Daylen's Instinctual Dump

Where are you and where are you going?
daylen
Posts: 1678
Joined: Wed Dec 16, 2015 4:17 am

Re: Daylen's Instinctual Dump

Post by daylen »

Who is more insane?

1. An agent that communicates to check their sanity?
2. An agent that attempts to summarize what other agents are communicating to check their own sanity?
3. An agent that thinks about whether or not this question deserves consideration?

daylen
Posts: 1678
Joined: Wed Dec 16, 2015 4:17 am

Re: Daylen's Instinctual Dump

Post by daylen »

Eventual

Objective: Construct a web app to search, scrap, predict, and alert users of events.

Meta: Such events are spacetime intervals on Earth that significantly(*) impact human agents (e.g. wars, pandemics, natural disasters, industrial accidents, artistic productions, technological inventions, scientific revolutions, political upheavals, migrations, extinctions, speciations, etc.).

(*) Initially based upon my own interpretation.

Stage 1: Develop a bare-bones database of historical events in an events.json file. The subject is what multiple sources/agents agree to name the event, the objects are uncommon nouns/verbs that are used by multiple sources to talk about the event, and the spacetime key is where and when the event happened.

{ "subject": { "objects": [ ... ], "spacetime": {} } }

Stage 2: Develop additional attributes that may further aid analysis. Possibly construct a script to auto generate event data from select sources.

Stage 3: Develop the app logic that translates attribute sets into geometry to be viewed by users.

Stage 4: Develop a user interface to search and filter these views primarily through event objects.

Stage 5: Develop web scrapers that gather data to construct current events.

Stage 6: Develop event constructors and app logic to continuously update the associated event views.

Stage 7: Develop an alert system.

Stage 8: Develop a learning algorithm to anticipate the possibility of near-future events.

Stage 9: World domination... jk
Last edited by daylen on Sun Oct 04, 2020 6:39 pm, edited 5 times in total.

daylen
Posts: 1678
Joined: Wed Dec 16, 2015 4:17 am

Re: Daylen's Instinctual Dump

Post by daylen »

Within' this frame, Existence is an unordered set of events emergent from agent communication (i.e. storytelling). History is typically referenced linearly. Eventual will help agents gain a more non-linear view of history by abstracting away semantical details into geometry/topology. In the later stages it may even supplement news or help to vaguely anticipate near-future events. The idea is not to develop a machine that thinks like humans but to develop a machine that thinks about what humans would rather forget. This complementary technology should ideally help agents transform themselves by attending to a wider view of spacetime and by finding relevant details quicker. Converting Si into Ni or Ni into Si.
Last edited by daylen on Sat Oct 03, 2020 4:36 pm, edited 4 times in total.

daylen
Posts: 1678
Joined: Wed Dec 16, 2015 4:17 am

Re: Daylen's Instinctual Dump

Post by daylen »

Agency

Objective: Construct an English communication parser that differentiates and types agents.

Meta: Allows Eventual to correct for bias in event reporting. May also tangentially aid personal development.

Approximate Functionality:
  • Differentiate POV.
  • Detect object references and attribute relevant characteristics using Eventual.
  • Use an event constructor to track the objective context.
  • Assign a probability vector to each statement/question for external/internal, denotative/connotative, and perception/judgement.
  • Differentiate statements/questions using progressive typing and frequency analysis.
  • Track possible agents using an agent constructor.
  • Return attention-topology and type-distribution.
  • Potentially dress output with a viewer for human readability and visualization.
Type Model:

external/internal - Based on reference to objects and their proximity within the event holarchy.

denotative/connotative - Based on how objects are referenced in conjunction.

perception/judgement - Based on object selection/attraction relative to other agents with similar event holarchies.

External Denotative Perception - Se
External Denotative Judgement - Te
External Connotative Perception - Ne
External Connotative Judgement - Fe
Internal Denotative Perception - Si
Internal Denotative Judgement - Ti
Internal Connotative Perception - Ni
Internal Connotative Judgement - Fi
Last edited by daylen on Mon Oct 12, 2020 6:15 pm, edited 5 times in total.

daylen
Posts: 1678
Joined: Wed Dec 16, 2015 4:17 am

Re: Daylen's Instinctual Dump

Post by daylen »

To elaborate on this model:

Each agent has an event holarchy. References to more peripheral objects outside an event deep in this holarchy indicates that the agent is bringing the objects with them [internally] whereas reference to proximate objects indicates [external] collection. External functions attend to objects closely bundled together within the most immediate event whereas internal functions attend to objects scattered throughout a more encompassing event (that was perhaps not so scattered in the past). By constructing a normalized event holarchy for each agent this will become a measurable quality.

