Community Archive

🧵 View Thread

🧵 Thread (11 tweets)

Placeholder
UltimApe@ultimape• almost 9 years ago

Damn, all I have are insane ones that just happen to map to tensors. https://t.co/qxCZSI5DyM

3 0
12/16/2016
Placeholder
UltimApe@ultimape• over 8 years ago
Replying to @ultimape

What even are Tensors anyway?Principal Component Analysis for Tensor Analysis and EEG classificationhttps://t.co/T9VJL6e8oR https://t.co/KsGeC216xW

Tweet image 1
2 2
12/22/2016
Placeholder
UltimApe@ultimape• over 8 years ago
Replying to @ultimape

Unintuitive search for how to understand tensors:"tensors explained with ants"...Oh hey: https://t.co/HK0QPJT6OD https://t.co/dk2NowQF3X

Tweet image 1
1 0
12/22/2016
Placeholder
UltimApe@ultimape• over 8 years ago
Replying to @ultimape

A 2d array of values (X,Y), where value represents height Z is theoretically a suitable surrogate for a kind of tensor transformation? Neat.

1 0
12/22/2016
Placeholder
UltimApe@ultimape• over 8 years ago
Replying to @ultimape

Makes me wonder if tile based data structures, or sparse-voxel-octrees might be suitable abstractions for storing compressed data.

1 0
12/22/2016
Placeholder
UltimApe@ultimape• over 8 years ago
Replying to @ultimape

Oh hey "Our OctNet hierarchically partitions the 3D space into a setof unbalanced octrees"Not even published yet!https://t.co/UfJNiZUqQ9 https://t.co/XpGJBSjuq3

Tweet image 1
3 0
12/22/2016
Placeholder
UltimApe@ultimape• over 8 years ago
Replying to @ultimape

... Now, if only there was a way to increase dimensional space into one that correctly correlates to time, and we get modeling causality?

1 0
12/22/2016
Placeholder
UltimApe@ultimape• over 8 years ago
Replying to @ultimape

How would ants model causality and time delimited phenomena?Well, maybe they count passage of time like bees?https://t.co/gtKM7NsSzs

1 0
12/22/2016
Placeholder
UltimApe@ultimape• over 8 years ago
Replying to @ultimape

"Ours can, by greatly reducing the search space through evolving compact, compressed descriptions of huge solvers."https://t.co/5you3EohsS https://t.co/Z9GhCifb75

Tweet image 1
1 0
12/22/2016
Placeholder
UltimApe@ultimape• over 8 years ago
Replying to @ultimape

"the real trick [..] is that measuring angles is a very technically challenging problem, so instead we measure time"https://t.co/9SEjEIGDt2

3 0
12/22/2016
Placeholder
UltimApe@ultimape• over 8 years ago
Replying to @ultimape

Oh hey, measuring time to get position?Pew pew: That's how light-pens work:https://t.co/gph8xDp14j

1 0
1/2/2017