Machine Learning Idea – Unsupervised Clustering With supervised models() via loss functions. What can a classification of f(x) = x look like. Smooth transitions?
Monthly Archives: January 2019
Machine Learning Physics – Why should the atom have a spherical coordinate system when it ?could self-organize to meet any problem. I Wonder If It Just Dynamically Alternates Between different coordinate systems spherical, polar, cartesian?, … // Per Lindholm 2019-jan-31
AI Wordbook – What if a black hole as a dense object takes advantage of all other ?dense objects like prime numbers. Prime numbers I guess are dense after the definition of primes.
Math Guess – Does There Exist Variables That Are Not Allowed To Solve The Equation Alone? Me
My Math Fermat Guess – Taking the derivate dx~dy in the equation x^n+y^n = z^n I guess the ratio dz/dx ~= (x^(n-1)+y^(n-1)) / z^(n-1) must be able to have a prime number for z which holds true for n = 2 not 3,4,5,… for 3x,4x,5x the ratio is 7/5
Traveling Salesman Problem A Step Closer? – Looking at the reflection in the sphere you ?could optimize the path i a circular map to plane fashion? At least the light going straight paths has to find the minimal path to go straight.
Machine Learning Physics Guess – Every state of matter must be evolutionary defendable. So space nebula gas as an entropy example would pick up energy fast. Iterate over vast distances fast. Thats my guess. A fast heat equation.
The Triangulum Emission Garren Nebula NGC 604
The pattern could be so a ?Fast diff. equation could take place.
A.I To Support billions of people and their livelihood. That’s is Big. Saving The World Many Time Over Idea – Make machines Like This Only Cheap and Big Enough To Fit A Small Livelihood. New Catagory Products Household To Livelihold
Fire Protection Innovation Guess – Active Firewall With Circulating Water. Cooled ?Under Ground. Like a water cooling CPU block. // Per Lindholm 2019-jan-26
Machine Learning Math – Key Finding? – Every Equation can be modulated with the 1.0*term replaced by Limit constants. So c0,c1,c2,… = model() and F.mean_squared_error(c0,1.0). As the solution progresses. c0 in the equation approaches 1.0 which corrects the equation // Per Lindholm 2019-jan-26
Machine Learning Physics My Thoughts – Battery Rewritten Becomes Threshold Battery. Innovating Different Thresholds And You Get Different Batteries. Could A Circuit Provide a Threshold In That The Electron Problem Changes. Until Discharging When The Threshold Becomes A Catalyst.
Energy Innovation – Should There Exist Metal Fuel Cells? Run on hot or cold metal?Since the by product ?will contain a little CO2 and steel holds a little carbon. Im wondering if there is a high carbon iron mix. A fuelcell metal. And since high carbon iron is brittle. Could this be a similar problem to dry old wood. It to works as a fuel // Per Lindholm 2019
Agriculture Innovation Guess – Should There Exist Something Like Soil Design. Similar to material design. Layers With Different Abilities Stacked Or Otherwise Working Together
Machine Learning Physics – A machine learning network Is Similar To Electrons In A Wire In That It ?Is Push To Push Mathematics. Then everything you can do with electrons should exist for networks. Coil Networks. Network To Network Induction,…
Machine Learning Physic Ultra Battery? – Why Guess – One Reason For Electricy Could Be To Heat Mass Quickly At Different Places At The Same Time? Then The Idea Is To Learn A Mass?Mesh Network That Also Transports Heat By Electricity? Then Tap That Electricity For Power. I think you need pressure // Per L
Machine Learning Math Guess – Could Singularity Free Math ?Be Dependent On Freewill. Some kind of elbow room or anti freeze. What if there is an associated normal distribution to every value iterated in a model(). This so it can iterate itself out again if in trouble // Per Lindholm 2019-jan-24
Machine learning – AI working as a clue giver. Just because its probably easier to do this than let AI design a theory by itself.
2D Recognition Art – Aided by Noise3D Blender. Creating fantasy characters is not easy but as a start you can create a random object to use your own recognition or why not develop a machine learning recognition algorithm to select one good start. Below a winged ant animal : )
Google Colab Idea – Please Extend Google Colab With A Linux ?VNC Instance To Be Run From The Phone,Tablet or PC from any operating system. A Personal Linux Cloud Computer Free For All. All the needed software already installed. With this poor people can have a computer like everyone else. To make a living, to get money for food That would be a major wow.
Livelihood India, Bangladesh, Rohingya Idea – Android App Tutorials Screencasted From The Phone Uploaded To Youtube,… . Billions of apps, billions of opportunities.
Find Land For The Rohingya – Make them self supportive. Jobs, education, agriculture. Then it does not cost much. So we need brokers to find land for these refugees.
Machine Learning Physics Fusion Guess – Assumption. Everything In The Sun Is Important To Fusion. Machine Learning GAN Networks Depend On Noise. Mean The Sun GAN Solves Its Problems Using Turbulence Which Looks Like Noise. This insight is important since 1 question GAN’s produce badly recognizable images. To my experience. Is a 100 questions GAN enough? // Per Lindholm 2019-jan-21
Machine Learning Physics – Is Dark Matter Another Universe Making A Dynamic Intersection With Our Universe? It should give a gravity effect? // Per Lindholm 2019-jan-20
Machine Learning MNIST Innovation – What if you learn 1 + 1 = 2, 1 + 2 = 3 etc. Complexity Learning? That is. Presented with Two Samples. ?Not Using A Algorithm. Just A model Transform Add Two Samples As If It Is A Recognition Problem. Funny pic..
