a solid basis reference.
Mip Map
Mip Map Pirms Mēneša
WOW, thank you for bring this course to LVlocal, The most high quality course online.
madie Pirms Mēneša
is the course on the website is free or charged?
Tony Rcd
Tony Rcd Pirms Mēneša
Thank you so much for sharing amazing contents with us !
Sergio Arroyo
Sergio Arroyo Pirms Mēneša
Peter Müller
Peter Müller Pirms 2 Mēnešiem
is it possible to rent artificial intelligence? For example an algorithm? Maybe you found a company, instead of programming the AI by yourself, you rent it or Buy it?
Brent Vaalburg
Brent Vaalburg Pirms 2 Mēnešiem
Getting kids to learn, one prize at a time.
Thelma Romero
Thelma Romero Pirms 2 Mēnešiem
i was super interested until i saw obama 🤮😤
Prabhat Singh
Prabhat Singh Pirms 2 Mēnešiem
Best content and lecture......Thanks alot Sir
Md Golam Rabbani Fahad
Md Golam Rabbani Fahad Pirms 2 Mēnešiem
One of the best online resources for Machine Learning. Whole hearted thanks to MIT and the teaching team.
PAGE2SITE Pirms 2 Mēnešiem
Stephane Pirms 2 Mēnešiem
Rajoo Ananth
Rajoo Ananth Pirms 2 Mēnešiem
Thx Prof Amini
Isaac Nazar
Isaac Nazar Pirms 2 Mēnešiem
All MIT students in the field of medicine, biometric, security and other fields can contact me for funding small projects. Keep coding!
육단계 Level6
육단계 Level6 Pirms 2 Mēnešiem
Ayn Ml
Ayn Ml Pirms 2 Mēnešiem
man, at 25 yoa? unbelievable. Keep it up Sasha! and thank you for incredible lectures
editor ijsmi
editor ijsmi Pirms 2 Mēnešiem
Great Video. EditorIJSMI@ijsmi.com/book.php
Sayed Mahmood Sayedi
Sayed Mahmood Sayedi Pirms 2 Mēnešiem
One day I will be setting in an MIT class, INSHA'ALLAH
Azaz Butt
Azaz Butt Pirms 2 Mēnešiem
The lecture was so inspirational. I'm working on Machine and Deep Learning and these lectures really help me to my tasks effectively.
Proxima Centauri
Proxima Centauri Pirms 2 Mēnešiem
What are the Prerequisites required ? Please tell ASAP
Murray Pendergrass
Murray Pendergrass Pirms 2 Mēnešiem
I like that the video was filmed in 2020 but they still chose to emulate Obama instead of Trump. Really cool video!
Mohamed Badr Ismail
Mohamed Badr Ismail Pirms 2 Mēnešiem
i can't download the slides form your site , this massage 'coming soon' appears
Walid Ibrahim Ali
Walid Ibrahim Ali Pirms 2 Mēnešiem
It is an informative presentation . It is the MITian's code of professionalism . Thank you
Nomad Traveller
Nomad Traveller Pirms 2 Mēnešiem
I have gone thru' various video on Deep Learning and found it very difficult to understand. Most of teachers try to explain in mathematical language and make it more complex for a beginner to understand. You nailed it, Alexander. Terrific explanation. The pace and quality of your teaching is truly fantastic. Request you to provide a detailed course of Deep Learning. Thank you.
Norbert Fejér
Norbert Fejér Pirms 2 Mēnešiem
Excellent introduction to DNN! You should upload more videos on this topic. It helped me a lot to understand the big picture of the system. Thanks.
Delludens Pirms 2 Mēnešiem
What exactly the math do one needs to understand this lecture?
Ravindu Nawanjana
Ravindu Nawanjana Pirms 2 Mēnešiem
Can I get slides and lab materials i visit the site but those are not available excellent course thanks alot.
spacepanda Pirms 3 Mēnešiem
the numbers aren't discrete or continuous, the variables are. The fact that the variable is in discrete or continuous form comes from the way it's values (represented for example by real numbers) are distributed, and yes, this distribution can either be a discrete choice {0, 1} or a continuous range (0:1). Might not sound important, but it ruins the way of how the regressor or classifier output should be understood. To put it short, the regression algorithm tries to approximate the real continuous distribution, the classifier - the real discrete one, but when you sample your output from it, well, the number is a number.
