Generalisation of reinforcement learning

What do we mean my generalization? By the [[Process and Reality]] “we are no exception to this nature we are example of it”. Taking this Idea we can say that everything is like us has some level of active participation in the nature and during this interaction these things are evolving too. To make this some what abstract Idea that every thing are like us evolving. I like to generalise this Idea that there is an agent rather every thing is an agent....

May 21, 2024 · 2 min · Shriman Keshri

Hopfield Network

This is a type of recurrent neural network. Where each perseptron are connected to other perseptrons. This Hopfield Network also introduced a new theory of Memory. [^1] This is also connected to the problem of 3x3 connected node game. where I was treing to find the [[Modularity]] . In this since the weight of the connection creats a gravity(force) on the state to get converge to a stable state which we call memory....

May 21, 2024 · 1 min · Shriman Keshri

Idea of separation

[[Triangular MDP]] is the simple realisation of Idea of Separation through RL. localisation of causation is the root of separation [[Randomness to separation]]

May 21, 2024 · 1 min · Shriman Keshri

Important ideas

Reinforcement Learning [[Consciousness]] [[Causal inference]] [[Evolution]] [[Ability to use tool]] [[Language]] ( giving identity to our feelings) Randomness [[Reprasentation]] [[learning]] Images of Important Ideas tools [[RL]] Randomness [[Exact sequencec]] [[attention]] [[action = observation]] Circle of time Different feeling Randomness/ Predictability. Time Space Separation/ Identity Intelligence/ consciousness. Drive ( unmoved mover) Creativity. Free will How log is this present? How past -> future? where is action action min comes from? where is love...

May 21, 2024 · 1 min · Shriman Keshri

Information processing system

The concept of Information processing system( IOP) I read in the book Superrecursive algorithms. Def: Something that process information, I will have input, output and Processing unit. Although in the book they assumed the [[triadic structure of IPS]]. While in my writing I don’t always assume that. While I am working on the equality between action (output) and observation (input), sticking on the triad structure of IPS can lead to Inadequate results....

May 21, 2024 · 1 min · Shriman Keshri

Jul 21st, 2023

I manage to setup Logseq in the Nix OS with the Xmonade desktop interface. this will not check on all the lines but at least the line in which I am writing and I think that is more thing good to have this feature activated while I am writing. this is not just fiction time I am an expert in this type of tool in my tool chain. [[diffusion model]].

May 21, 2024 · 1 min · Shriman Keshri

Local-global

Examples. Global $\to$ Local : ( You want to achieve ____ what should I do now? ) Ex. Lagrangian Machanics $\to$ equation of motions ( Calculus of variations) Ex. Reward function $\to$ policy (reinforcement learning) #RL principle of least action Local $\to$ global : [[Dynamic programming]] Ex. all the types of simulations. Evolution and Local-global #Evolution Evolution: What is evolution? Local global: what is local local? Sub-representation and Local-Global #[[sub-representation]] ?...

May 21, 2024 · 1 min · Shriman Keshri

localisation of causation is the root of separation

What I experience is that the thing we see separately is only based on the fact that stuff near can affect each other, and things that are fare away can’t affect each other. We say I am acting on $A$ and not in $B$ because they are separated by space or time. Say I acted on this today, and I acted on that tomorrow. Somehow all these things I think are head to answer are coming and connecting each other and I don’t know form where should I start....

May 21, 2024 · 1 min · Shriman Keshri

Markove decision process

RL interaction Dynamics Markov decision process is a mathematical model. The above interaction dynamics can be studied using the MDP. So here, the MDP comes into the picture. Remark: Because of a lack of knowledge, we use probability in this. #Randomness Remark: When we fix the policy, the above process becomes automatic. It will just run and run.

May 21, 2024 · 1 min · Shriman Keshri

Memory

How do our brain stores the information? it stores its information not in the form it captured it but in the form in which is the most useful for the future. Data Storage #Memory I am in search of general Database management system. This should Coherent with the World view. It will be just like a world i.e. it will have a input and output. Do it need a language? [[Memory is a communication]]...

May 21, 2024 · 1 min · Shriman Keshri