Deep Learning

Understanding and using deep learning networks

This is machine translation

Translated by Microsoft
Mouseover text to see original. Click the button below to return to the Original version of the page.

Best of 2018 – Deep Learning Edition 2

Posted by Johanna Pingel,

Today, let’s journey back to the "Best of 2018". I’ll highlight specific posts and features from 2018 that are worth a second look. 

Deep learning for ...

We’re excited to report we’ve been spending lots of time here on new functions, examples, and videos for different deep learning application areas. We still tend to focus a lot on deep learning for images since:
  1. The research is still leaning in that direction and
  2. It’s really interesting
But now, we are diverting more attention to other areas as well, first and foremost: signal and time series data. Along these lines, there are a variety of examples to check out:

Deep learning examples in the doc

  • An interesting article in the 2018 “News and Notes” publication contained an article highlighting deep learning for signal, image and text.

 

A sample page of "News and Notes" featuring deep learning

There are even more resources on "deep learning for..." in the following sections.
Videos

It was a great year for the creation of webinars and videos:

  • Gabriel and I did a 30 minute video on Image Processing and Computer Vision, which of course has deep learning in it! This is a great intro for anyone just getting started.
  • I created a “What is Computer Vision” video, that just went live recently, (special thanks to our video team!) This is for the beginner crowd interested in computer vision, so this may not be as relevant to this group!
  • Along the same line, there's a new What is Machine Learning video, created by Loren Shure. This is a quick and thorough look at Machine Learning that's great for the beginner crowd, or anyone that hears terms like 'Machine Learning' and 'Deep Learning' and wants to learn more.
  • A new webinar is available for deep learning for signal data called: "Deep Learning for Signals and Sound"
  • A new webinar is available for Text Analytics.
Here's a recap of these videos with links in the images:

Image Processing and Computer Vision Computer Vision Video Signals and Sound Webinar Text Analytics Webinar

"New Stuff"

We have also been hard at work creating new content. I’m just going to bucket this as “new stuff.”

It's nice to have the relevant deep learning functions printed and posted to your office wall.

Deep Learning is not the only cheat sheet available. A successful Machine Learning Cheat Sheet is also available for you to use.

Here are 5 cheat sheets you may find useful: (The images are links to the PDFs)

Deep Learning Machine Learning Time Series Preprocessing ↓ ↓ Live Editor Import / Export Data
↑ I strongly recommend this one!
    • Solutions page updates: We’re continuously updating our content on the web: We have a new look to our IPCV solutions page and our Deep learning solutions page. There are links to new content on both of these pages.
 
    • One final thing to mention is a new approach to examples, making them more interactive. We released a Deep Learning Browser Example, which allows you to try out a simple example to experience what deep learning with MATLAB would look like. This is great for beginners and people who don't have access to MATLAB yet (no license required to try this example). Tell your friends!
 

New Features!

Every year, we have 2 releases of our product. This year brought us R2018a and R2018b. (We’re pretty predictable on the names).

  • In R2018b, we decided to try something different and used a video approach instead. Gabriel did an excellent job highlighting the important features of 18b, along with some of the key features from recent releases. You can find the video on the solutions page, or this link directly to the video. Please be aware this video plays automatically, which made me jump out of my seat with my speaker up too loudly!

Final Thoughts

As we head into the new year, I like to take stock of what went well during the past year and what we can do better. My goal for the new year is to increase the frequency of posts and start posting shorter content as it becomes available. I’m looking for suggestions on what you’d like to see more and/or less of.

Also, do you agree with my top 4 list? Anything else I should have mentioned? Leave a comment below!

Happy Holidays and see you in 2019!!

MathWorks is ready for the holidays!

2 CommentsOldest to Newest

yaogang replied on : 1 of 2
Reinforcement Learning in MATLAB? Like the python library rl_coach providing Reinforcement Learning algorithms and interface to various environments. link: https://github.com/NervanaSystems/coach
Anoush Najarian replied on : 2 of 2
Congrats on this, thank you for your work on the blog in 2018, Johanna, can't wait for 2019 posts! Love all the awesome Deep Learning examples! It's going to be a fun holiday break!! https://www.mathworks.com/help/deeplearning/examples.html