Year: 2022

BLACK HOLES AND BRAINS. “In this paper we shall attempt to take one more step and try to see if a neural network can possess some additional informational characteristics that are usually considered to be the defining properties of black holes, such as the Bekenstein's area law entropy.” https://onlinelibrary.wiley.com/doi/full/10.1002/prop.201800007?casa_token=GdPRXs_tGLEAAAAA%3Ag-pxPMtkcleTYDUNK3mYGfpEzKRyc2LqQknTgFFB7dkYFDL8rDe2s0f4_0Q0oKjXD4_AEgFPGWolVTQ View in LinkedIn

INFO STORAGE. “It was shown that a theory of a quantum field inhabiting a d‐dimensional sphere and experiencing an angular‐momentum‐dependent attractive interaction exhibits a sharp enhancement of memory storage capacity in a critical state in which the lowest angular momentum mode is macroscopically excited.” https://onlinelibrary.wiley.com/doi/full/10.1002/prop.201800007?casa_token=GdPRXs_tGLEAAAAA%3Ag-pxPMtkcleTYDUNK3mYGfpEzKRyc2LqQknTgFFB7dkYFDL8rDe2s0f4_0Q0oKjXD4_AEgFPGWolVTQ View in LinkedIn

GAPLSSNESS. “This excitation lowers the energy thresholds of the higher angular momentum modes, so that the theory delivers a set of the gapless modes with their number scaling as the area of a ‐dimensional sphere.” https://onlinelibrary.wiley.com/doi/full/10.1002/prop.201800007?casa_token=GdPRXs_tGLEAAAAA%3Ag-pxPMtkcleTYDUNK3mYGfpEzKRyc2LqQknTgFFB7dkYFDL8rDe2s0f4_0Q0oKjXD4_AEgFPGWolVTQ View in LinkedIn

SOURCE OF STORAGE. “The micro‐states of these modes house an exponentially large number of patterns, with the corresponding entropy given by the area of the same ‐dimensional sphere.” https://onlinelibrary.wiley.com/doi/full/10.1002/prop.201800007?casa_token=GdPRXs_tGLEAAAAA%3Ag-pxPMtkcleTYDUNK3mYGfpEzKRyc2LqQknTgFFB7dkYFDL8rDe2s0f4_0Q0oKjXD4_AEgFPGWolVTQ View in LinkedIn

THE MODEL. “The quantum brain network that we shall construct in the present paper is nothing but a neural network implementation of the same model. This is achieved by means of identification of the neuron‐describing degrees of freedom with the number operators of the angular harmonics of the quantum field.” https://onlinelibrary.wiley.com/doi/full/10.1002/prop.201800007?casa_token=GdPRXs_tGLEAAAAA%3Ag-pxPMtkcleTYDUNK3mYGfpEzKRyc2LqQknTgFFB7dkYFDL8rDe2s0f4_0Q0oKjXD4_AEgFPGWolVTQ View in LinkedIn

MEGASCALE THOUGHT MACHINES. “These machines, referred to as Matrioshka Brains, consume the entire power output of stars (~10(26) W), consume all of the useful construction material of a solar system (~10(26) kg), have thought capacities limited by the physics of the universe and are essentially immortal.” https://lnkd.in/gxfU_Yr View in LinkedIn

This annual verbena has half of its petals painted red, and radially arranged facing inwards to the center — a feat of genetic orientation instructions. View in LinkedIn

GEOMETRY AND NEURONS. “Such mapping provides a way in which the sense of locality and geometry can be brought into something intrinsically non‐local, such as the neural network.” https://onlinelibrary.wiley.com/doi/full/10.1002/prop.201800007?casa_token=GdPRXs_tGLEAAAAA%3Ag-pxPMtkcleTYDUNK3mYGfpEzKRyc2LqQknTgFFB7dkYFDL8rDe2s0f4_0Q0oKjXD4_AEgFPGWolVTQ View in LinkedIn

EXTREME ANALOGY. “However, in our construction we shall not simply copy the quantum field model, but rather arrive to the desired neural network by a bottom‐up approach. For this purpose we design a neural network in which the synaptic connections are closer cousins to the gravitational interaction as compared to the ones studied in [1].” https://onlinelibrary.wiley.com/doi/full/10.1002/prop.201800007?casa_token=GdPRXs_tGLEAAAAA%3Ag-pxPMtkcleTYDUNK3mYGfpEzKRyc2LqQknTgFFB7dkYFDL8rDe2s0f4_0Q0oKjXD4_AEgFPGWolVTQ View in LinkedIn

CYBERNETIC BRAIN. “Namely, we study a network in which the synaptic connections among the pairs of neurons are energetically negative and also proportional to the products of their energy thresholds.” https://onlinelibrary.wiley.com/doi/full/10.1002/prop.201800007?casa_token=GdPRXs_tGLEAAAAA%3Ag-pxPMtkcleTYDUNK3mYGfpEzKRyc2LqQknTgFFB7dkYFDL8rDe2s0f4_0Q0oKjXD4_AEgFPGWolVTQ View in LinkedIn