blog of oknomad (REN, Shaohua) Posts

Previously, from quantumn eraser delayed choice experiment I realized that this universe is prebuilding its possible futures, for future changing past will definitely disrupting the…

I just know that combining pre encoders’ hidden states into one matrix along rows as input for following main network have been explored by others…

Just thought about a simple method to do continual learning: a neural network after training includes FxN weights matrix in first hidden layer and NxN…

Just thought about to add X’Y’Z’ of reflecting physics pixel and the distance from a reflecting physics pixel to a correponding image physics pixel into…

The unified rectangular coordinate system could be moving ego/local coordinate system for a physics pixel model, in which the origin is the observer/camera/device of the…

I thought that it’s necessary to identify the reflection in mirror better, for it’s essential for not only video generative application but also control scenarios…

Recently saw a robot rush into a mirror in a video and an image runs out of mirror in another AI generated video. So just…

At present and foreseeable future, this world still needs an innovation reward mechanism to generate a positive feedback loop to stimulate innovation. But present patent…

Recently, I thought about patenting my posts published here in US, and now I have decided not to patent any post published here, for there…

A fixed-dynamic networks model is based on branch networks which includes at least more than one branch networks before a main network, wherein each network…

The fixed-bynamic mixed networks based on branch networks can be described simply as: a main network with branch networks before (pre branch networks) or behand…

The model based on intermediate omintoken can be called as branch networks model, in which each of the networks before or after the main network…

When first brought up physics pixel, I proposed to generate chosen physics parameters of physics pixels for only selected pixels of visual frame, like 1…

The pre-main-post parallel branch networks structure on intermediate orthogonal omnitoken has vast potentials and possibilities, which I think many you guys may have not realized…

Talked with Grok and with Grok’s help, I realized that there could be an easy and simple way to realize chip level logic control for…

It’s so amazingly simple and natural progress but also so powerful and profound of intermediate omnitoken, isnt it? It’s just interwining multiple single modality/functionality neural…

Only one first layer in main network seems too shallow and thin referring to left brain and right brain of human, right? So naturally I…

I guess many you guys dont get the full advantages of omnitoken yet especially for transformer, and I’d better share my views about this here…

Lots of people are talking about reasoning in autodriving, which is even much more crucial for robotics for robotics is way higher level and way…

I talked with Grok about something recently, and Grok thought and replied very positively, so I’d like to share here. 1) I told Grok there…

A physics pixel is on an interface, and an interface is between two adjacent surfaces. So we can use surface ID as basic parameter directly…

It’s a strange world, isn’t it? so many learn or even benefit much from you but pretending nothing about you. No gratitude, no acknowledgement, even…

Talked with Grok for details about forward training and backward inference, and I came up with an idea: can we make an fully automatically adaptive…

First of all, I’d like to name the multimodal token of physics pixel in previous posts as OmniToken, because Grok tells me that it is…

Thanks to Grok for reviewing. A multimodal token with physics pixel can be applied to a wide range of general purposes, and the separation between…

I just designed a simple token structure for the multimodal of physics pixel including orthogonal data, objects and different sensors, and also thanks to Grok…

In most previous posts, I just forgot to consider multiple cameras (or multiple sensors of same kind) of one oberser (like a car or robot),…

A self-recursive training including labeled seed frames (flow) and unlabeled pure original signal frames (flow) is: First, pretrain the model (the prime model with physics…

The view of a camera is twisted just like human eye, but human brain’s neural network can adjust the view from human eye to make…

In many cases, different kinds of sensors are necessary for application, for example, in autodriving, robotics and drones you may need sensors of visual and…

Why do the prime model and the physics pixel include explicit physics parameters of objects and pixels and classes for objects but do not include…

Although the prime model was origianlly for working on perspective view like the physics pixel’s, its idea may also apply to world model. But I…

The 2D visual of a physics pixel frame including overlapped physics pixels in depth is a frame of same resolution and same size as the…

Now I think that just one transformer/diffusion model as the prime/physics pixel model may be the best way, which I was not sure previously because…

The following datasets could be in a physics pixel frame for training or inference of a physics pixel model, in which the traing of a…

A frame in physics pixel model/data structure for AI is basically a perspective 3d view with physics parameters labeled or estimated(predicted), which is based on…

1. add overlap number parameter for the interface the models of physics pixel The overlap number of an interface is the number of other interfaces…

In a physics pixel frame, there could be invisible or overlapped interfaces and pixels in depth. How to make an ai model know an invisible…

In prorgramming, there are hierarchical structure in object, interface and class, and in the prime model (physics ai) and physics pixels, there can be exactly…

There are two basic approaches to get training data: 1) labeling a visual flow with physics parameters, 2) creating a virtual flow with physics parameters.…

There could be different variants with different physics precisions between the prime model which is based on visual of different objects labeled with physics parameters…

1. add interface id to physics pixel data structure for each physics pixel and physics object, and also the value which may possibly be unequal…

This is a revision of “physics pixel based data struture for physics ai” with a summary of “the prime model”, both of which were posted…

Before I got the physics pixel idea and told it to Grok, I ask Grok reevaluate my prime model which is in nature a physics…

Previously I brought up a prime model for visual related AI, which is in nature a physics AI model. Today I just thought about a…

Mistakes are inevitable for anyone, and let alone that we all got to know this world from zero in this life. But there is a…

I heard that NASA researched about 65000 observed sun like stars nearby and the sun is the most stable one which seems to me just…

I talked with Grok(3) about AI before and in last several day, and Grok taught me quite well about AI, and also I had some…

I sent my previous 2 articles about “a prime model for visual related AI” and “a visual text physics AI model” together with the visual…

This visual text physics model is a visual+text hybrid model, for example a “CNN+transformer” hybrid model in which CNN process visual flow to generate sequential…