Page 3 of 3 FirstFirst 123
Results 21 to 23 of 23

Thread: smart videogames

  1. #21
    Join Date
    Apr 2012
    Location
    Illinois
    Beans
    220

    Re: smart videogames

    Continued:

    So the level generator that utilizes directed evolution would run dozens of smaller simulations that would determine the overall level (and it might run one final glitch-testing simulation).
    The reverse-processing I discussed is a reversal of means and ends. The speaker's end (his choice of words) becomes the listener's means to his end (grasping the idea of the speaker). This level generator would generate ends and means to be tested against those ends, and then those that pass would probably undergo further testing or mutation and testing. Thus the level generator is also using a kind of reverse-processing, because it is generating causes from effects rather than effects from causes. It is "working backward", as a psychologist might say. Thus it can create examples of psychological or biological perspectives, examples that would truly be "puzzles" whose comprehension depends on human thought.

    Edit:
    I remembered this ArsTechnica video.
    How Gamers Killed Ultima Online's Virtual Ecology
    It's not the same idea, but it matters if you want the level to evolve in its inception and its continued self-perpetuation.

    Late Edit: Ethical Ambition of Self-Awareness:
    The assumption was that knowledge is power, that humans with knowledge will determine humanity's fate, and that self-awareness is a central kind of knowledge. Indeed, I have wondered how people can have "impulse control problems" if their physical movements are still voluntary, and whether self-awareness might pertain. However, our default might be to replace unawareness with optimism.
    Sometimes people are unrealistic optimists who engage in motivated cognition that minimizes threatening information rather than the threatening reality. Especially, they do if they are made to think they are powerless (Re: dissonance theory: "freely chosen"?). Moreover, illness can be interpreted as a threat to self-esteem (Re: Terror management theory?), which people try to maintain (social comparison theory). Thus, one ironic possibility is that our disagreements' potential to generate conflict is mitigated by our optimism, ... that a realistic society would have less disagreement but would also lack the cushioning of our unrealistic optimism. It might be very important that we understand motivated cognition. Our cognitive capacity probably followed after our basic drives, although we do have "acquired tastes" that show a possible bidirectional interaction rather than a unidirectional, subservient interaction. I'll try to address this, but didn't want to bump again.
    Last edited by haplorrhine; 3 Weeks Ago at 07:41 PM.

  2. #22
    Join Date
    Apr 2012
    Location
    Illinois
    Beans
    220

    Re: smart videogames

    With this new problem, I knew I would need time to find intersecting ideas. An important intersection might be "the trial" or "the experiment." I cannot see a difference except that "trials" are normatively described as "succeeding" or "failing", although this same normative terminology is in the binomial probability theorem. As humans, our means are usually some application of a practiced (or shaped) behavior or a conceptually guided combination of various practiced behaviors, and our ends are usually some lasting state of affairs called a "success." However, ends can be means to other ends and, at the end of the day, one's ends can be anything, including more knowledge or more practice. If an experiment is a trial, its success state is the acquisition of new knowledge within the experimenter's own mind.

    In the case of evolution, each organism is like an experiment, in the broader context, or a trial, in the narrower context of that organism's own subjective success through its own trying/trialing... try-aling? Then again, "experiment" might still imply an active goal whereas "case" is a more passive term. In any case, the "trials" don't need to be conducted consciously. In the brain, a kind of individualized natural selection occurs through the process of neuronal death, but our biological pathways also utilize feedback mechanisms that establish a sort of trial. Cellular metabolism and cell signaling consist of many intertwined cause-and-effect interactions that are called biological pathways, and negative feedback isn't entirely unlike the perceptual feedback utilized by a thinking/perceiving organism. If an axis like the HPA or the RAAS is working properly, a downstream effect will trigger a down-regulation of some precursor to prevent a run-a-way process. It's like the body's saying, "It worked, you can stop now." Thus, in this sense, even non-conscious organisms can be said to have "goals" that succeed or fail or that are aligned or misaligned with a broader goal. This broader goal would be homeostatic maintenance of the organism itself, but, in the specifically human case,: sometimes we humans create self-fulfilling prophecies, perceiving our subjective success as a perceptible outcome that is actually objectively a failure in the sense that the intermediary result doesn't actually bring one closer to the ultimate goal. Social psychologists talk about the fundamental attribution error or self-fulfilling prophecies, but I think these are describing the very simple cognitive process of determining causation, determining the existence or non-existence of and the nature of the causative relationship between a person and a behavior or the perceptible result of some behavior.

