Plants have at least 20 different senses (including electromagnetism), but do not have a centralised nervous system, thus it was assumed they did not 'learn' things in any individual sense and simply had to make the most of their environment and genetics.
Research by Monica Gagliano has found Pavlovian learning in plants, associating the direction of a fan with a light source.
Pavlovian learning in plants
Other research from Gagliano determined that plants do not just seek out water using their sense of moisture-gradients, but also use the sound of water to determine where to put down roots with the greatest chance of finding water.
(Thus, soundproof water/sewerage pipes would be less likely to break from root-infiltration)
Plant roots use sound to locate water
Finally, the long-term persistence of plant-learning and how it relates to the plant's environment.
Persistence of plant learning
The research is somewhat controversial, though I think the experiments were well designed and the research could be as applicable to the home garden as it could be to decentralised learning in AI systems.
"Appel and Cocroft found that recordings of the munching noises produced by caterpillars caused plants to flood their leaves with chemical defences designed to ward off attackers."
"...the ability of many plants to sample soil for nutrients and toxins via their roots, responding in ways that maximize growth and minimize risk; they do this by analyzing and responding to chemical gradients in a way the human body cannot.
As for sight, there is the well-known phenomenon of phototropism, in which leafy stems grow toward a light source to enhance photosynthesis. Add to this the less familiar ability of roots to detect light that penetrates soil, and then avoid it."