We've created a summary of key topics covered in this course to help you decide if it's the right one for you. Click individual badges to see more courses on the same topic.
Harvard University, the Massachusetts Institute of Technology, and the
University of California, Berkeley, are just some of the schools that you have
at your fingertips with edX. Through massive open online courses (MOOCs) from
the world's best universities, you can develop your knowledge in literature,
math, history, food and nutrition, and more. These online classes are taught
by highly-regarded experts in the field. If you take a class on computer
science through Harvard, you may be tau...
Harvard University, the Massachusetts Institute of Technology, and the
University of California, Berkeley, are just some of the schools that you have
at your fingertips with edX. Through massive open online courses (MOOCs) from
the world's best universities, you can develop your knowledge in literature,
math, history, food and nutrition, and more. These online classes are taught
by highly-regarded experts in the field. If you take a class on computer
science through Harvard, you may be taught by David J. Malan, a senior
lecturer on computer science at Harvard University for the School of
Engineering and Applied Sciences. But there's not just one professor - you
have access to the entire teaching staff, allowing you to receive feedback on
assignments straight from the experts. Pursue a Verified Certificate to
document your achievements and use your coursework for job and school
applications, promotions, and more. EdX also works with top universities to
conduct research, allowing them to learn more about learning. Using their
findings, edX is able to provide students with the best and most effective
courses, constantly enhancing the student experience.
This introductory computer science course in machine learning will cover basic theory, algorithms, and applications. Machine learning is a key technology in Big Data, and in many financial, medical, commercial, and scientific applications. It enables computational systems to automatically learn how to perform a desired task based on information extracted from the data. Machine learning has become one of the hottest fields of study today and the demand for jobs is only expected to increase. Gaining skills in this field will get you one step closer to becoming a data scientist or quantitative analyst.
This course balances theory and practice, and covers the mathematical as well as the heuristic aspects. The lectures follow each other in a story-like fashion:
What is learning?
Can a machine learn?
How to do it?
How to do it well?
Take-home lessons.
Instructors
Instructors:
Yaser S. Abu-Mostafa
University
University:
Caltech
Instructors
Instructors:
Yaser S. Abu-Mostafa
University
University:
Caltech
Reviews9/10 stars
24 Reviews for Learning From Data (Introductory Machine Learning)
Rankings are based on a provider's overall CourseTalk score, which takes into account both average rating and number of ratings. Stars round to the nearest half.
The lectures are great, and the class covers a number of essential concepts in
ML in much gory detail. Unfortunately, everything else in this class was
rather disappointing. The translation to edX platform was an afterthought, and
the homeworks are a mess: there's no opportunity to practice unless you come
up with practice problems of your own, problem statements can be a bit on the
undecipherable side, and with just one attempt there's no chance to recover
from your mistakes. Lack of immediate feedback doesn't help there either. If
you want to audit, this will be a great experience. Otherwise, prepare for
some pain.
This is really an excellent course. It gives a real understanding of the basic concepts and methods in the world of machine learning. But this understanding is achieving through hard work, challenging tasks are available. And complexity is not an end in itself, tasks are chosen so that the solution leads to an improvement in the conceptual understanding of things. The lion's share of tasks requires setting up a computational experiment, so without good programming skills this course can become an excessive load.
The lecturer talks about the material not dispassionately, but as something very pleasant and interesting for himself. This greatly enhances the effect of perfectly prepared lectures.
This is a great course. It's carefully structured and taught. Professor Yaser is a fantastic instructor. That being said, it's not an easy course. The homeworks require a good amount of time and there's no "retake" on wrong answers. Nonetheless, it does make you feel smarter after finishing the job. Highly recommended. Thanks again professor, Edx and Caltech for such an top level opportunity.
These lectures are the best. Prof. Mostafa explains involved math very easily. If you want to build an intuition and at the same time understand the math behind completely, then this is a course for you
This was the best course I've taken on edX so far. I very much like the fact that this is a full fledged university lecture and not just a summary of one.
I really appreciate the course creators generosity to share their expert knowledge with everybody for free of charge. In my humble opinion it is those people who drive the progress of the human race.
I've done a few courses which including machine learning topics before, but never achieved this level of understanding as in this course. Mostly, they just introduced some ideas of machine learning with practical applications, which was interesting but could not grow in my mind due to the lack of a solid foundation. A few weeks after the courses ended, the learned things where already gone again.
