Nodes of Thought2019-01-21

• I had a feeling something was about to be born.
• But I didn’t know what it was.
• I felt that if I started working at this point, I would be swept away by my thoughts and fail to grasp the “something” that was so close at hand.
• So I drew it anyway on my iPad.
• I draw it, travel to the station, draw it again on the train, and then organize it on this page.
  • Nodes of Thought

This part of what I wrote on Facebook on my phone while traveling:.

• Associating while typing words is different from what comes out while drawing a picture.

• I was translating my book on how to learn into English so that I could share it with English speakers, but what I really need to do is to share it with Tayupinko (people).

  • I have a feeling something will be born. happen.
  • consult a dictionary On the actdictionary
  • fruitful metaphor # fruitful
    • If metaphor is highly developed within metaphorical space (lots of Networks between concepts), it can have the effect of providing a warp portal to other spaces and Facilitates bonding of distant objects.
  • Mapping non-computers analogy to concepts in the computer science domain has worked well

• Translation, or “Slow Reading,” is effective in assembling knowledge.

• Coupling is discovered after the fact.

  • Concepts tied to the figure and inefficient in text
    • Is there a case where the link is to metaphorical space?

• 

  • 2019-01-20 Facebook and may be related to

    • Metaphor that chaos is already dead when it is pushed into a one-dimensional symbol string, and that a seed may be born from the corpse.

    • Oh well, the problem is bigger because it can only represent one-dimensional permutations of symbols.

    • What meaning is expressed when two words are followed by one another?

    • The basic premise is that they acquire language while observing the same world, so they will be reading the same books and so on, and pointing.

  • Supplemental Explanation:
    • When humans acquire language, they do so while being in the same space as the person teaching the language and observing the same things (joint warning).
    • Learning occurs when pointing by eye gaze or pointing at precedes speech, and adults provide feedback on the output of that pointing.
    • If a machine acquires language, it will be while reading the same books as a human or something.
  • 

    • The “stacking operations” part of this needs to be supported.

      • There are no permalinks in the human experience.
      • oblivion , blurring , disassembly , decay , fermentation
        • Forget the unimportant part = abstraction.
      • What is important?
        • Random wiring is fine at first, just learn after the fact.

    • Associations are not similar searches.

    • 

    • One type of similarity is not the only metric similarity search

    • Relationships are diverse

    • One way is to create multiple metrics by dimensionality reduction from a metric.

    • Equivalent to variation in “what counts.”

    • A learning process that takes 10 years

    • Human learning takes more than 10 years.

    • What happens when a machine is designed to take 10 years to learn as well?

    • Must be interruptible

    • Lower layer calculator will be upgraded over the course of 10 years, need a movable design

  • I did some copying of a sutra-like action myself in the process of translating it into English.
    • 
      • [Translation is one means of slow reading.
    • Cases in which Japanese words expressing “important concepts” do not fall easily into English
      • Example “Thinking while writing”
      • Expressed by adding an additional page and writing more explanations
        • [Different symbol spaces have different symbol distributions.
        • Why not try to express yourself in a language with a smaller vocabulary?
          • Basic English - Wikipedia or
        • Grammery says, for example, that words like “important” tend to be “overused, so why don’t you use another word?” He says, “It tends to be overused, so why don’t you use a different word? However, if we want to convey the same concept to non-native readers, we should not unnecessarily use different words to express the same concept.
        • Isn’t it enough to actively use very good / good / bad / very bad?
    • What is the abstraction phase?
    • be unbroken

• - What is abstraction phase? - Connect keywords that appear in multiple locations - Find and unify similar expressions - Example: overall picture / whole picture

  • What is the output/verification?　→ Information to facilitate abstraction.
    • The first output in the development of intelligence is pointing at.
    • Learn from that feedback.
      • pointing at = indicating pointer
      • Example: Indicating a page with a book
    • This is not beneficial from the start.
      • Babies don’t start out doing useful work.
    • What specific feedback is needed?
      • Important Questions
      • But I tend to be biased toward examples.
    • What does that feedback look like?Clean Questions
    • Information to facilitate abstraction in the subject
    • What does abstraction look like?
    • A - Extract the important parts - handle give - Links to other information
    • What does Scrapbox give humans?
      • Bracketing = markup of key phrases that you think are important
        • Provide a means of referring to pages with detailed information in concise key phrases
          • Same functionality for page titles
          • Unlike mere permalinks on Twitter, Facebook, etc., a human can read the meaning from the title itself.
      • link
        • The connection is expressed in color.
        • Similar key phrases are suggested by ambiguous search when creating links

• 

• The associations you have when you are drawing are different from the associations you have when you are writing.

  • Why not?
    • The question is instructive.
      • Some beneficial, some not.
      • Should be low cost and negligible
  • What’s the difference?

• Painted, difficult to search for current status

  • Search is a means of [Human Augmentation
  • Difficult to convert all to language
  • OK if some vertices are searchable: sweet potato search.

• The associations you have when you are drawing are different from the associations you have when you are writing.

  • Why not?

    • Easier to express “links” and “relationships” in pictures than in text.
    • What is easy?
    • No time axis, no restrictions on one-dimensional representation
      • Even the process of reorganizing this section into bullet points is inconvenienced by the restrictions.
    • Similar to the case of different Facebook threads joining later.
    • Coupling is discovered after the fact.

  • Concepts tied to the figure and inefficient in text

  • 

    • This of course means [Human + computer = augmented human
    • associative connection Something like that.
    • Is it possible to be Link to Metaphor Space?

---

This page is auto-translated from /nishio/思考の結節点2019-01-21 using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I’m very happy to spread my thought to non-Japanese readers.