Fundamentals of Learning and Motivation

south central schutzhund training

~ “Fundamentals of Learning and Motivation” ~
(Excerpts from the college text by Frank Logan (Prof. U of NM)

Introduction:
1. Learning is the association of a response with a stimulus.

2. Shaping is defined as successive approximations toward a final desired response.

3. Auto-shaping is acquired responses without any systematic training involved.

4. A Behavior Chain (a sequence of responses) should be learned backwards to be most effective.

5. The attractiveness of a stimulus decreases with overexposure

Positive Reinforcements: (reward training)
• All rewards are stimuli that reduce drives. Any event that leads to drive reduction functions as a reward.

• Frequently rewarding a poor approximation of the desired response may cause the imperfect response to become so well learned that the organism fails to make progress beyond it.

• When rewards are given on regular or fixed intervals, response tends to be confined to those intervals; when rewards are given on irregular or unpredictable intervals, response tends to be steady but slow; when rewards are always available, but several responses are required to get it, response, once it begins, is rapid.

• Performance generally increases as the reward increases. However there is a point of diminishing returns, at which more reward will result in only marginally greater response.

• The longer a reward is delayed, the poorer the performance of a response. Some performance can be maintained with a delayed reward, but at some point the delay is becomes a detriment to performance.

• If a reward is too infrequent, the organism may be reluctant to complete a response lest it go un-rewarded again. However, occasional reward produces a higher level of overall performance than a consistent reward.

Conditioning
• Conditioning is cumulative: each trial builds upon the strength already gained through previous experiences

• Distributed practice produces faster learning; learning trials massed to follow each other closely result in a greater number of experiences necessary for conditioning than when occurrences are distributed more widely over time. (ie cramming is ineffective and wasteful.)

• For rewards to be effective they must occur shortly after a response. Whenever a response is closely followed in time by a reward, the tendency for that response to occur in the future is increased.

• An organism does not have to be aware or recognize the relationship between the response and reward for the principle to work.

Negative Reinforcements: (aversion training)
• The concept of reinforcement is called negative only because it reduces or terminates stimulus conditions rather than presenting them.

• Negative reinforcement is as automatic as positive reinforcement.
However, outcomes will differ: a rat controlled by positive reinforcement (food) displays distinct eagerness when approached by the experimenter, while the rat controlled by negative reinforcement (shock termination) will make distinct efforts to avoid the whole experiment, ie, biting the experimenter.

An aversive event is unpleasant and one that the organism seeks to remove. It is characterized by the following;
• It elicits some unlearned motor responses
• It elicits an internal, implicit fear response
• It has not only emotional consequences but are stimuli in the sense that new (possibly unwanted) responses can be associated with them.

Aversion learning
Aversion learning has emotional consequences. If one uses adverse controls over another, it is important to be responsive to the other’s efforts to reduce it. Otherwise one may be teaching a persistent sense of helplessness. Even worse, this sense of helplessness carries over into areas that originally were not subject to aversion learning.

Aversion learning results in
• Escape Learning: the organism can do nothing to avoid the adverse event; it can only terminate it as quickly as possible.
• Avoidance Learning: the organism is given a warning signal, and if it responds in time, as required, it can prevent the adverse event.

Avoidance responses are extremely persistent. Even “enforced reality testing” ie, force to face a fear, is not all that effective. Once the force situation is removed, avoidance responses resume.

Response Persistence:
Forgetting: defined as the decrease in performance of a learned response since the time of learning (not due to any sensory, motor, or motivational changes in the organism)

An organism that learns two responses in a similar context will perform the first response more poorly than if the second response had not been learned.
(ie, a university marine biologist complains that every time he had to learn the name of a student, he forgot the name of a fish…)

New learning interferes with the ability to remember old learning. (Cramming is ineffective.)

Last Modified January 11, 2006

Visitors count started July 13, 2005