>coursera.org/learn/machine-learning >Free tutorial course provided by Dr. Andrew Ng, Co-founder of Coursera. >If you end this course, You can now code OCR by yourself.
Past attempts at establishing a machine learning general have failed. Sup Forums just doesn't have the background and interest.
Oliver Lee
I'll just make this into README.md and publish on GitHub.
Hudson Bell
How strong of math do you need to really dig into this?
Angel Bell
It's too bad, but Sup Forums has changed dramatically over the past 7 years. It's mostly just fags/kids with their consumer grade electronics.
Juan Barnes
Get a degree in Mathematics. OCW is ok
Leo Cox
I noticed this too, even though I hang around /diy/ more for actually technology related discussions. That part of Sup Forums actually talks about real technology and practical projects associated with them. Usually outsiders would expect Sup Forums place talking about design and development of electronic devices and applying science and tech in practical scenario but instead we get this endless scene of corporate worship of consumer tech and bait threads associated. I'm quite surprised there isn't a dedicated /ceg/ (Consumer Electronics General) sub board to help this board stay relevant to it's actual intent of discussing technology and design/development associated with it. I don't get it when /vg/ is created to mitigate the flooding of Videogames "insert game name here" General thread that Sup Forums once had, but Sup Forums doesn't get the same treatment despite all the consumer tech/goods generals thread this board gets daily.
Grasp how day-to-day activities are powered by machine learning Learn to 'speak' certain languages, such as Python and R, to teach machines to perform pattern-oriented tasks and data analysis Learn to code in R using R Studio Find out how to code in Python using Anaconda
These are useless to me though, I know Python and care little for R ( I know C). I also already know wtf machine learning is...
Juan Long
I got one to thumb through while waiting for jenkins, it was pretty decent.
Called Machine Learning for dummies, just saw it at B&N while wandering on the weekend some time back.
Uses primarily python, worked at a reasonable pace, and did an ok job explaining. I have a problem of buying tech books and reading them, sometimes work will pay for them if I add them to the library when I'm done though.
Jacob Smith
>tfw in 1st year comp sci
i hope im not too late to ride the machine wave
Samuel Flores
Anyone else using ML methods to do quant things?
John Evans
The suggestion stands, if you knew machine learning you would have passed the course but struggled. 99% of struggling in learning is because your foundation is missing something, perhaps a sentence slipped you or it was phrased in a way you couldn't make sense of but passed over it.
As as result when you get further on you are missing something and stuck. Coursera are pretty idiotproof and if you tried to dig around a bit you wouldn't be here asking for suggestions.
That said, the book is cheap, solid, and will bring you up to speed. Those points you mentioned from what I believe is the cover, are literally just advertising minor points, if that was the book it wouldn't be about machine learning, now would it?
Adrian Jenkins
How do you organize your deep learning papers? My backlog is ever growing.
Samuel Reed
Dang, way to make a person interested feel stupid - guess this shit ain't for me if there are people who don't struggle with the course.
I struggled at the coding portions - the quizzes were easy enough. But the assignments in Octave was loosely based on the concepts from the course and more to do with vector manipulations
The points I mentioned are the summary from the description... ie: the thing that is suppose to convince people to buy it. If you say it's more than that then sure... I'll give it a gander.
Benjamin Murphy
I didn't say anything about your intelligence, just persistence or dedication.
Also if you're struggling with the coding sections of that course then I suggest you get a bit more practice, and I say this not as a dick insult but as someone who wants to help you. I wouldn't respond if I didn't want to.
The course wasn't high level, it only expects as much as a college grad could do. Entry level pretty much.
Ian Turner
Read the Ian Goodfellow book if you already have a solid undergraduate level CS and math background.
Justin Hughes
Can you share it? Torrent or 4lpha.one or google drive, anyways.
Samuel Cruz
gotcha, can I just ask one thing then - is vector manipulations really that big of a thing for machine learning?
I stopped trying simply due to that alone. I never used linear algebra in industry after graduation (about 4 years since) so I'll probably have to retake a course on it - granted my dislike of matrix manipulation is probably strong enough to make me not learn how to do Machine Learning
Samuel Ramirez
Can you zip them up and put them on mega please.
Brandon Miller
I only save the ones relevant to my research/project so it could be biased. It's mostly arXiv papers posted on /r/ml so I don't think it would be of more use than things like arxiv-sanity.com/ where you can tailor to your need
Chase Reyes
You may want to brush up on your foundation, then try again. What you're describing are pretty basic in any type of programming outside of code monkey jobs.
Lucas Richardson
What side projects are you guys doing in addition to your research? I'm annotating data to train an object detection model that identifies intimate interracial relationships in photos on social media.
Chase Powell
>What you're describing are pretty basic in any type of programming outside of code monkey jobs
Where exactly does Linear Algebra and Calculus come into play in programming except for data science??
Christian Wright
Computer graphics
Nathaniel Bennett
Which is a meme field now as half of the papers at SIGGRAPH this year are deep learning applications.
Caleb Hill
Why?
Levi Kelly
I work in the financial sector making and maintaining small applications which act as a middleware between what others do in desktop interfaces and a larger processing server and database system.
Perhaps I've been doing this so long I have forgotten what is easy. Just keep on pushing and you'll get somewhere. If you want it bad enough you'll get it. If you do not that you will find some other interesting fancy or programmatic status symbol.
Matthew Martin
bump for interest
Ian Watson
Why can't machine learning into chess?
Top chess programs these days (The top three change around, but for the most part, Stockfish, Houdini and Komodo shit on everything but each other) still use more classical AI algorithms that require lots of domain knowledge. Will neural networks/SVM's/etc. ever be as good at evaluating chess positions?
I understand that alpha-beta and other pruning techniques are better than monte carlo for the search because the search space is somewhat manageable (compared to something like Go), but for some reason the evaluation part feels like there shouldn't be a reason not to be as good. I mean, people can easily get a database of high quality games online to train on, etc.
Jordan Hill
What's the best way to study without forgetting things you learned?
Go to /sci/ if you want to talk about ML and related.
Liam Carter
First of, modern chess AI engines are good enough to win a game against any player in the world, for a practical standpoint chess is already solved.
Also the good thing about chess is that we have 100% of the information so we can check the game by brute force.
Of course what we want is to find a winning strategy, that is a strategy that always wins if playing whites. The state space of chess is insanely big. The problem is not the theory. We know that there exists a deep neural net model that converges to a winning strategy for chess, we just can't find it with current computers.