Tasks in Informatics on Pattern Recognition and Other Senses

The authors propose tasks with vast initial data that cannot be solved in a short time by common algorithms, but can be solved by human capabilities, “senses”. AI initially claimed it didn’t have such senses, but then, after being trained by the authors, it rejoiced and declared that it now understood how humans see. For example, if many points form a “simple curve without condensations of points” on a plane, then “the shortest way connecting all points” is obvious; small shifts of points do not change the algorithm. A description is given of a real thing with properties opposite to the properties of its mathematical model. AI could predict this property using the sense of reality. Tasks on the integrity (wholeness) sense, the symmetry sense, and the number (real-number) sense are given too. Therefore, tasks must be designed so that the AI cannot solve them.