lesson 7 Flashcards
(42 cards)
Generality of Learning Laws (Initial Question)
Do the same rules of learning apply across different types of US (appetitive vs. aversive) and across different species?
Evolutionary Argument for One General Learning Process
Evolution is thrifty; if one learning system can solve multiple prediction problems, there’s no need for separate, costly systems.
Evolutionary Argument Against One General Learning Process
Argument: Different ecological challenges (e.g., finding food vs. avoiding predators) might require specialized learning systems for optimal outcomes, even if it’s more costly.
General Process Approach
he idea that there are some fundamental laws of learning that apply broadly across different species and situations.
Serial Reversal Task
An operant task (often in a Y-maze) where the rewarded choice alternates periodically. Tests the animal’s ability to learn the “rules of the game” (e.g., win-stay/lose-shift) beyond simple association.
Win-Stay/Lose-Shift Strategy
Rule: Continue with the same choice if it was rewarded; switch to the other choice if it was not rewarded. Optimal strategy for serial reversal tasks with unsignaled reversals.
Perseveration (in Reversal Learning)
The tendency to continue making a previously correct response even after the contingencies have reversed and it is no longer rewarded.
Learning Set Paradigm
A learning procedure similar to serial reversal but uses a new pair of stimuli for each discrimination problem. Tests the ability to learn the general rule that one stimulus in each pair is consistently correct.
Performance on Learning Set Tasks Across Species
Observation: Different species (and even different tasks within a species, e.g., visual vs. olfactory for rats) show varying levels of performance, suggesting potential differences in learning abilities or strategies.
Species-Specific Sensory Abilities and Learning
Influence: Differences in sensory organs (e.g., echolocation in bats, UV vision in bees) can make some species appear superior at tasks relying on those senses, but this doesn’t necessarily reflect general learning ability.
Cognitive Differences and Task Performance (Example: Ebbinghaus Illusion)
Observation: Different species can perceive the same stimuli differently (e.g., pigeons see the Ebbinghaus illusion reversed compared to humans), which can affect their performance on learning tasks involving those stimuli, independent of learning rules.
Salience of Stimuli Across Species
Influence: Different types of stimuli (e.g., odors for rats, visual cues for birds) have different levels of salience for different species, affecting how quickly they learn associations involving those stimuli.
Learning Strategies Across Species (Example: Pigeons and Memorization)
Observation: Some species (like pigeons) may rely on memorization of individual stimuli and their outcomes rather than learning general rules, especially when memory capacity is high relative to the complexity of the rule.
Unique Cognitive Skills (Example: Language in Humans)
Observation: Some species possess cognitive abilities (like human language) that appear to be unique and can fundamentally alter how they learn and solve problems.
Taste Aversion Learning (Introduction to the Lesson)
Significance (Early Research): Uncovered surprising properties that challenged the generality of existing learning laws and prompted a re-evaluation of learning principles.
Biological Function of Learning
Role: Learning serves to help organisms predict and prepare for biologically significant events, and evolutionary pressures shape what and how organisms learn.
Exaptation in Learning Mechanisms
Definition: A learning mechanism that evolved to solve one specific problem may be sufficiently effective for other problems as well.
Adaptive Specialization in Learning
Definition: The evolution of a learning mechanism specifically adapted for one problem, potentially at the cost of efficiency in solving other types of problems.
Relative Validity Effect (Generality)
bservation: Demonstrated across diverse species (rats, rabbits, pigeons, honeybees, humans in categorization/causal judgment), suggesting some fundamental learning principles are conserved despite independent brain evolution.
Backward Blocking (Human Learning)
Definition: A learning phenomenon in humans (where prior learning about a US-B association can interfere with learning about a CS-US association even when the CS is presented before the US and B). It was not initially predicted by classical conditioning models but is prompting extensions to those models.
Propositional Learning vs. Associative Learning
Debate: Some argue that conditioning involves learning propositions (declarative knowledge) about event relationships, while others emphasize associative links. The Perruchet effect (dissociation between CRs and verbal reports) suggests the latter might involve non-declarative processes.
Long-Term Goal of Animal Learning Research
Aim: To uncover general laws of learning by examining the scope and generality of principles observed in the lab.
Can you explain why food-poison pairings should be especially salient using our prediction lens? Jot down your thoughts and then click the button to see my thoughts.
Using our prediction lens can help explain the efficacy of taste aversion learning. Eating poison can be very dangerous – there are plenty of foods that will kill you if you eat them even once. Animals therefore don’t have the luxury of learning slowly what is good and what isn’t. Anything that makes you feel sick needs to be completely avoided from then on. It therefore makes sense for evolution to have made taste aversion learning very fast.
Taste Aversion Learning
A type of learning where an animal associates the taste of a food with subsequent illness, leading to avoidance of that food. Ecologically important for survival.