In this course – if you are willing to invest the time (and you will need to) – you will really dive deep into the fundamental concepts of machine learning and hence truly build step by st...
This was the best course I've taken on edX so far. I very much like the fact that this is a full fledged university lecture and not just a summary of one.
I really appreciate the course creators generosity to share their expert knowledge with everybody for free of charge. In my humble opinion it is those people who drive the progress of the human race.
I've done a few courses which including machine learning topics before, but never achieved this level of understanding as in this course. Mostly, they just introduced some ideas of machine learning with practical applications, which was interesting but could not grow in my mind due to the lack of a solid foundation. A few weeks after the courses ended, the learned things where already gone again.
In this course – if you are willing to invest the time (and you will need to) – you will really dive deep into the fundamental concepts of machine learning and hence truly build step by step a very strong insight in the realms of machine learning and computer science. And since these brilliantly thought Ideas are “as simple as possible, but not simpler” I really believe that they will last and enable us to pursue further knowledge.
The course is not easy. But it is manageable even for non computer scientist, as I am – coming from physics - but a good mathematical understanding is extremely helpful as I can confirm. Besides that, the only thing you need to have in order to profit most from this course, is passion. Everything else is managed really good by the course team.
I can only recommend this course to anyone who is willing to learn about learning a machine to learn.
Thank you again!
Be ready! This is a real course. Not watered down. Very challenging. You will learn a lot! Note the effort required in the overview - I probably averaged 15 hours a week on just the homework. Note the prerequisites. A healthy combination of theory, analysis, and practical application. The professor and the TAs are very active on the discussion boards, as are the students, so you get *great* interaction with a great community.
This course is a great introduction to the world of machine learning and the instructor is really good in his lectures.
The course covers some important aspects in ML and is a good start if we want to continue in this domain. Unfortunately, because it is an introduction, it is not based on real and concrete problem of ML and the homeworks are all theoretical.
About the level needed to pass this course, even if it's written "basics in calculus, matrix and probability", I think it is a plus to be used to advanced mathematical notions.
Great theoretical course. Many brilliant explanation for the hard material.
Pay attention this is excellent theoretical course with a taste of practices
Great course, by all means! Great teaching style and challenging homeworks. The course elaborates a lot on the theoretical framework, which is insightful for the practical applications.
I have taken this course before it was available on edX via the caltech website but, when I saw that this course is now available here also, I decided to write a brief review anyway because I found it a great experience and hope many others will too.
The lectures for this course were quite challenging but also very entertaining. It is very well structured and the instructor, prof. Yaser S. Abu-Mostafa, illustrates nearly all concepts using simple examples often with some visual aids (graphs, figures, sketches) and speaks slowly and clearly making the content much easier to digest. Having participated in this course I truly believe that more than deserves all the teaching awards he has received over the years. The lectures are rather brief so after watching one for the first time you may think that it could have been more in depth but, I can assure you, the lectures contain everything you need, nothing more and nothing less - this is...
I have taken this course before it was available on edX via the caltech website but, when I saw that this course is now available here also, I decided to write a brief review anyway because I found it a great experience and hope many others will too.
The lectures for this course were quite challenging but also very entertaining. It is very well structured and the instructor, prof. Yaser S. Abu-Mostafa, illustrates nearly all concepts using simple examples often with some visual aids (graphs, figures, sketches) and speaks slowly and clearly making the content much easier to digest. Having participated in this course I truly believe that more than deserves all the teaching awards he has received over the years. The lectures are rather brief so after watching one for the first time you may think that it could have been more in depth but, I can assure you, the lectures contain everything you need, nothing more and nothing less - this is something you will greatly appreciate when preparing for the final.
If you do not intend to use the book, you will probably end-up re-watching lectures or at least portions of them before and while doing the homework (that is how I did it). I remember the homework being quite difficult and the answers not always very precise (it's machine learning, in some cases results will vary slightly from one attempt to the other). Definitely set aside plenty of time during the weekends, the estimated 10-20h a week is not exaggerated! I would watch the lectures on Saturday mornings and give the homework for that week a shot in the afternoon, on Sunday morning I would wrap up what was not finished and give exercises where I got stuck a fresh look. You may get frustrated at times when you get a wrong answer but I think the lack of 2nd chances makes all the correct answers so much more rewarding. The professor supplies plenty of inspiration and motivation during the lectures to make you want to really understand the material and work on the homework till you get it right!
