Problems Bongard Problems in which arbitrarily small changes to examples can switch their sorting vs. Problems Bongard Problems in which small changes to examples never switch their sorting.

Left examples have the keyword "discontinuous" on the OEBP.

Right examples have the keyword "continuous" on the OEBP.

"Small change" is a relative judgment, but, in general, modifying only a few pixels should always be considered a small change.

In a "continuous" Bongard Problem, no small change should outright flip an example's sorting. It IS allowed for a small change to make an example sorted slightly more ambiguously.

"Continuous" implies "perfect" (left-BP913).

"Continuous" implies "pixelperfect" (left-BP947).

"Gap" (right-BP964) implies "continuous".

Continuous Bongard Problems tend to either be "fuzzy" (right-BP508), or otherwise not be "allsorted" (right-BP509) or have a "gap" (right-BP964).

See BP1140, which is about any (perhaps large) additions of detail instead of small changes.

BP1 is discontinuous because it's possible to change nothing slightly by adding a pixel to end up with something.

Left examples have the keyword "discontinuous" on the OEBP.

Right examples have the keyword "continuous" on the OEBP.

"Small change" is a relative judgment, but, in general, modifying only a few pixels should always be considered a small change.

In a "continuous" Bongard Problem, no small change should outright flip an example's sorting. It IS allowed for a small change to make an example sorted slightly more ambiguously.

Continuous Bongard Problems tend to either be "fuzzy" (right-BP508), or otherwise not be "allsorted" (right-BP509) or have a "gap" (right-BP964).

BP1, for example, is discontinuous because it's possible to change nothing slightly by adding a pixel to end up with something.

"Continuous" implies "perfect" (left-BP913).

"Continuous" implies "pixelperfect" (left-BP947).

"Gap" (right-BP964) implies "continuous".

See BP1140, which is about any (perhaps large) additions of detail instead of small changes.

"Continuous" implies "perfect" (left-BP913).

"Continuous" implies "pixelperfect" (left-BP947).

"Gap" (right-BP964) implies "continuous".

Left examples have the keyword "discontinuous" on the OEBP.

Right examples have the keyword "continuous" on the OEBP.

In general, modifying only a few pixels should always be considered a small change.

In a "continuous" Bongard Problem, no small change should outright flip an example's sorting. It IS allowed for a small change to make an example sorted slightly more ambiguously.

Continuous Bongard Problems tend to either be "fuzzy" (right-BP508), or otherwise not be "allsorted" (right-BP509) or have a "gap" (right-BP964).

BP1, for example, is discontinuous because it's possible to change nothing slightly by adding a pixel to end up with something.

Left examples have the keyword "discontinuous" on the OEBP.

Right examples have the keyword "continuous" on the OEBP.

In general, modifying only a few pixels should always be considered a small change.

In a "continuous" Bongard Problem, no small change should outright flip an example's sorting. It IS allowed for a small change to make an example sorted slightly more ambiguously.

Continuous Bongard Problems tend to either be "fuzzy" (right-BP508), otherwise not be "allsorted" (right-BP509), or have a "gap" (right-BP964).

BP1, for example, is discontinuous because it's possible to change nothing slightly by adding a pixel to end up with something.

Right examples have the keyword "continuous" on the OEBP.

In general, modifying only a few pixels should always be considered a small change.

In a "continuous" Bongard Problem, no small change should outright flip an example's sorting. It IS allowed for a small change to make an example sorted slightly more ambiguously.

Continuous Bongard Problems tend to either be "fuzzy" (right-BP508), otherwise not be "allsorted" (right-BP509), or have a "gap" (right-BP964).

BP1, for example, fails this because it's possible to change nothing slightly by adding a pixel to end up with something.

"Continuous" implies "perfect" (left-BP913).

"Continuous" implies "pixelperfect" (left-BP947).

"Continuous" implies "perfect" (left-BP913).

"Continuous" AND "perfect" (left-BP913) implies "pixelperfect" (left-BP947).

"Continuous" implies "imperfectionscanmatter" (left-BP913).