Ajay Kumar Reddy
Ajay Kumar Reddy Pirms 3 Mēnešiem
Mom, I am attending MIT!! Great Lecture Series Alex.Thank You!
이이라 Pirms 3 Mēnešiem
6:20 Novelty 에 high value adding 해주는 건 지구상에서 MIT 의 undergraduate program course 딱 하나네..... 교수 채용할 때도 SCI 논문의 NOVELTY가 중요한 게 아니라는 게 상위권 대학 교원 채용...
David SL
David SL Pirms 3 Mēnešiem
Can't believe this guy is only 25..
Halil Durmaz
Halil Durmaz Pirms 3 Mēnešiem
Incredibly good presentation skills, structure ! The best intro material...30:12 I guess f(x) should be replaced with f^(x) since its related with predicted value
P Didy
P Didy Pirms 3 Mēnešiem
whats this the fucking commercial channel after yak yak yak meaningless dribble. wow. Next time, try to say something less limited in thought.
Nguyễn Bảo Dung
Nguyễn Bảo Dung Pirms 3 Mēnešiem
Golden, diamond, all the best thing is converted into knownledge. It's really my privilege here to be able to watch a recent MIT course, eventually in 2020. I'm following ML field, and you just make me feel so happy and motivated to move forwards. All these knownledge are truly golden. Thank you so much.
nafil Al-Alawi
nafil Al-Alawi Pirms 3 Mēnešiem
Amazing Leture
Sarmistha Das
Sarmistha Das Pirms 3 Mēnešiem
Indeed One of the finest deep learning lecture I've ever heard. Thank you so much Professor Alexander Amni .
Ma. Emilia Puente
Ma. Emilia Puente Pirms 3 Mēnešiem
Thanks so much. People from around the world are watching and learning.
Asad Anees
Asad Anees Pirms 3 Mēnešiem
Thanks Alexander for great community services. Best deep learning lecturers
NALIN PRABHATH Pirms 3 Mēnešiem
any prerequisites required to watch this playlist?
ayoola fakoya
ayoola fakoya Pirms 3 Mēnešiem
This just sealed the concept of DL for me. Amazing lecture.
William Wang
William Wang Pirms 3 Mēnešiem
Do you teach this to the ECE students? I feel right at home with this.
DigitalSCRM Pirms 3 Mēnešiem
Wait, so data science is the same as deep learning? I'm lost
Xin Niu
Xin Niu Pirms 3 Mēnešiem
Hi, Thanks so much for making this course. At 15: 28, why the activation function makes the decision boundary non-linear? In logistic regression, we also have a sigmoid activation function, but the decision boundary is linear? Is the non-linearity caused by the multi-layer structure?
Aditya Pirms 3 Mēnešiem
If I get into MIT this year (I really hope I do but the competition is crazy; and understandably so for such an amazing institution) the first thing I’m doing is signing up for your class
Abdo Y
Abdo Y Pirms 3 Mēnešiem
Hey everyone, who need this written video. I can write it for you in any language you want, very reasonably priced, pay after receipt, even if you need to write another LVlocal video, welcome
Abdo Y
Abdo Y Pirms 3 Mēnešiem
This is my address (youness291988@hmail.com)
Madhav Pirms 3 Mēnešiem
Dear Prof. Amini, deeply impressed by the quality of content and delivery of your lecture. Feel privileged to listen to such a dedicated teacher and scientist.
Aditya Since 2006
Aditya Since 2006 Pirms 3 Mēnešiem
this lecture is very useful and informative cause it is DEEP............
