Raspberry Pi Innovation – I got the idea for a multi screen monitor. Like 8 inch tablet or one that fits below the screen. Could be solved with a screen holder for mobile or tablet screens. A no brainer idea. // Per
Monthly Archives: May 2019
Machine Learning Electric Plane Innovation – Use machine learning model() to mitigate beam bending of the wings. To know when to use power that is. Every reltative position and event in the plane body must coincide with the motors. // Per
Machine Learning Electric Plane Innovation – Use machine learning model() to mitigate beam bending of the wings. To know when to use power that is. Every reltative position and event in the plane body must coincide with the motors. // Per
Important nature medicin Innovation – Curry spice mixed with milk and boiled water. The yellow color I guess has to do with cleaning your body of harmul stress.
Food Innovation – White Point Cabbage and doritos chips. Wow what a feeling. Don’t eat much though. Pure gold. // Per
Machine Learning Battery Guess – I wonder. If you ?decrease the pressure for a volume with a surfplus of electrons. Will the universe teleport extra electrons to that place to attract ions and restore? Some similar problem could exist where the universe need to compensate. //Per
Machine Learning – Could mass rotations be thought of as classifications of machines. Since vectors in a field give solutions to diff. equations and triangles in a field give rise to mesh objects. So rotational circles should give something. Gears predict a certain machine. Vibrations predict new machines? //Per
Machine Learning Innovation on the Rasberry Pi – Trace the hidden weight matrix with plt.scatter(http://model.l1.W.data [:,0],http://model.l1.W.data [0,:]) . Could lead to generalization if you compare the direction and stop the points takes. // Per
Raspberry Pi 3 as my only computer. Just since its possible. // Per
Browser Innovation – What about firefox and chromium with built in python interpreter. So you can code other languages than javascript in the browser. No change to web html but since you can solve problems with python. Why not. // Per
Machine Learning Innovation – What if .Adam or .SGD has a 97%efficiency rate. How do we model a soft update optimizer?
Machine Learning Innovation – What if .Adam or .SGD has a 97%efficiency rate. How do we model a soft update optimizer?
Math Innovation Guess – Could the continuum be augmented with free fall limited functions. Functions similar to chains. Limited freefall functions. // Per
Math Innovation Guess – Could the continuum be augmented with free fall limited functions. Functions similar to chains. Limited freefall functions. // Per
Novice Programming Innovation – Python Pyzo mark lines and run selection. Can debug inside for loops and functions. Way good.
aspberry Pi Innovation – Pyzo interactive python IDE. Really very good. Running on Pi3 berryboot MATE16. Install with $sudo pip3 install pyzo // Per
Raspberry Pi – Running machine learning on the Pi with pyzo a light weight python ide. Chromium was a little bit slow for google colab. They really should have a lightweight client for the terminal and desktop for the rpi.
Machine Learning Fusion Sun Guess – Could you take advantage of a classification flow. Flow from indata to reductions classifications. Just meassure where the reduction points are and make something ?similar. // Per
Machine Learning Tests – Since anything can be input to function. There is much important to predicted. An image can produce a histogram. What about the inverse. Learn histrogram data sample to corresponding image. // Per
Machine Learning Innovation – Could you ask a model if it forgot or lost its sample value. Like 0/1 question. So another model_memory(X) tells if model_brain remembers a sample. It predicts a number between 0..1
Machine Learning Innovation – What if you need to use a radio multiplexer -demultiplexer setup for machine learning models(). Looks like a solution for similar problem. // Per
Machine Learning Math Guess – Top Number object I guess would be similar to a black hole. You cant hole+1 since it sucks in everyting.
Machine Learning Math – Classification Thought – Why discard errors if you could move them. They carry information so otherwise you risk another error appearing. // Per
Peace Medicine For a Tiger – A little Curry, Milk and Boiled Water in cup and sip a little. Perfect
Machine Learning Battery Idea – Since you need energy left in the battery to be ion-move-ready there should exist concentrated heat battery. Where the mass model adapts its move paths to a energy source. Like laser heated dots? // Per Lindholm
Machine Learning Math Adapts – You get Not-Linear from F.relu(x) where x could be a neg. number but the result is positive. The same way you could do for complex number plane. 4 sets in each quadrant is filtered so you have 4 not-quadratic solutions. F.relu4(x+1j*y,quadrant=0) // Per
Guess – A reason why bananas make themselves vulnerable to disease is attracting insects. The fruit A.I system is intelligent. From this the conclusion is that we need more insects that eat fruit. Or otherwise they will become super fragile. // Per
Machine Learning Infinite Equal – Adapt the model after an inf. number of equal signs. Could this be done? If the number of equal signs are random then the model can jump over a lot of tests or adapt to that enviroment. Equal signs are loss functions // Per