    Cause and effect interactions are basic, but we can describe them in various ways. I think the sets in set theory can be ordered tuples or non-ordered Boolean sets, or non-Boolean fuzzy sets for that matter, because we humans use these basic structures to ascertain cause and effect. Non-order clusters can outline complex phenomena with unclear mechanisms, phenomena that effect us but are not affected by us. If a cross-case examination reveals a repeated ordering, the ordering can suggest the direction of causation, the "pathway." If many things depend on one thing, we might call this a "dependency," although many dependencies might be things that we uncover rather than things that we can witness in real-time. Not only can these describe, more generally, the hypotheses that proceed from the analysis of varying cases and then precede experimental testing, but, more specifically, they can describe the biological pathways of the organisms that find success through varying themselves. Moreover, these can also describe how we give eachother directions. Most man-made instructions are unidirectional sequences, but they could also be branched and conditional in the same way that a probability tree diagram can utilize conditional probabilities to describe a new, downstream probability distribution for each joint/fork in the tree.

    In that vein, I have a potential error to point out. I proposed "reverse-processing mechanisms" within the mind, but specific vs generalized forms of attention might be equally applicable. One thing that seems clear is that we humans needed a way to reflect on our own thoughts. If our activities proceeded unidirectionally toward completion, we would lose the bigger picture of what we had done. We don't lose ourselves in the moment, but somehow obtain some kind of feedback of the broader, overall sequence that was just executed. Maybe we reverse the same processes that were carrying us forward. Maybe we direct our attention in a certain way that recovers the content. In the case of attention, I think that attention can be specific or general. For example, if I see words, my attention is specialized enough to see a message, but not so specialized that I see the wrong message. In general, the refinement of attention tends toward greater specification/specialization, but this specialization can avoid errors at the expense of possibly discovering a better method. I think this specific/general principle could also be applied to the biological concept of functionality.

    Simulation Feasibility
    I was imagining a fully evolved gameworld that is always different, but this might too demanding for the machine. There might be some tricks: punctuated equilibrium can cause the organisms to fixedly establish certain combinations or traits until that organism is facing extinction again; and moreover the organisms can become even more fixed during gameplay so that each organism is a perfect copy of its kind. However, the issues are more.
    A gameworld that doesn't interact will be boring, and a gameworld that does will be very complex. If the player leaves a situation alone for a time, it might be useful if the game's memory can do the same. The game might use experimental or modeling techniques like twin studies, factor analysis, principle component analysis, etc., occasionally halting the simulation to ensure that no stone is left unturned and that the game has some kind of simplified mathematical model for what happens after some unexpected event is perpetrated by the players. In life, we more readily identify stable objects with stable properties. Functionality varies with time in predictable ways, and structure is basically fixed. However, anything can be de-stabilized, and it is this flux or transition that will probably be the challenge.

  3. #23
    Join Date
    Jul 2008
    Location
    The Left Coast of the USA
    Beans
    Hidden!
    Distro
    Kubuntu

    Re: smart videogames

    This is beginning to sound like a blog.

    The Ubuntu Forums are not an appropriate place for blogging.

    Closed.
    Please read The Forum Rules and The Forum Posting Guidelines
    My Blog
    A thing discovered and kept to oneself must be discovered time and again by others. A thing discovered and shared with others need be discovered only the once.
    This universe is crazy. I'm going back to my own.

Page 3 of 3 FirstFirst 123

Tags for this Thread

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •