Greedy, Brittle, Opaque, and Shallow: The Downsides to Deep Learning

Re:Greedy, Brittle, Opaque, and Shallow: The Downsides to De... by gaius on Saturday February 10, 2018 @03:13PM (#16349324)

Calling it "deep learning" was the first mistake. It makes it sound a lot more profound than it really is. "Machine intuition" is the term I prefer.

[ Parent | on HN ]

Re:Greedy, Brittle, Opaque, and Shallow: The Downsides to De... by aoeusnth1 on Saturday February 10, 2018 @03:27PM (#16349391)

I like "machine perception". We're not much past the point of sensory-motor development.

[ Parent | on HN ]

Re:Greedy, Brittle, Opaque, and Shallow: The Downsides to De... by enraged_camel on Saturday February 10, 2018 @03:43PM (#16349468)

I have heard it called computer vision.

[ Parent | on HN ]

Re:Greedy, Brittle, Opaque, and Shallow: The Downsides to De... by mkl on Saturday February 10, 2018 @03:59PM (#16349549)

That's different. Some deep learning models are used for computer vision, and some computer vision involves deep learning. Neither is an essential part of the other.

[ Parent | on HN ]

Re:Greedy, Brittle, Opaque, and Shallow: The Downsides to De... by contrarian_ on Saturday February 10, 2018 @06:07PM (#16350162)

"linear transformations with soft thresholds sandwiched in between"

[ Parent | on HN ]

Re:Greedy, Brittle, Opaque, and Shallow: The Downsides to De... by therein on Sunday February 11, 2018 @02:07PM (#16354180)

LT-STIB = Linear Transformations w/ Soft Thresholds In Between

[ Parent | on HN ]

Re:Greedy, Brittle, Opaque, and Shallow: The Downsides to De... by salty_biscuits on Saturday February 10, 2018 @05:55PM (#16350110)

How about "long skinny networks" over "short and fat networks". Deep maps to profound in my mind rather than just talking about lots of layers.

[ Parent | on HN ]

Re:Greedy, Brittle, Opaque, and Shallow: The Downsides to De... by Quarrelsome on Saturday February 10, 2018 @04:57PM (#16349857)

I prefer "just throwing cycles at the problem".

[ Parent | on HN ]

Re:Greedy, Brittle, Opaque, and Shallow: The Downsides to De... by floatboth on Saturday February 10, 2018 @05:15PM (#16349953)

You can throw cycles in various ways (like with regular algorithmic code that implements various heuristics or something). So, "just throwing cycles at statistics"

[ Parent | on HN ]

Re:Greedy, Brittle, Opaque, and Shallow: The Downsides to De... by gaius on Sunday February 11, 2018 @09:14AM (#16352756)