"Imperfectionscanmatter" (left-BP913) AND "continuous" implies "pixelperfect" (left-BP947).

"Continuous" (left-BP963) implies "imperfectionscanmatter".

"Imperfectionscanmatter" (left-BP913) AND "continuous" implies "pixelperfect" (left-BP947).

"Imperfectionscanmatter" (left-BP913) AND "continuous" implies "pixelperfect" (left-BP947).

"Continuous" (left-BP963) implies "imperfectionscanmatter".

Left examples have the keyword "continuous" on the OEBP.

In general, modifying only a few pixels should always be considered a small change.

In a "continuous" Bongard Problem, no small change should outright flip an example's sorting. It IS allowed for a small change to make an example sorted slightly more ambiguously.

Continuous Bongard Problems tend to either be "fuzzy" (right-BP508), otherwise not be "allsorted" (right-BP509), or have a "gap" (right-BP964).

BP1, for example, fails this because it's possible to change nothing slightly by adding a pixel to end up with something.

"Imperfectionscanmatter" (left-BP913) AND "continuous" implies "pixelperfect" (left-BP947).

Problems that are NOT "imperfectionscanmatter" (right-BP913), meaning little pixel-based changes should not change an example's sorting, are likely "continuous".

"Imperfectionscanmatter" (left-BP913) AND "continuous" implies "pixelperfect" (left-BP947).

Problems that are NOT "imperfectionscanmatter" (right-BP913), meaning little pixel-based changes should not change an example's sorting, are likely "continuous". However, other changes (more conceptual, besides just changing pixels) might also be intuitively small changes depending on the Bongard Problem.

"Imperfectionscanmatter" (left-BP913) AND "continuous" implies "pixelperfect" (left-BP947).

Problems that are NOT "imperfectionscanmatter" (right-BP913), meaning little pixel-based changes should not change an example's sorting, are likely "continuous". However, other changes (more conceptual, besides just changing pixels) might be considered small changes depending on the Bongard Problem.

Left examples have the keyword "continuous" on the OEBP.

A "small change" might be defined differently in an intuitive way depending on the Bongard Problem, but, in general, modifying only a few pixels should always be considered a small change.

In a "continuous" Bongard Problem, no small change should outright flip an example's sorting. It IS allowed for a small change to make an example sorted slightly more ambiguously.

Continuous Bongard Problems tend to either be "fuzzy" (right-BP508), otherwise not be "allsorted" (right-BP509), or have a "gap" (right-BP964).

BP1, for example, fails this because it's possible to change nothing slightly by adding a pixel to end up with something.

Left examples have the keyword "continuous" on the OEBP.

A "small change" would be intuitively defined differently depending on the Bongard Problem, but, in general, modifying only a few pixels should always be considered a small change.

No small change should outright flip an example's sorting. It IS allowed for a small change to make an example sorted slightly more ambiguously.

Continuous Bongard Problems tend to either be "fuzzy" (right-BP508), otherwise not be "allsorted" (right-BP509), or have a "gap" (right-BP964).

BP1, for example, fails this because it's possible to change nothing slightly by adding a pixel to end up with something.

Left examples have the keyword "continuous" on the OEBP.

A "small change" would be intuitively defined differently depending on the Bongard Problem, but, in general, modifying only a few pixels should always be considered a small change.

No small change should outright flip an example's sorting. It IS allowed for a small change to make an example sorted slightly more ambiguously.

Continuous Bongard Problems tend to either be "fuzzy" (right-BP508), otherwise not be "allsorted" (right-BP509), or have a "gap" (left-BP964).

BP1, for example, fails this because it's possible to change nothing slightly by adding a pixel to end up with something.

"Imperfectionscanmatter" (left-BP913) AND "continuous" implies "pixelperfect" (left-BP947).

Although NOT "imperfectionscanmatter" (right-BP913) means little pixel-based changes should not change an example's sorting, it does not imply "continuous", because other changes (more conceptual, besides just changing pixels) may be considered small changes (depending on the Bongard Problem).