Group Abstract Group Abstract

Message Boards Message Boards

6
|
701 Views
|
19 Replies
|
11 Total Likes
View groups...
Share
Share this post:
GROUPS:

[WSG25] Daily Study Group: Neural Networks for AI

19 Replies

Just click Submit, and we will receive your completed quiz. Certificates will be issued after the deadline of November 7. Thanks for your participation!

POSTED BY: Jamie Peterson

Hi All; I have completed the Quiz wsg72 format on my Profile and push the SUBMIT bottom: it is already ok so, or I have to send quiz page to some address? Many thanks! Best.

Gianluca Argentini - Italy

POSTED BY: Arben Kalziqi

Happy to hear it!!

POSTED BY: Arben Kalziqi

Yet in the webinar it appears that students are interacting with the lecturer. This perhaps means that the webinar is recorded and not live?

Posted 1 month ago

Thank You - I thought I was missing something simple - it worked perfectly - By the way Great Course I am learning a great deal and it is very much appreciated!!!

POSTED BY: Randy Janke

Hm—if the light is perpendicular, we can go ahead and think of it as lasers. But in that case, I don't think that you can even physically tell where the individual components came from just by looking at the surface on which they intersect, so I don't know how a neural net could do it. Maybe I misunderstand something, though!

POSTED BY: Arben Kalziqi
Posted 1 month ago

Hi,

I have a probably convolution or correlation layer type question.

Imagine a bay window with 3 panes. In front of these window panes from a good distance are lamps positioned that the emitted light is perpendicular to the particular window pane. These light rays are converging on the back wall of the room in a small area. The question is, if someone makes an image of the light on the wall can a neural network separate where portions of the lights were coming from? Lamps are the same and they are from the same distance from their particular window pane.

My guess is is that somehow the different directions - 45, 90, 135 degrees - has to be picked out, but how?

POSTED BY: Janos Lobb

Do you mean that you can't post Q&A? We don't have public chat enabled in our webinars, generally speaking.

POSTED BY: Arben Kalziqi

my chat is disabled, so I cannot interact in the class. Can you fix it, please?

That's happened to me too :). The definition for dcf is indeed hidden in a closed cell somewhere in that notebook. Here's the definition itself:

dcf = Blend[System`PlotThemeDump`$ThemeDefaultMatrix, #1] &

This is grabbing one of the default color schemes—the one used for MatrixPlot, for example—and making it available for your use. (It's also converting it from a scheme with discrete colors into a continuous one.)

POSTED BY: Arben Kalziqi

Hi Gerald! Sorry for the late reply; something seems funky with my notification settings for this thread.

Like NetBlockPlot3D, this is a custom resource that either Stephen or the blog team created for one of his blogs, so there's no real documentation that you can grab. However: I've added a notebook to the downloads folder where I pull out the colors/color functions used in the notebook that should be instructive!

I'll also include it here.

POSTED BY: Arben Kalziqi
Posted 1 month ago

The auto correction turned "dcf" into def - sorry I did not check first before sending

POSTED BY: Randy Janke
Posted 1 month ago

I am having a little trouble with the Machine Learning Basics notebook with regards to the option def for the ColorFunction - when I run the code the output is all red - when I change the option to one of the builtin options such as "AuroraColors" the output looks good - I am wondering about where or how the def was defined. I think I am missing something simple. Thanks

POSTED BY: Randy Janke
Posted 1 month ago

Thanks for the response. I did some more digging and found this section in a old blog post about support for Core ML as a target device. I decided to give it a try using examples in the Wolfram Documentation. In my tests the GPU was 3X faster than the processor, while CoreML was 6% faster than the GPU.

POSTED BY: Jon Rogers
Posted 1 month ago

In Notebook WSG72-1-MachineLearning.nb, the Veronoi mesh input cell includes within the With[...] instruction an iconized ChatTechColors . When I try to get information about ChatTechColors by mousing over it and selecting the left icon, there is no information in the popup window. Selecting the right icon yields this message: "ChatTechColors: The source notebook cannot be opened." If possible, please make the source code and documentation available.

POSTED BY: Gerald Dorfman

Hi Jon! Totally fair to ask questions beyond the scope of the course—especially if you kindly indicate that you're aware that you're doing that ;).

My understanding is that (currently!) CoreML can be used when running a neural net (or using a built-in function that uses neural nets "transparently" to the end user), but it can't be used when training a neural net.

I can't say that I'm aware of MLX, but a quick look through some internal information seems to indicate that it is in use in at least some places!

POSTED BY: Arben Kalziqi
Posted 1 month ago

I have a question about Mathematica and machine learning on Macs. This may be out of scope for this study group, if so please let mew know.

I am trying to get my head around how Mathematica utilizes Apple's Core ML framework for processing machine learning tasks on a Mac (or if it does at all). This is on my mind because I have started reading about Apple's new machine learning framework MLX. If I understand this correctly the Core ML and MLX frameworks are tools for developers to get the best machine learning performance on Apple Silicon. Are these frameworks relevant for Mathematica?

POSTED BY: Jon Rogers

Looking forward to this Study Group next week! We have a great group signed up, but seats are still available. Sign up here.

POSTED BY: Jamie Peterson
Reply to this discussion
Community posts can be styled and formatted using the Markdown syntax.
Reply Preview
Attachments
Remove
or Discard