Objects referenced in conjunction tend to have certain syntactical patterns. Some of these patterns will be more causal/literal/denotative and others will be more correlative/analogous/connotative. This is perhaps the most challenging dichotomy to measure.

Agents with similar event holarchies will tend to reference certain objects in tandem as is (perception) yet may distinctly reference these same objects as they should be (judgement).

daylen
Posts: 1678
Joined: Wed Dec 16, 2015 4:17 am

Re: Daylen's Instinctual Dump

Post by daylen »

Merger of Economics and Evolution

I can see how both economics and evolutionary theory could be thought to solve different problems or study different objects, yet I tend to see these fields as two sides of the same coin. Let me explain.

Economics: (resources, interactors)
Evolution: (replicators, interactors)

Both deal, in my mind, with interactors. Interactors are generally thought to be organized in a hierarchy or holarchy if you prefer. The primary interactors for both are human organisms or agents. In economics, other emergent levels include markets, firms, economies, etc.; in evolution, these levels include populations, species, ecosystems, etc..

The kicker is that I also view the primary objective of both to be strikingly similar. One way to phrase this merged goal is to answer the question of "why is this interactor the way it is?" or simply "why this interactor?". Economics seeks answers in the form of economic efficiency (..because it is economical) and evolution phrases answers in the form of adaptation (..because it is adaptive). Both dealing with what levels of interactors are ultimately selective/causal of other levels.

The primary difference being that economics deals in resources that exist external to the interactors and evolution deals in replicators that exist internally. Resources are typically assumed to be finite and diminished through use (plenty -> scarcity) whereas replicators are assumed to be infinite insofar as they are passed from generation to generation.

Both are split into micro and macro branches yet are quite different in their pursuit of each. Economics roughly branches into (portfolio, economy) and evolution roughly branches into (genome, species)(*) due to their importance in selection and/or ease of modeling. The goal being to find causal links from interactors back to resources or replicators. Resources deplete due to the preferences of an agent or the interest rate of an economy; and replicators persist due to the fitness of an organism or species.

(*) Choosing genotype as opposed to gene here because a perfect replicator is not also an interactor and is therefore not a level of selection. Otherwise evolution becomes what Gould would call "bookkeeping" and Dawkins would call "selfish replication".

The tools each employs are quite similar due to their shared focus on agents/organisms which can choose to trade or sexually reproduce. Game theory helps formulate agent-agent interactions of competition or collaboration. Generally, communication is abstracted away yet both can potentially be aided through such a complication. Viewing preferences or instincts as innate without an understanding of their higher-order interactions between agents results in a dead-end for both. This is where a model of cognition/communication/computation can serve as the glue which brings them together resulting in a more satisfying explanation for "why this interactor?".
Last edited by daylen on Mon Oct 12, 2020 11:13 am, edited 1 time in total.

Mister Imperceptible
Posts: 1384
Joined: Fri Nov 10, 2017 4:18 pm

Re: Daylen's Instinctual Dump

Post by Mister Imperceptible »

daylen wrote:
Thu Oct 08, 2020 5:09 pm
The goal being to find causal links from interactors back to resources or replicators. Resources deplete due to the preferences of an agent or the interest rate of an economy; and replicators persist due to the fitness of an organism or species.
Or

The preferences of an agent or the interest rate of an economy are due to resource expansion/depletion

And/Or

The fitness of an organism or species (or lack thereof) is a random happenstance byproduct of the most persistent replicators

daylen
Posts: 1678
Joined: Wed Dec 16, 2015 4:17 am

Re: Daylen's Instinctual Dump

Post by daylen »

Resource expansion could be modeled by imperfect information of the territory. Random happenstance is boring.

..but yeah, sure you can invert if you want. :P

Mister Imperceptible
Posts: 1384
Joined: Fri Nov 10, 2017 4:18 pm

Re: Daylen's Instinctual Dump

Post by Mister Imperceptible »

My empirical skepticism does not prevent me from attempting to be a discerning economic or sexual agent 8-) :twisted:

daylen
Posts: 1678
Joined: Wed Dec 16, 2015 4:17 am

Re: Daylen's Instinctual Dump

Post by daylen »

(economics, evolution) = econvolution

.. as will now be how I refer to such thought processes from here forth ..

