Facebook Wants to Teach Machine Learning

When you think of technical education about machine learning, Facebook might not be the company that pops into your head. However, the company uses machine learning, and they’ve rolled out a six-part video series that they say “shares best real-world practices and provides practical tips about how to apply machine-learning capabilities to real-world problems.”

The videos correspond to what they say are the six aspects of machine learning development:

  1. Problem definition
  2. Data
  3. Evaluation
  4. Features
  5. Model
  6. Experimentation

None of the videos are longer than 10 minutes, so you’ll invest less than an hour. The videos focus less on a specific product and more on the architecture and implementation strategies. That’s valuable, but this probably isn’t your only machine learning tutorial.

Quite a bit of these videos cover things we think are pretty obvious engineering axioms applied to machine learning. For example, a recurring theme is that you need to have a way to evaluate the system and do testing to see that things you change are actually making things better. Still, there are some things that are specific to machine learning.

Facebook has been in the news a lot lately, mostly not in a good way.  However, their research arm quietly turns out things ranging from Torch — a scientific computing framework with machine learning, to speech recognition and synthesis.

It seems like a lot of companies want to teach you about machine learning, including Google. You can even run TensorFlow in your browser.

21 thoughts on “Facebook Wants to Teach Machine Learning

  1. “And in a related story, Facebook reported today that their AI infused bots taught themselves how to shut down Facebook’s user data sharing algorithms to the external works … MZ said ‘must have been a programming error’”.

  2. “It seems like a lot of companies want to teach you about machine learning, including Google. You can even run TensorFlow in your browser.”

    Naturally. Takes awhile to grow a new PhD.

  3. No amount of testing and “validate” the “correctness” of self-teaching AI. It’s just not possible. They can test until the end of the universe, and it still won’t prove it has no bugs. OK,so that does sound like a lot of the code running the planet right now, but it’s for different reasons.

  4. A lot of the fear of ‘AI’ comes from the fact people don’t understand how it comes to decisions/findings. This at least may help those who want to learn, how it does so.

    It’ll also show how hugely fragile the whole learning system can be, and how easily it can become biased in any particular direction, and maybe people will learn that sometimes the bias the machine appears to take, is because there is a natural bias that way. Once we start to believe that, maybe we’ll finally start fixing some of the worlds problems.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.