3 ms·

huh that sounds exactly like machine learning overfitting vs generalization

• shadowgovt 3y ago
  Interesting analogy and it may not be wrong. The acquired immune system basically includes a feature where it can squirrel away some of the white blood cells that fought an infection in the past at the site of the infection. This is a subset of the cells that were activated by the chemicals that were found during the infection. But the immune system doesn't have a great way to distinguish the chemicals that actually cause illness from ancillary chemicals that may have been present during the infection as well. This is one of the root causes of allergies... you get sick with, say, rhinovirus at the same time that the pollen count is really high and your immune system defeats the rhinovirus but squirrels away both the white blood cells that were activated by the rhinovirus and the ones that were activated by the basically harmless chemicals on the surface of pollen grains. Then the next time allergy season comes around, your nasal passages and lungs fill up with pollen grains and the white blood cells close to the surface that were squirled away due to that old rhinovirus infection activate and trigger your immune system to respond to the pollen.
• closewith 3y ago
  Well, over-fitting is a term borrowed from biology (and generalisation from economics), so not surprising in grand terms. Maybe the real surprise for some is that not everything was invented in the valley.
  • make3 3y ago
    that's definitely not what I meant or implied; it makes sense that the immune system is a classifier of sorts