.... until a new synthesis is formed ....

daylen
Posts: 1678
Joined: Wed Dec 16, 2015 4:17 am

Re: Daylen's Instinctual Dump

Post by daylen »

Minor alteration/clarification of super-strategy space (i.e. Strategies):

Nodes correspond not to strategies but to interindependent activities with a conscious/unconscious boundary. The overall network is an atomic strategy that can be interpreted within a macro-event, and the individual nodes are activities that can be interpreted within a micro-event.

Leading to a merger of Eventual and Agency(*) with more tangible activities/events such as fairs, auctions, celebrations, hospitalizations, interrogations, and so forth.

Econvolution will link up by describing the topology of Strategies and may even provide local metrics. At least that is my guess for how it will best fit in for now.

(*) Communication parser, type modeling, bias detection, etc.

daylen
Posts: 1678
Joined: Wed Dec 16, 2015 4:17 am

Re: Daylen's Instinctual Dump

Post by daylen »

Workflow Update

A glimpse into my workflow for future reference/analysis.

Image

This first image is of the oldest drawings which are associated mostly with the construction of Eventual. Along with several topological/geometrical patterns corresponding to views in the web app representing event holarchies.

Image

This more recent image has a hard-coded pastel structure with gaps to be filled in with soft-coded pencil structure. Pretend the Eventual topology is rectangular to better fit in with the encoding below. Agency is associated to a distribution given that is one of the primary outputs.

Image

This last image details how Strategies can be linked to Econvolution geometrically/algebraically. Broken-line graphs can be composed onto these forms to represent an interpretation. For instance, the game between two agents in this example could a have broken line drawn connecting a resource point representing an implication that resource has to the game (generated within the micro-event). Furthermore an additional broken link to one of the strategy graphs may implicate that resource in that agent's strategy during this game.

R: rectangle holarchy
C: circle holarchy
P: point distribution
Last edited by daylen on Sat Oct 10, 2020 8:49 pm, edited 6 times in total.

daylen
Posts: 1678
Joined: Wed Dec 16, 2015 4:17 am

Re: Daylen's Instinctual Dump

Post by daylen »

Spacetime Topology

Question: What data structures or topologies are natural when representing events?

Constraints:
  • prefer simple html/css views
  • prefer simple json format
  • extensible
  • intuitively converted from external references
  • boundaries correlate with actual geographical/geological boundaries
Techniques:
  • The surface of Earth can be mapped onto a rectangle by removing two points at the poles then cutting along a line connecting them. This line is likely to be somewhere across the pacific ocean. Morph surface as needed to fit a particular view.
  • Use multiple morphs in order to cancel out view distortions and provide a more accurate aggregate Ni derivation.
  • Rectangles are easier to scale across views yet circles and points can be easily placed relative to a local rectangular frame.
  • Provide a search page as well as a page starting at the most macro event. On each sub-page provide corner links back to these pages.
  • Gradually transition into event drag interactivity for hands on scaling.
  • Grids that split up the continents is an example of a space-grid.
  • Timelines with overlapping event boxes. Vertically scaled according to impact.
  • Composing circles of varying radius to represent impact within regions of a space-grid.
  • Circular timelines with points as events. Hover over points to see detail and click to redirect.
  • Concentric circles representing points in time (rehoboam style).
  • Point distributions of resources or agents may be applied within certain events optionally.
  • Some event pages may provide an option panel that can be hidden.
Last edited by daylen on Mon Oct 12, 2020 9:21 pm, edited 2 times in total.

daylen
Posts: 1678
Joined: Wed Dec 16, 2015 4:17 am

Re: Daylen's Instinctual Dump

Post by daylen »

Focal Levels of Econvolution

Agency will require some constraints on objectification while parsing. Econvolution will help provide these constraints by delineating existence into a focal set of interactors which appear to dominate selection. Inversely, my own parsing system is initially responsible for constructing/discovering such focal points (i.e. the bootstrap system for Agency itself). Assuming there is an underlying power-law governing what levels are selective of other levels; or in other words, assuming that a minority of moderately closed systems interact with their own kind such that the interaction of all other systems is nearly determined from those interactions. Some determination is to be avoided by truncating spatial scale on both ends: systems smaller than cells or larger than cities are to be avoided except when referenced in narration. Furthermore, earth is to be the only planet considered. Each of the focal levels of econvolution can be described in terms of intermediate levels or solely in terms of a finite set of replicators and resources (plus some glue/logic). All interactors have both differentiation and integration mechanisms. As a side note, interactors are holons or simultaneous whole/parts.