Aashish Rimal
Aashish Rimal Pirms 3 Mēnešiem
love from nepal
Louis Szeto
Louis Szeto Pirms 4 Mēnešiem
Prof Amini, I want to on 44:51, ask why mini-batch gradient descent is using their average loss, but not the minimum among the local minima collection from mini-batch
Bijan Ghofranian
Bijan Ghofranian Pirms 4 Mēnešiem
Dear Alexander Amini Aziz, I thought I was a good teacher but damn you are way better than me by several order of magnitudes. You have a superb skill to break down a complex subject and gradually delve into finer details that is easy to follow and comprehend. Thank you so much for sharing this great presentation, I am so motivated that I will keep watching the full series.
分数指数幂 Pirms 4 Mēnešiem
2020年的课程 很新啊
Sravani Katasani
Sravani Katasani Pirms 4 Mēnešiem
Can anyone please tell the prerequisites for this course.
Beltus Nkwawir
Beltus Nkwawir Pirms 4 Mēnešiem
I had watched this video some months ago and I thought it was complex, Watching it today, it's like being fed with ice cream while you lie on your spine. High quality. Easy to understand. Impressive. Thanks Alexander Amini.
Sanem Deekshith Goud
Sanem Deekshith Goud Pirms 4 Mēnešiem
and here my univ teaching c in turbo c compiler , life sucks
FOSALGO Pirms 4 Mēnešiem
Nice Video
x dg
x dg Pirms 4 Mēnešiem
this is the clearest introduction to deep learning I have ever watched.
Ratandeep Gupta
Ratandeep Gupta Pirms 4 Mēnešiem
Such a wonderful lecture. It was the most valuable lecture i ever got on this topic. It gave clearest ideas. Thank you.
Steven On The Road
Steven On The Road Pirms 4 Mēnešiem
Fascinating, brilliant! Though deep learning is far away from my knowledge or academic background, your lecture is cristal clear and opens a door for me. Thank you!
Shobikhul Irfan
Shobikhul Irfan Pirms 4 Mēnešiem
Semoga ada bahasa Indonesia nya di subtitle
Ananya Mittal
Ananya Mittal Pirms 4 Mēnešiem
Can anyone help in telling from where can I learn TensorFlow from scratch?
Vu pr0
Vu pr0 Pirms 4 Mēnešiem
i have not learned AI, so can i learn this course now?
Muhammad Jahanzaib
Muhammad Jahanzaib Pirms 4 Mēnešiem
Can I win these GPUs😁
Mohammad Uzair
Mohammad Uzair Pirms 4 Mēnešiem
Tanmay Mitra
Tanmay Mitra Pirms 4 Mēnešiem
undoubtedly one of the best and most effective presentation for an introduction to a complex subject like Deep-Learning, available on youtube for free viewing and learning. Kudos and best wishes Alexander.
Matt Coakes
Matt Coakes Pirms 5 Mēnešiem
Man, these concepts are said in plain english, but so effectively. Even in a few sentences I feel like I learn so much. Fantastic quality
Gbadamosi Idris
Gbadamosi Idris Pirms 5 Mēnešiem
Wow really cool...
kxmode Pirms 5 Mēnešiem
Could the dense arguments be comma separated to save space? For example, tf.keras.layers.Dense(n1, n2, n3)
ahmhmd1990 Pirms 5 Mēnešiem
I believe this lecture summarized a whole course of artificial intelligence for me hahaha. Right to the point and clearly explained. I am familiar with the mathematical concepts, and this lecture was like the thing I needed to connect the dots for me.
Manan Lad
Manan Lad Pirms 5 Mēnešiem
Why am I getting privacy error when I try to visit link in the description.
Sushil Dubey
Sushil Dubey Pirms 5 Mēnešiem
Good introduction to Deep learning in short video.
Siba Prasad
Siba Prasad Pirms 5 Mēnešiem
Thank u teacher Alexander Amini, this will be extremely helpful who are new to this field! Thank you sir, i got every concept what you have taught.
AlphaGo - The Movie | Full Documentary
1. Algorithmic Thinking, Peak Finding
6. Monte Carlo Simulation
Skatījumi 1 milj.