Cancer Innovation Guess – Cancer is a ?classification problem. I made an earlier unrelated discovery. What can you do with loss = MSE( b + c , b). If b is blod and c is cancer then c should be zero. But if everything must be classfied. // Per

June 15, 2019 Per Lindholm

If you cheat the body with gen manipulation. Mix multiple b types. Then cancer will start to develop. c different from zero. Actually the ?only solution to c = model( ). This will help classfication for the body model with inputs c and b. // Per

Machine Learning Test – My trick worked better than expected converged L.Convolution2D(10,10,3) x 4 to 99.17% MNIST // Per

June 14, 2019 Per Lindholm

idx1 = to_cpu(y.data ) < 0.999

idx2 = to_cpu(y.data ) > 1e-3

idx = idx1*idx2

Machine Learning – Interesting result. Some sort of classification separator. Adapt x,y so that they can be classified and the same time be part of a circle /// Per Lindholm

June 13, 2019 Per Lindholm

Machine Learning Newtons Law Is an ?Existens Law. Keep( 1 / inf ) existing and Keep( 1 / 0 ) existing. For 1/0 -> inf. Here you need to shed a lot of extra material. For this mass generates forces to move both away from each other. So the mass can keep existing in a number sense. // Per Lindholm 2019-jun-12

June 12, 2019 Per Lindholm

achine Learning Adapt Math – How to classify if value is an x-cord or y-cord. Add a model_x_zero() resp. model_y_zero() to x and y. They will shape so they can solve the problem and at the same time be close to zero. // Per

June 11, 2019 Per Lindholm

Machine Learning Innovation Guess – Extra Functional Classes. Recalculate with some random extra input. Break calculation if at some point. // Per

June 11, 2019 Per Lindholm

MAchine Learning Math adapts – Solvning Singular? How? I guess by adapting the solution with singular_relu(x). It outputs 0 when undef else x. // Per

June 11, 2019 Per Lindholm

Peace Idea – The simplest thing you can do is to work for peace. This way you don’t have to think so much.

June 10, 2019 Per Lindholm

Peace Idea – The simplest thing you can do is to work for peace. This way you don’t have to think so much.

Regreening The Desert With Bitcoin Money For People, Seedlings and Water? What if buying a coffey is as easy as regreening the desert. Simple choice in my opinion why we need crypto currency. Instant climate change mitigation // Per

June 10, 2019 Per Lindholm

Regreening The Desert With Bitcoin Money For People, Seedlings and Water? What if buying a coffey is as easy as regreening the desert. Simple choice in my opinion why we need crypto currency. Instant climate change mitigation // Per

Machine Learning Speculation – Why wave-forms? Since the problem of fitting a matrix is like a play-springboard. One things goes right and another goes wrong. It could be why wave forms exist. Adjust the springboard for zero total error // Per Lindholm

June 10, 2019 Per Lindholm

Machine Learning Thought – Is square matrices inefficient? Inspired by DNA a 2xN matrix is the best? How do you adapt machine learning calculations to this? Going backwards and forwards transformations f(g(x))

June 10, 2019 Per Lindholm

Machine Learning – How do you prevent dice throwing in Machine Learning. Data sees no differences I think between random dice and updates that seem looking random. Adjust the update vector so it looks less random. // Per stuck at 97%

June 9, 2019 Per Lindholm

Somehow turn random dice into static modelled dice?

Machine Learning – Magnetic field inspired model(). The idea to let mass model layers transform input X to output y under the common loss with the outer field lines. Through another model with a differential equation or mask as extra loss. // Per

June 6, 2019 Per Lindholm

Machine Learning Physics Guess – Should not the magnetic field be turned off some time before fusion starts? Otherwise you missconfigure the mass? // Per

June 5, 2019 Per Lindholm

Amazing Machine Learning Math – Where there is smoke there is fire. Meaning that normal probability function should not be so clean in the long tail. It should smoke out. // Per Lindholm

June 4, 2019 Per Lindholm

Machine Learning Test – If f = modelF( x ) and x = modelG( f ) as its inverse then I could adapt everything so that y = modelG(X) where X is the MNIST number and y its classfication. testing

June 4, 2019 Per Lindholm

Machine Learning Test – If f = modelF( x ) and x = modelG( f ) as its inverse then I could adapt everything so that y = modelG(X) where X is the MNIST number and y its classfication. testing

Machine Learning Physics – Since Energy Can Solve Calculation Problems Maybe There Are Energy Who Can Only Cause Calculation Problems. Different Ion Energies perhaps. Energy classified comming from interaction with ions of different kind.

June 4, 2019 Per Lindholm

Machine Learning Physics – Since Energy Can Solve Calculation Problems Maybe There Are Energy Who Can Only Cause Calculation Problems. Different Ion Energies perhaps. Energy classified comming from interaction with ions of different kind.

Machine Learning ML/AI – Could the set of low hanging fruit ideas be classified? No-brainerIdea( X ) = 1 or 0. If so you could work out important inventions and innovations. // Per

June 4, 2019 Per Lindholm

Machine Learning ML/AI – Could the set of low hanging fruit ideas be classified? No-brainerIdea( X ) = 1 or 0. If so you could work out important inventions and innovations. // Per

Food Waste – In Store Sorting Machine. Keeping fruits, vegatbles and potatoes fresh // Per

June 4, 2019 Per Lindholm

Machine Learning Physics – If everything can and is classifiable then water can decide the truth in Schrödingers cat thought experiment. This means water can teleport energy. Maybe all matter in general can teleport if something needs to be classfied. // Per

June 4, 2019 Per Lindholm

Machine Learning Water Science Guess – Since water is related to life. I wonder if it can generate and teleport different energies so that all problems concerning life can be solved. Here you could tap water calculations ?by analyzing heat perhaps.

June 1, 2019 Per Lindholm

Machine Learning Water Science Guess – Since water is related to life. I wonder if it can generate and teleport different energies so that all problems concerning life can be solved. Here you could tap water calculations ?by analyzing heat perhaps.

Machine Learning Physics Guess – Since energy can produce intelligent thought. IQ as a function does not exist. It is merely different energies that can solve many problems.o energy packets can be classified into what problems they can solve?

June 1, 2019 Per Lindholm

Machine Learning Math – Errors are very frequent as is the ok results. So why cant a function change value depending on if there was an error. F = (F1,F2) where F2 is the function @ error values. // Per

June 1, 2019 Per Lindholm

Made a fermat number. 3^3+4^3+5^3 = 6^3 // Per

June 1, 2019 Per Lindholm

Machine Learning Math – Detachable math – If the target equation is x2+y2 = z2 then I guess there is no reason to use it more than necessary. So there should exist intermediate equations like x2 – y2 = z2 so you can control the path more exactly // Per

June 1, 2019 Per Lindholm

Machine Learning Math – Detachable math – If the target equation is x**2+y**2 = z**2 then I guess there is no reason to use it more than necessary. So there should exist intermediate equations like x**2 – y**2 = z**2 so you can control the path more exactly // Per