back to the article index
Page feed
Computer-Assisted Evolutionary Epistemology for the E-CAP 2009
Problematic
- Initial formulation
- What would the creative process (eventually applied to epistemology) require to make it impossible to automate? Even impossible to partly automate and delegate?
- Alternative formulations
- Can there be a creativity algorithm?
- What would it be? Evolutionary Algorithm? Need something else?
- Is evolutionary algorithm the equivalent of creativity?
- If life=cognition, does creativity=evolution?
- The creativity fractal, the geometry of innovation by the evolutionary process
Note : clarify how creativity fits with epistemology, hypothesis construction, inventing, innovating, ...
Requirements
Structure
Identifying the key building hypothesis H and the reached deduction D. Is each H required, sufficient, re-enforcing another, based on statistics (forming a PLN?), evidences, internal domain coherence(forming a domain like mathematics?), reliability? Why is D important (predictive range) and how can it be falsified by a test T. As long as T is false, how can D be used (usability)?
- sum-up
- context
- limited resources
- evolution!
- life!
- cognition
- emergence of a survival mechanism (a physical strategy)
- cognitive rat-race
- result of the encompassing situation (limited resources) and the competitive advantage (evolution and the survival of the fittest, not a rat race limited against peers but also against nature itself with the aging process)
- inter-individual cognitive rat-race (classical)
- intra-individual cognitive rat-race as the more we know, the more he have to manage and thus the better organization we need
- information has a due date
- cognition processes producing them too
- history of the movement
- evolution of the physical cognition tools
- internal (biological)
- external (technological)
- evolution of the formal tools
- intertwined progresses
- a required co-evolution
- no potential nature vs nurture debate
- software is delegated cognition
- binding cognition, creativity and epistemology
- proposition of solution
- algorithm applied!
- formalized structured hypothesis tree
- submitted to selective pressure
- added value
- support as formalization
- automation of tasks
- ability to generate faster
- experimental protocol
- take a creativity test and measure with and without the tools
- note : if the tool provide directly answers without requiring supervision/input it is not support to the user but directly "creativity" (on the basis that the tool did not have the solution before the test)
- conclusion
- critics
- potential consequences of a validated hypothesis
Questions
- Is life OR evolution the cognition process
- evolution seems to be an information acquisition process itself
- by natural selection
- also with Munz famous metaphor as organism=environment models (cf notes)
- ...less debating or explaining, more using!
- EE in the lab
- predicting results and then possible procedures
- since at that level what would be manipulated would be hypothesis and the encompassing process
- generate hypothesis and their related tests
- how is that different form
- importance of being able to isolate the core functional element
- intelligence = la capacite a produire un model (manipulation d'information) et a l'exploiter pour augmenter ses chances de survie, eventuellement grace a l'evolution a l'ameliorer
- discussion the recursive process for writing this very paper
- because it can apply to itself
- but as footnotes only (everything secondary should go there), eventually as an opening
- each paper being for an author a new affordance to his model (as an API)
- for his own more global cognitive scaffolding
- but also an API for others to be able to use it too
- thus publishing is making it "connectible" by others
- "the membrane idea it was inspired by the Protocells book and http://www.chemoton.com/eng9.html as a functional depiction"
- example of the argument of having a "beautiful" theory
- it makes it "portable" (you can think about it anywhere)
- it's this very need for portability that requires beauty and that pushes for the (usually difficult) synthesis process
- where are the relevant experiments?
- UNICOG : INSERM-CEA Cognitive Neuroimaging unit
- interviews of factory workers during the early stage of the industrial revolution who argued that mechanization wouldn't change their own work because what they were doing was too complex, any reference?
- "artificial intelligence" is not a premise
- link between agronomy/agriculture and evolutionary epistemology. It seems logical though because it's probably one of the initial goal of natural epistemology and it uses evolutionary algorithms (even if without computers).
- gradually breeding the species with the traits desired
- an argument from Cziko in his 1st or 2nd book
- why so many nature-like (natural epistemology, biomimicry) work/process?
- because we need compatible interfaces (see our waste/energy network, same principle from the cell up to the large cities)
- because nature established over time efficient ways
- because we might not be that creative
- is it possible to cut the problem in pieces?
- what can be studied later on?
- in 2009, is EE actually the conservative paradigm based on progresses in related fields?
- which one? thanks to which experiments?
- What are the invariants of creativity?
- What are case of creative behavior that are NOT using EE/EA?
- How can that be detected?
- What would be the process providing the same result but in the impossibility of relying on the mechanisms suspected to be at work?
- would the visual tree suggested in bibliography and finding an isolated node disprove the hypothesis?
- why conserving biodiversity (or at least putting it into databases before they disappear) is fundamental?
- because nature took a very long time finding those solutions through the process described here
- there are no resources to waste, they have already been "spent" for those solutions
- Is it better than the team of monkeys and their infinite amount of time typing Shakespeare?
- case/experiment for simultaneous discovery/innovation from independant labs
- no copying (since independant) BUT because of the environment being ready (available resources) and pushing for the discoveries
- possibility to:
- witness past cases
- recreate the conditions to try to prove it (wouldn't prove that the individual doesn't matter BUT that the environment is part of the conditions thus that selective pressure plays a role?)
- inspiration : (in the same way that life emerged in chemicals substrate not because of individual chance of smartness but because the conditions were there)
- CAEE, (Genetic Algorithm/Evolutionary Computing/Neural Networds) are very costly on computing resources
- FPGA, Cloud Computing, ... resources become widely available
- What is the Carbon-14 of ideas? just publications?
Leads
Creativity to remain or become independent from the existing (legal) system
Content moved to Drive.
- culturally : the tree of life and the tree of knowledge come from the same seed
- organisms are epistemology embodied, their evolution lead to the evolution of models
- thus having the same structure
To integrate
