SELF ORGANISING MAPS: INTRODUCTION

[@mikkelgedehansen9479]

For a channel titled "Art of Visualization" I honestly expected a more...-visual explanation. Aside from that, great explanation.

[@almoh]

Can you share the code? also, describe the hyperparameters?

[@erobusblack4856]

these work wonders for emotional capacity for conscious AI 💯😏

[@viratkumar6328]

How did we compute the lateral distance?

[@FreeMarketSwine]

I didn't see the SuperDataScience username at first, but this needs to make its way into the Udemy Deep Learning course with Kirill and Hadelin. The explanation of dimensionality reduction as shown in this video is not in the course from what I can tell. It would also help to create a 3D visualization that could be rotated to show clusters of nodes. I didn't get it until I saw this video, so thank you.

[@FreeMarketSwine]

Also, based on your explanation, it seems that images like the ones on this page (www.ai-junkie.com/ann/som/som3.html - referenced in the Udemy course) are misleading because the width of the neighborhood is not dictated by the 2D grid but by the location of the neurons in higher dimensional space, correct?

[@chilliking3424]

can you please share the codes? thanks

[@osamaaljarrah9851]

I'm looking for code that compress the data

[@vigneshpadmanabhan]

how to save and test on external data?

[@ngocanle3037]

can you share the code ?

[@wolfisraging]

Great explanation

[@ArtofVisualization]

Thanks for your feedback Rishik!

[@rupeshkamble9105]

Great Work, Thanks a lot

[@Jkauppa]

only forward evaluations in evolution neural networks

[@Jkauppa]

if you in any way label, you dont have an unsupervised analysis

[@Jkauppa]

therefore soms are not unsupervised, if you have a goodness value

[@Jkauppa]

fake algorithm

[@Jkauppa]

like you are doing some sort of pca algorithm by linearly fitting the weight as the data x average

[@Jkauppa]

what you are fitting in the data, when you should not fit anything, only similarity

[@321MrMateus]

Where is the map? XD

[@ArtofVisualization]

Did you watch till the end?))

[@wolfisraging]

Mateus Maia, 6:06

[@321MrMateus]

My Internet bugged lol

[@islandparadise]

It's fascinating stuff! BUT, please don't record a video stoned. Rule of thumb: if one sounds stoned at 1.25x speed, one definitely sounds boring. Please improve and keep churning out good content!

[@ArtofVisualization]

:D I will tell Bruno your opinion :D

[@islandparadise]

Thank you :)

[@otsogileonalepelo9610]

😂😂

[@nikitasinha8181]

Thank u

[@sachinmotwani2905]

My training took 0.07 seconds

[@leoniloris6595]

kinda useless to explain SOM's and then use a pre-built library