TLDR
Introduced in (Dawkins 1976), a meme is an idea, behavior, style, or usage that spreads from person to person within a culture. A meme bank would be a zoo for cataloging and breeding memes.
Meme bank
Disrupting innovation - using brainstorming, TRIZ, Mcluhan’s tetrad (McLuhan 1988), and so on.
Creating a bank of memes seems almost like an idea out of science fiction. Sure there are ontologies and so on, but memes are supposed to be alive, so this is as much a project about algorithms as it is about creating artificial life.
Could a meme bank be used in a procedures like TRIZ to come up with new solutions to open problems or to improve existing ones. Could they be used with Marshal Mcluhan’s tetrad to try and predict change and even disruptions of new media and technology?
Use cases
- attacking a problem
- approaching creativity using generate-and-test
- choosing an approach
- identifying similarity and difference between algorithms etc.
- genetic programming
- generalizing in mathematics
What constitutes a meme?
For the meme bank one might want to collect memes from the wild but also to perhaps establish a meritocracy and allocate resources to memes that seem to be more useful. In science ideas are often formalized as theories. This statement and behavior of theories have been studied and later formalized using the interrelated mathematical branches of formal languages, logic, set theory and category theory. In other domain like fashion or popular culture we may lack such rigor.
The more formaly stated a meme is the fewer minds it can inhabit. But perhpas could
Are memes alive ?
Memes have a number of properties:
- Birth:
- Memes, must orginate with someone
- Spreading:
- The then gain traction over time.
- Decline:
- Many ideas can also lose traction and become discredited or supplanted by more powerful ideas.
- Can they evolve?
- If memes are alive can they breed !?
- How should memes be represented ?
Meme’s don’t have DNA, but they often have a complex genealogy. This suggests that emergent memes contain some aspects of simpler memes that may have vanished from our minds. Why should we care about such basic ideas? It would seem that having a more basic idea is going to be easier to work with. At least in the sense that a simple meme might be easier to utilize or generalize or combine than a complex one.
Specialized type of memes.
- Words (have etymologies)
- Scientific theories
- Data structure
- Frames.
- Scripts.
- Classes.
- Algorithm.
- Design Patterns.
- Religion.
- Media
- Art Clearly we would treat words algorithms differently.
Algorithmic memes:
Problems
If you want to collect memes you should look at problems, as memes we often present as parts of the solutions.
A bank
Mathematical:
- morphism
- equivilence
- mapping
- identity
- order
- reflexive
- transitive
- relation
- axioms
- continuity
- compactness
- maximum and minimum
- distance
- game.
- topology.
- matrix.
- group.
- probability distribution
- grammar
- fsm
- markov chain
- gausian process
Logic
- deduction
- causality
- correlation
- sylogism
- normal form
- skolemization
- completeness
cognitive
- generate-and-test
- sample
- stratified
- snowball
- optimization
- filtering
- smoothing
- search
- Depth first search
- Breadth first search
- Beam search
- MCMC search ?
- Adverserial search ?
- ab search
- Space filling curves
- Hilbert,Peano, Lebesgue, Moore, Sierpinski.
- Space filling trees
- Similarity and distances
data structures:
- sets
- categories
- graphs
- trees
- lists
- linked lists
- array
- tables
- dictionary
- state space
- topology
In my bank:
decomposing PCA into its consituents
vitrtebi alg - converting it to boolean view of transitions + a likelyhood for one happening
is an analysis sing using there is a solution to one with none
variational autoencoder + gan =
sampling + a space filling curve
error propagation though sampling ?
propergating error gradients through a sampling step ?
are there other repramtrization trick we can
spacewise-seperable convolutional layers - can we represent larger
convolutinos as products of 2 or 3 matrices
negative sampling
inverse sampling
follow the leader
follow the regularized leader
regret
recursive matrix alg
non-negative matrix factorization
widen dataframe from wikidata
References
- Steven S. Skiena - The Algorithm Design Manual 2nd Edition
- Thomas Kuhn - The Structure of Scientific Revolutions
- Christopher Alexander - The Timeless Way of Building
- John E. Hopcroft and Jeffrey D. Ullman - Introduction to Automata Theory, Languages, and Computation, Addison-Wesley Publishing, Reading Massachusetts, 1979. ISBN 81-7808-347-7.
Citation
@online{bochman2021,
author = {Bochman, Oren},
title = {Meme Bank},
date = {2021-12-30},
url = {https://orenbochman.github.io/posts/2020/2020-12-30-meme-bank/2020-12-30-meme-bank.html},
langid = {en}
}