cdgore

I've heard people say this since the program is so large, but the same could be said for any large research university that has a CS master's program.

Anecdotally, I got my master's degree in CS at USC and had a good experience overall. I definitely learned a lot and was able to fill in some gaps in my knowledge of CS theory and machine learning. I was also able to get offers from Google and Facebook from their on-campus recruiting programs, and I know several other people in my cohort that did the same.

I'm seeing the same thing in NYC.

Merkle trees are used extensively in distributed computing. Here's an example: You have many instances serving assets of some kind replicated in several data centers. They are structured in an arbitrary number of branches with many assets in each branch:

  (root)
    |   \
  (foo)  (bar)---(baz)----[even_more_assets]
    |     |
    |  [many_more_assets]
    |
  [many_assets]

How do you make sure the asset trees in the instances are synced up, serving the same assets?

You could compare every bit across every instance, but that would be very inefficient. You could compare the hashes of all the assets at each node of the tree, but you would still end up making many unnecessary comparisons.

You can simply hash each node of the tree based on the hash of its child nodes / assets. This forms a Merkle tree. That way, you can simply compare the root hashes across instances, and if they match, then the whole asset tree must be the same. If they don't match, compare the hashes at the next level of the tree until you find the discrepancy, and sync the nodes / assets where the hashes don't match. Merkle trees can be a very powerful tool for reaching consensus quickly in distributed systems.

Google Cloud had a speech api and supports 80 languages. There is a demo: https://cloud.google.com/speech/

>> "Google's building models with billions of parameters that require much more than 200MB, and that are really, really good at scoring data. I have to believe either that a) Apple is not telling us everything, or b) they haven't figured out a way to bring their customers the most powerful AI yet. (And the answer could very well be c) that I don't understand what's going on...)"

The 200 MB figure quoted appears to refer only to the model stored locally on the phone. In my experience, 200 mb translates to a few million parameters in one or more sparse matrices.

The figure on the whiteboard in the background says "Hey Siri small". I take that to indicate the model that does feature extraction and prediction for some queries, such as "set a timer for 20 minutes", while there is a larger, more general model for other use cases in the cloud.

Kiip | San Francisco, CA | Full-time, onsite | http://kiip.me

Kiip is building a mobile ad network focused on contextual relevance.

Our stack: Python, Go, Scala, Java, Objective-C, C, Erlang Backend platforms: AWS, Postgres, RabbitMQ, Celery, Redis, Redshift, Graphite, Statsd

Responsibilities:

* Extract and analyze existing traffic, logs, and usage metrics to proactively scan for issues or improvements for monitoring purposes * Create, improve, and update systems to improve scalability performance and capacity * Contribute ideas and solutions as needed as Kiip's products are developed through 2016 and beyond

We are currently hiring senior backend engineers: http://www.kiip.me/careers/details/job_20160128170214_1RZCXK...

For more information, apply online or email [email protected]

That's pretty sad, seeing Detroit turn from a bustling city to abandoned fields and parking lots

Ann Arbor is about 40 miles from Detroit, and the university there has one of the top AI programs. I believe they are building an autonomous car testing ground for the auto industry as well http://www.technologyreview.com/news/531301/town-built-for-d...

Great article. I've recently been looking into doing parallel optimization with python and found this article on convex optimization useful as well: http://www.ibm.com/developerworks/cloud/library/cl-optimizep...

I've known several people who have transitioned into the tech industry without a CS degree. Particularly good examples that I can think of got their degrees in physics or some type of engineering; I even knew a guy who got his degree in political science. Really, anyone who is interested and motivated enough can learn programming, analysis, and good Unix practices.

There was also a video presentation by Peter Norvig posted a few days ago explaining research at Google done in collaboration with Geoffrey Hinton from the University of Toronto on deep learning at Google: http://news.ycombinator.com/item?id=4733387