Cell:
Overlaps with the organism level for unicellular organisms. Differentiated by cell walls and integrated through junctions, molecules, inflammation, and so forth. All cells carry, protect, and read from the highly stable replicators which are contained in every interactor on this list. Unicellular organisms often use their organelles as a multicellular organism might use their organs or organ systems (e.g. for movement, assimilation, or respiration). Multicellular organisms constrain cell level selection primarily by cutting off access to the germ-line leading to the termination of all somatic cell-lineages when an organism dies. Cancerous mutations thus have no way of being transmitted into the next generation. Cells depend on an influx of several different molecular resources to persist and divide. The majority of unicellular organisms replicate asexually yet some reproduce sexually.

Organism:
Multicellular organisms reproduce sexually and make use of recombination to help disperse adaptations throughout a population while simultaneously offering selective flexibility. Organisms are differentiated via an equivalent to skin and integrate via a large variety of mechanisms (e.g. audition, olfaction, vision, etc.). Organisms are the only level at which agency can exist. Although several organisms are capable of higher-order attention, it will be assumed for the sake of simplicity that all agents are humans and all humans are agents. Agent <==> Human. This relation forms the primary bridge between economics and evolution. Human agents are currently both highly selective of replicators and of resource depletion.

Species:
Species are considered the primary level of selection in macro-evolution. Genetic drift is generally weak relative to punctuated equilibrium, therefore species are thought to have definite beginnings and ends (i.e. speciations and extinctions). On geological time-scales it would even appear that species have sharper boundaries than organisms do. Species can bifurcate during what is called speciation. This could be due to a geographical barrier as in allopatric speciation or not as in sympatric speciation. Species tend to specialize within an ecosystem in order to minimize competition for resources with other species, but some species do become more generalized by utilizing a larger pool of resources (as do humans). Mutualism is also common between species and in some cases can lead to inter-species dependence. Small and isolated ecosystems may be especially vulnerable to cascading extinction.

Organization:
An organization must consist of at least one agent. In most cases, every agent will simultaneously be their own organization and be apart of at least one multi-agent organization. Leading to a similar overlap with organism-organization as was found with cell-organism. Within modern tax systems, organizations have fairly clear beginnings and ends. Organizations are differentiated by property/ownership through cryptography and/or contracts; and organizations integrate by trade through tit-for-tat or time-dependent derivatives. Mergers are possible at this level and are somewhat akin to consumption at the organism level except that some replicators are likely to be transferred to the dominant organization. Some combination of agent hierarchy and heterarchy is present in every organization. Organizations are oriented towards resource extraction, distribution, and combination. Resource combinations are traded as products (i.e. intermediate or final). Minimal focus is typically spent on the underlying replicators within an organization but within econvolution these replicators are present. Some examples of replicators at this level may include manufacturing techniques, technological patients, and social mirrors.

City:
Cities are somewhat more challenging to differentiate yet are generally presumed to be separated by a combination of zoning, governance, and population clustering. Cities primarily integrate though transit, pipelines, cables, and EM radiation. In terms of resources, cities are somewhat like giant pump and storage systems. Cities scale predictably while tending to replicate certain technologies and infrastructural patterns from other cities that were once of similar scale. Memes are an important form of replicator at this level given how population density correlates with communication frequency. Organizational operations are yet another form of replicator spreading between cities.

Fun Fact: (cell, organism, species, organization, city) forms a palindromic acronym (i.e. COSOC).
Last edited by daylen on Tue Oct 13, 2020 12:57 am, edited 15 times in total.

daylen
Posts: 1678
Joined: Wed Dec 16, 2015 4:17 am

Re: Daylen's Instinctual Dump

Post by daylen »

Recombination

Genetic:
When two organisms sexually reproduce an organism with a unique permutation of genes.

Technic:
When two or more technologies/techniques integrate to produce a unique permutation of techs.

Memetic:
When two or more memetic sources integrate to emit a unique permutation of memes.

