Joseph Van Name
1.59K subscribers
20:49
I tried to disprove the existence of a higher infinity and produced these tables instead.
Joseph Van Name
43 views • 10 hours ago
21:17
Ways to append an ordinal to a classical Laver table of size at most 128: 2-adic ordering
Joseph Van Name
28 views • 3 days ago
15:29
Critically simple Laver-like algebras with 2 generators
Joseph Van Name
61 views • 5 days ago
9:31
All 6854 ways to append an ordinal to a classical Laver table of size at most 64: 2-adic ordering
Joseph Van Name
59 views • 5 days ago
9:31
All 6854 ways to append an ordinal to a classical Laver table of size at most 64
Joseph Van Name
44 views • 5 days ago
15:28
Hill descending algorithm produces up to size 56 by 56 Hadamard matrices with 2 circulent cores
Joseph Van Name
97 views • 7 days ago
5:56
Hill descending algorithm produces up to size 44 by 44 Hadamard matrices with 2 circulent cores
Joseph Van Name
114 views • 7 days ago
10:53:14
All 39194 superreduced multigenic Laver tables with 2 generators of size at most 50: Over 10 hours
Joseph Van Name
110 views • 9 days ago
1:09
Finding a vector on a torus that remains on the torus after a unitary transformation
Joseph Van Name
169 views • 13 days ago
1:07
Artificial intelligence produces a bent vector of length 256.
Joseph Van Name
105 views • 2 weeks ago
1:20
AI finds a bent vector of length 256 by moving along a vector field in the real projective space
Joseph Van Name
171 views • 2 weeks ago
1:35
This AI finds a bent vector of length 64 by training 64 complex matrices.
Joseph Van Name
194 views • 2 weeks ago
1:04
AI finds a bent vector of length 256 using a non-random complex-valued initialization
Joseph Van Name
163 views • 2 weeks ago
0:43
AI finds a bent vector of length 256 using a non-random initialization
Joseph Van Name
288 views • 2 weeks ago
0:50
This AI finds a bent vector by moving in the direction of a vector field that is not a gradient.
Joseph Van Name
2K views • 2 weeks ago
2:15
Evolutionary computation produces a bent vector
Joseph Van Name
469 views • 2 weeks ago
1:20
Artificial intelligence produces a 20 by 20 Hadamard matrix.
Joseph Van Name
2.3K views • 2 weeks ago
5:50
A neural network with an unusual activation function learns the sine function.
Joseph Van Name
3.2K views • 3 weeks ago
13:41
Training a neural network on the sine function.
Joseph Van Name
46K views • 3 weeks ago
11:52
More machine learning recovering handwritten digits from the discrete Fourier transform magnitudes
Joseph Van Name
209 views • 3 weeks ago
11:07
Machine learning recovering handwritten digits from the discrete Fourier transform magnitudes
Joseph Van Name
411 views • 3 weeks ago
20:00
Human MNIST Challenge: Can you recognize the Fourier transform of handwritten digits?
Joseph Van Name
374 views • 3 weeks ago
1:17
A neural network learns then suddenly and catastrophically fails on all data during training.
Joseph Van Name
820 views • 4 weeks ago
3:21
Weight matrices during training of a neural network that computes XORs of 40 bits.
Joseph Van Name
157 views • 4 weeks ago
2:00
Comparison of 5 activation functions in a neural network memorizing data (3 non-linear layers)
Joseph Van Name
197 views • 1 month ago
1:05
Factorizing the all 1's matrix as a product of three 0-1 matrices using gradient descent
Joseph Van Name
469 views • 1 month ago
2:08
Discrete Fourier transform of reversible linear cellular automaton with low period
Joseph Van Name
620 views • 1 month ago
3:18
We use gradient descent to find a 5-coloring of a graph.
Joseph Van Name
634 views • 1 month ago
0:49
Visualization of Boolean 3-satisfiability solved using gradient descent
Joseph Van Name
173 views • 1 month ago
0:31
Maximizing the spectral radius: A discrete optimization problem solved using gradient updates.
Joseph Van Name
139 views • 1 month ago
Load More