Machine Learning Math – In ML math rules does not matter. Its about possibilities. Two rule matrix multiplication with a row and col vector for each matrix A. Why does not matter. The important thing is that ML adapts. // Per
Monthly Archives: April 2019
Machine Learning Water Science Guess – Create a gitter matrix of water. Meassure the real time temperature. The assumption is that heat is needed for water iteration. Follow the energy is like a GAN question in the water matrix. For solving ?any question. // Per Lindholm
Machine Learning Physics ?Relativity Guess. I wonder if light traveling at lightspeed can affect time. Looking at a patch of a light emitting object. I would be very difficult to get a weightmatrix update if not synced. So updates in the patch need to wait for each other. // Per Lindholm
Machine Learning Air conditioner – Could There Exist Cooling Electrolyte? Anode, Chathode ?
Machine Learning EV/Train? innovation – In car racing speedway you need to have ?hot tires. So why not invent an inboard hair dryer or hot air car tire dryer. Could this work for car road handling?
Machine Learning Electric Vehicle Innovation – If water does not like vibrations. Why not use water or some other fluid and measure an onboard container using accelerometers or electric resistance and use the data as input for vehicle settings on the motor and ?suspension. // Per
Machine Learning Battery – How do you construct a battery with an internal ion sorting algorithm. ?Since nothing is random. Different imperfections could perhaps choose among ion ?variations.Could entropy generation be used by the network to create a remember pattern for function.
The Most Important Peace Idea – Water Desalination Education Q and A Live Shows On Youtube? Innovation Needed In Desalination. Problem space for desalination are not unique to desalination it can borrow from other fields?
Other important topics regarding climate change innovation could run on live youtube shows also I think.
Machine Learning Physics – I wonder if the electron path in an atom creates a machine learning weight matrix mask. To change the functionality of the atom at the given instant required. // Per Lindholm 2019-04-24
Machine Learning Battery – Could the ion use its model of the world to meassure the ?speed of light in different directions to find its place in the battery? Or how does it find its place since nothing is random. // Per
Machine Learning Physics Guess – What use have electrons of loss functions? I wonder if it uses energy to finish in light speed time. The number of venn diagram fits is then proportional to the heat generation. To get the level of universe accuracy // Per
Water Desalination – Since salt and water go together for a long time. Water have time to adapt. Thinking in life optimation terms. A fast sort mechanism for water would be water that sticks to stones and rocks. This water is aimed for evaporation and therefor fastsorted.
Machine Learning Beyond Quantum Computers – Universe Computer via Electric Computings Effect (problem space) generating feelings ( answer space ) Get Electric effects by stacking CPU in 3D. Calculations per volume cubes? // Per Lindholm.
Machine Learning Intelligent Physics – Classify acceleration and velocity from entropi as a function of time and space positions. I got the idea from the possibiliy of a venn diagram of a,v,pos but they need to be independent. So how to meassure directly? Entropy_pattern(t,x,y,z) // Per Lindholm
Agriculture Innovation – Water falls downwards so could there exist merge watering of crop land. Using highspeed rotational effects or gravitational effects could there exist flying cropfields. To be used in a merge watering scheme. Recreate the motion of the dance spinning thing on earth generate zero-g?
Machine Learning Water Science Guess – What if water has stored in each molcule or set of molecules the entire set of earth related loss functions? Then you calculate advanced things without knowing the true target data examples. // Per Lindholm
Machine Learning Guess – By Association Water has to do with life so detecting infections or ?cancer should have something to do with water. So I think we need a machine learning water model() and water scientists. // Per
Machine Learning Physics – Padding Input X with another model2([1,0]) could be similar to empty physical space. Its not really empty but a faint iteration of mass to make calculations better // Per Lindholm
Machine Learning Innovation – Super Easy Bandwitdh increase – Two! loss functions. loss = F.mean_squared_error(y,yt) ; loss += F.mean_squared_error(y,1000*yt) // Per 98.3 vanilla 100 unit 3 layer model. Just go from 2 and higher to 1000.
Machine Learning Phyics Fusion Guess – Is the active weight function overfit? Showing streaks compared with a simple MNIST problem suggests this could be a general problem. Overfitting in Physical Plasma Fusion // Per
Machine Learning Invention – Moving Wheel Axle – ?Less Friction if the wheel can find its sweetspot position. // Per Lindholm
Machine Learning – Black holes are Super Dense Weight matrix in regular matrix. Looking at space I wonder if black holes are like a circular masks in machine learning. Just because masks serve a ?function they have to also exist in space. A more dense weight matrix is the answer? More complex problem solver! // Per Lindholm
Machine Learning Generalization Guess – Could a way towards generalization for a model() be to apply multiple problems. X1 = mnist, X2 = line lenghts, X3 = number of point circles. To be solved at the same time. 3 different kinds of input 3 different targets // Per
Business Idea – Sell Pair Phones. Two equal phones for all kinds of possibilities. Always have one phone fully charged or lend one of your phones to your husband who lost his phone on the vacation. // Per
Refugee Livelihood – Sell Vegan sandwich toppings. Vegetables or little meat toppings. Billion dollar idea?