Level of Influence

Cell: usually none for unicellular organisms
Organism and Species: genetic
Organization: primarily technic
City: primarily memetic

User avatar
Jean
Posts: 1288
Joined: Fri Dec 13, 2013 8:49 am
Location: Switzterland

Re: Daylen's Instinctual Dump

Post by Jean »

To diferentiate cities, you could use the inverse of the population density in the same way that we use altitude to differentiate watershed.

7Wannabe5
Posts: 6299
Joined: Fri Oct 18, 2013 9:03 am

Re: Daylen's Instinctual Dump

Post by 7Wannabe5 »

Geoffrey West’s “Scale” might be relevant. He researched factors for businesses and cities.

I am currently reading “The End of Average: How we succeed in a World that Values Sameness”, which is relevant to the science/math of personality typing. The gist of the author’s argument, as far as I am ferreting it out
thus far, is that if you are considering a property with single dimension such as “height”, comparison to the average cam be useful, but when multiple not-necessarily-well/correlated properties are measured and combined into a property such as “size” then too much information is lost for the average of this property to be useful reference.

Anyways, I know you are better at math than me, so maybe you could figure out the answer to the question of whether or not , for instance, two different individuals who were typed as ENTP, might have very few answers in common on questionnaire? IOW, might it be the case that, for instance, one “extrovert“ might be analogous to “large” human who is average height with very big belly and head circumference while another “extrovert” might be analogous to “large” human who is very tall with narrow shoulders and large hips?

sid3
Posts: 32
Joined: Sat Jan 25, 2020 1:40 pm

Re: Daylen's Instinctual Dump

Post by sid3 »

I would like to see the population of a city as a function of the area of a specific section that could be user selected, space-time location, and population density. Possibly this could lead to allopatric and sympatric speciation models for people and suburbs in cities. I am also curious to know where difficulties arise as theory is converted to code as well as general progress.

daylen
Posts: 1678
Joined: Wed Dec 16, 2015 4:17 am

Re: Daylen's Instinctual Dump

Post by daylen »

@Jean
Thanks for the tip. I'll have to keep that in mind.

@Sid
As far as I can tell, cities spawn and grow in fairly predictable patterns. As such the actual degrees of freedom are quite small so cross-fertilization of tools such as speciation to this level is not likely to be that interesting. Nevertheless there is enough variation here that warrants a closer look, especially for understanding migrations.

At some point in the near future I plan to share all my code on github where it will be periodically updated. I will post here when I decide to do that. For now I am working in four different directories (Agency, Eventual, Econvolution, Strategies) with minimal web templates and a few other miscellaneous files. Nothing exciting yet. I figure a more elaborate version of the post above on focal levels can be converted into an informative site for econvolution (the easiest site to build thus far). Eventual will probably be next in line for building a bare-bones site strictly focused on historical events for now. Agency and Strategies both have a lot of back-end work to be done and are therefore the most challenging corners of the square. Overall, I do not really know what I am building yet, but I figure it will be a multi-year project and will likely influence my future lives (i.e. if each life is around 5-10 years).

@7w5

"Scale" flickered into my mind a couple times while writing that post. I may want to give it a reread one of these days.

Averages, or in general, statistics always erode information from the underlying data-set. This is typically considered a good thing given that no human wants to interpret every single data point, so instead, statistics are used so that the human can interpret just those numbers (averages, variation, skew, median, mode, etc.). A data-set can be split categorically (perhaps with some overlap) then statistically analyzed within each category to output independent statistics on each category.

When typing you can split up the data/answers into E/I, N/S, T/F, J/P (4-dimensional) as most tests tend to do. In which case, multiple ENTP's could answer all the E/I questions differently so long as the ratio of E-answers divided by I-answers is greater than one(*). The model considered above for agency would measure internal/external, dennotative/connotative, and judgement/perception for each statement and assign one of the eight functions then aggregate all those assignment-statement pairs into a distribution to be interpreted and/or used as input to another algorithm.

Either way there is always going to be a trade-off between retaining information and eroding it for the sake of interpretation. Just depends on how complex/simple we want to see humans as (which likely varies with time/context/application).

(*) Answers could also be weighted.

7Wannabe5
Posts: 6299
Joined: Fri Oct 18, 2013 9:03 am

Re: Daylen's Instinctual Dump

Post by 7Wannabe5 »

Gotcha. The author suggested that education and self-improvement were contexts in which greater individualism should be promoted.

Post Reply