Machine Learning progress – Last layer is the prediction of the loss. As predictions are easy when loss is constant zero. It helps as an extra loss function. Last layer is the predicted loss. // Per
Machine Learning Insight – Input to label holds no visible logic to the model. Its kinda like presenting accurate guesses. Why not add logic to the labels. Like watching someone perform something to a result. Video labels.
Association Intelligence – Could nano material structures enable heat-radio? Radio is communication so is heat in a way.
Artistic Intelligence – What inventions could exist from painting with heat. Steer ions and electrons? // Per
Electric Car Battery Guess – Efficiency is the same thing as machine learning model accuracy. There must exist generalized efficiency. // Per
Solve the energy logic to 99% by physical super models.
Thanks Eben – Seamless move No os reinstallation between rp3 and rp4. Using barryboot 2.0 for rp4. Just attached the usb ssd with an installed MATE.
Mega Peace Idea – Train on a wheel with breaks based rail sytem. Replace railway cheaply! Put railway anywhere you like with wheels on. Mega Peace Projects. Puts the breaks on peace. Make it permanent. // Per
Machine Learning Math – Multi Loss function selection based on the most interesting results. Select only those loss functions that makes an interesting result. Testing…
Machine Learning Intelligence – Intelligence is capturing all the possibitilites in programmatic mathematics. // Per
Incredibal calculation? Predicted gravity from solar planet distances – Using an augmented classification helper. X = X + model2(yt) loss = F.mean_squared_error(y,yt) loss += F.mean_squared_error(model2(yt),model2(yt)*0) Then model2(yt) is not used in the prediction test. // Per
Could increased heat mean that the material want to randomize the quantum residuals even more. Help the material by blowing turbulent air?
Battery Innovation – Could there exist an adapted material as the spacer in a battery pack. Batteries emit heat so there are ?quantum residuals going with it. How do you chanel those so the next battery is not suffacated. Blow air around the batteries?
Could you put fire proof sensors in the fire to get a xray like data of the various gas levels? A forest fire black box recorder?
Probabalistic fire extinguisher – Forest/Bush fire and Water Science – If we assume everything is machine learned causal. Get detailed image maps of water effect on temperature, gases levels. Images that machine learning scientist could use to derive a water put out fire model.
Water Science – If we assume everything is causal. Detailed image maps of temperature, gases levels. Water targets. Its behavior in water etc. Could reveal howto stop those fires.
Important – Forest Fire Science – Record the Forest/Bush Fires for machine learning science we can develop Super Fast Measures for fighting those big fires. // Per
Invention – The washer keep breaking. Gives me an idea. Clothing Washer without moving parts. Like a dishwasher. // Per
Machine Learning Incrimental Progress – using F.arcsin((F.normalize(self.l1(x)**2-1))) for activation works better than expected. Over 98.3% on MNIST MLP. // Per
Compression Idea – Lossy (JPG) compressed HTML or text files and AI or machine learning to error correct the uncompressed text. Wow // Per
Education Innovation – Blaming the school system is a long futile wait? There should exist an open source community that makes online school courses in mathematics. A wordpress site could cost as little as 2 dollars a month.
Idea – Since everything must be classifiable. Defects could be uniqueness givers . Like an adress? // Per ‘Detours’ might lead to better batteries https://futurity.org/batteries-defects-lithium-ions-2244392/… via @FuturityNews
Like a computer memory has memory adresses.
Math Innovation Idea – Differentiated pow(x,a). Why assume that all factors are exactly the same. x**2 = x0*x1, x**3=x0*x1*x2. In machine learning math you could generate a new value from a model.
Use x*x as a limit target for x0*x1
Super Idea – Freeze water to heat the house and buildings. Capture the heat from the efficency loss. Save water save the world. Buy standing freezers and ?put a fan on the backside to the heat spreader. Fill it with desalinated water bottles or plain. Water heat pump for the win!
Home Experiment – Guess. With less arctic ice the water should freeze at a slower speed. Maybe the freezing water gets quantum residuals teleportated that causes heat and a longer freezing time. Put water in cocacola bottle in the freezer and measure the time all over the world.