I would recommend having some prior exposure to programming using Matlab or similar because you will be on your own on that front (as the professor says, the course content is not dumbed down for popular consumption).
Wondeful course content for all of those interested in Machine Learning that are searching some understanding of learning , rather than exercising number of methods without deeper context. Some minor issue is audio transcription which sometimes is not correct.
Absolutely a must for anyone thinking of learning Machine Learning. It
requires basic calculus and theory of probability knowledge, programming
experience (any language, but Matlab or R or Python are better choices), and a
lot of attention and time spent on its materials and homeworks. The book by
professor Yaser Abu-Mostafa et al. is very helpful resource (Learning From
Data, A Short Course). Expect spending at least 5-7 hours per week or more
depending on your background. This is very different from other Machine
Learning courses that immediately focus on practical aspects of the subject.
Learning From Data starts with founding concepts of Machine Learning as
mathematical and statistical problem, and gradually introduces to algorithms
without loosing its focus on the principles built. I consider both its
approach and materials the most sound introductory course in Machine Learning
available.
I took first instance of this course (started at end of 2013) on edX. Course
was fantastic and require lots of effort of mine to have it completed. It is
basically as prof. Mustafa said in introductionary speach : focus is on
understanding. This is very important and course deliver its promise
successfully. Nowadays there is countless of techniques in machine learning
you can learn them , but how to apply them for learning, considering data we
are given is not an easy thing, also what does it mean to learn, all of this
concerns are covered in detail in this course. I'm software developer and
completing this course enabled me to work on machine learning projects which
otherwise would not be hardly possible. So if you are to learn/understand
machine learning this course is the best way to spend your time in area of
machine learning. Highly recommended!
This is the best MOOC I have ever experienced. The instructor is very clear
about his goals and every lecture feels like a nicely wrapped present. The
care taken in making these lectures is palpable and I am truly grateful for
the team that made this possible. The new gold standard for anyone creating a
MOOC.
I've read quite a bit about machine learning, but this course was a very
interesting addition. It explains why machine learning methods work. A great
insight if you want to be more than an "ML engineer". I even couldn't wait and
watched all lectures from Youtube instead.
My opinion is that this is an incredible course. Dr. Mostafa asks and answers
the questions: "What is learning?" and "How do we know that we have learned?"
He also shows you the limitations of machine learning, and his exercise sets
are excellent. He demonstrates the theoretical foundations of machine
learning, each lecture builds upon previous lectures, and it is clear that one
is guided by someone who has figured out a very direct route to basic
understanding. I highly recommend this course!
One of, if not the, best course I have taken ever, in any format. My
background is biomedical with 'hobbyist' programming experience, with no
formal machine learning training. I found the lectures engaging, entertaining,
clear, and VERY educational, providing what I felt was a solid foundation for
further learning. The book (by the same name) was helpful to reinforce the
material.
Overall I say it is an excellent course. Prof. Yaser was very articulate in
explaining the concepts. The assignments are really challenging and and it may
be time-taking, but it is worth solving (Trust me!). I am really happy that i
did this course and more happy that i could pass the course! :) It sure
enhanced my knowledge in the area of machine learning. In short, I recommend
this course.
This class is HARD. But I like the challenge. Usually, courses from elite
universities are watered down when they become a MOOC. Perhaps there is a more
lenient grading curve for MOOC participants, but as near as I can tell, this
is the same material the CalTech students get. (Reference point: at CalTech,
getting 770 out of 800 on the math part of the SAT means 75% of your
classmates scored higher than that). I am currently taking the Stanford
Machine Learning class (which others have mentioned is watered down from what
Stanford students get) and I have taken the Berkely AI class and this CalTech
class is definitely harder than those two.
The lectures did not prepare me for the programming on the first homework
assignment. I'd just have to learn how to do it on my own, defeating the
purpose of taking the course. The lectures are just long uninterrupted blocks,
and they could desperately use interactive questions to incrementally develop
your understanding. I'm going to hope the Coursera class on Machine Learning
does just that.
Excellent course on the fundamentals of machine learning. It doesn't aim to
give a survey of all the various approaches, but rather gives a solid
understanding of the basics which you can use to analyze any of the machine
learning approaches on your own. The professor has a very clear and engaging
style.
Rankings are based on a provider's overall CourseTalk score, which takes into account both average rating and number of ratings. Stars round to the nearest half.