image - In Inverted index and pointing, I wrote that what is needed for pointing behavior is a mapping from observation to location, similar to the inverted index needed for search engines

  • Consider one-word sentence output as the next developmental step after pointing behavior
    • Primitive as an algorithm
      • Finds the position (document ID) from the observed words and selects a word in that document for output.
    • This would use the transposition of the mapping W from observation to location
    • AutoEncoder
  • This would eliminate the need for what was once a “place” to be discrete. - Location Continuity

This page is auto-translated from /nishio/一語文の出力 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.