12 Teorías del condicionamiento clásico

Name: 12 Teorías del condicionamiento clásico
Uploaded: 2020-11-05T20:18:43.000Z
Duration: 1 h 13 s

Introduction to Classical Conditioning

Overview of Classical Conditioning Theories

The video introduces classical conditioning, focusing on theories that explain the association between two independent stimuli.

It highlights Pavlov's traditional view, where the conditioned stimulus (CS) replaces the unconditioned stimulus (US) in triggering responses.

Mechanisms of Learning

The discussion emphasizes how associations are formed between CS and US, leading to conditioned responses.

A key question raised is about the mechanisms that establish these associations within an individual's mind.

Pavlov's Model and Its Assumptions

Key Assumptions of Pavlov's Model

Learning occurs when outcomes are unpredictable; surprise is essential for learning to take place.

Effective learning mechanisms aim to reduce prediction errors by aligning expectations with reality.

Associative Strength and Limitations

The model posits that maximum associative strength is limited by the total associative value of all present stimuli.

Conditioned inhibition is viewed as opposing conditioned acquisition, with values assigned accordingly (+1 for acquisition, -1 for inhibition).

Learning Mechanism Explained

Fixed Associativity Concept

Wagner’s model assumes that a stimulus's associability remains constant across trials, unaffected by learning experiences.

Psychological vs. Physical Intensity

Unlike other models, this one does not consider changes in associability based on psychological intensity but rather on physical intensity.

Understanding Prediction Error in Learning

Formula for Associative Strength Increase

Learning can be summarized through formulas indicating that increases in associative strength depend on the interaction between stimulus associativity and US intensity.

Discrepancy Between Expectations and Reality

The learning process hinges on discrepancies between what a subject can learn (lambda) and what they expect based on prior experiences with stimuli.

Understanding the Learning Model

Key Concepts of Alpha and Beta Parameters

The model utilizes alpha and beta parameters, which range from 0 (absence of intensity) to 1 (maximum intensity). This results in decimal values when applied in simulations.

For example, if alpha is 0.5 and beta is also 0.5, the product yields a value of 0.25, indicating that the associative strength increment for stimulus A will always be one-fourth.

Learning Expectations and Outcomes

The difference between what a subject receives versus what they expect determines their learning potential; specifically, it’s one-fourth of this difference that can be learned in any given situation.

The knowledge a subject has at any moment about stimulus A is equal to their previous knowledge plus what they have learned from the latest experience.

Application of Free Parameters

Using free parameters allows the model to explain negatively accelerated learning curves by ensuring subjects learn a constant proportion of what remains to be learned.

In examples provided, lambda represents maximum learning potential (1 when present, 0 when absent), maintaining consistency throughout various scenarios discussed.

First Trial Analysis

During the first trial where sound precedes an electric shock, the subject's expectation is zero since there’s no prior experience linking sound with shock.

The maximum discrepancy occurs as the subject experiences an electric shock (value = 1), while expecting nothing; thus, they learn significantly from this initial encounter.

Incremental Learning Process

In trial one, with alpha at 0.5 and beta at approximately 1.05, the resulting increment in associative strength for stimulus A becomes 0.25 due to multiplying by expected outcomes.

After trial one concludes, what the subject knows equals their previous knowledge plus new learning (resulting in an increase).

Progression into Subsequent Trials

By trial two, expectations shift as subjects anticipate receiving some level of shock based on prior experiences; hence discrepancies decrease.

Consequently, increments in associative strength become smaller but still contribute positively to overall knowledge accumulation over trials.

Explanation of Negatively Accelerated Learning Curves

The model explains negatively accelerated learning because subjects consistently learn less as they approach maximum understanding—incremental gains diminish over time while overall knowledge increases steadily.

Eventually, associative strength reaches its peak when it equals lambda (1), indicating no further surprises or learning opportunities remain for the subject.

Understanding Associative Learning through the Example of Maguey Worms

The Initial Phase of Conditioning

The ingestion of maguey worms has led to discomfort, establishing a strong associative force between the worms and the feeling of malaise.

In the second phase, there is an expectation of discomfort when consuming maguey worms, which sets up a conditioning scenario involving both maguey worms and chapulines (grasshoppers).

The Role of Expectation in Learning

Surprise and learning are influenced by the difference between what is experienced (discomfort from worms) and what is expected (no discomfort from chapulines).

If both stimuli (worms and chapulines) yield similar outcomes, no new learning occurs; thus, if one expects malaise from worms but not from chapulines, no association with chapulines develops.

Implications for Associative Strength

This model illustrates that prior expectations can overshadow potential predictors; if malaise is anticipated from one stimulus (worms), other stimuli may be disregarded.

The associative strength during initial exposure to worms was high (1), while it was zero for chapulines. Consequently, no learning about chapulines occurs in this context.

Mechanisms Behind Conditioned Inhibition

The model explains conditioned inhibition: if a subject anticipates negative outcomes due to one stimulus, other stimuli present may lose their predictive value.

Parameters within this model can be ignored as they only affect learning speed without altering what is learned.

Practical Application of the Model

Applying this model intuitively allows for understanding how experiences with different stimuli interact over time.

Mixing trials where maguey worms are paired with tequila leads to increased associative strength for worms when they cause malaise but decreases it when they do not.

Dynamics of Associative Strength Over Repeated Trials

As trials continue with mixed pairings, the associative strength for maguey worms increases while that for tequila decreases until reaching maximum or minimum values respectively.

Eventually, the goal is to balance these associations so that total associative strength equals zero when neither stimulus causes malaise.

Limitations of the Model

For effective learning to occur without oscillation in associative strength values, specific conditions must be met regarding expected outcomes.

Despite its predictive power and application in artificial intelligence contexts, this model has limitations such as assuming fixed intensity values for conditioned stimuli.

By structuring these notes around key concepts discussed in the transcript while providing timestamps linked directly to relevant sections, readers can easily navigate through complex ideas related to associative learning.

Understanding Latent Inhibition and Associative Learning

The Concept of Latent Inhibition

Latent inhibition refers to the phenomenon where repeated exposure to a stimulus without consequences makes it harder for an individual to later associate that stimulus with a significant outcome, such as food.

Theories of Stimulus Sociability

According to latent inhibition, individuals take longer to learn that a previously neutral sound predicts food compared to those who have not experienced the sound alone.

The Macintosh theory posits that the sociability of a cue increases when it is a good predictor of outcomes and decreases when it is less predictive than other cues.

Contrasting Perspectives on Attention

This theory suggests we pay more attention to stimuli that are reliable predictors of their consequences, which seems reasonable.

Conversely, Pierce and Hall propose that attention decreases for strong predictors and increases for uncertain outcomes, suggesting we focus more on ambiguous stimuli.

Hybrid Models in Associative Learning

Recent theories combine elements from both perspectives, indicating initial attention towards strong predictors which may diminish as learning becomes automatic.

This hybrid model illustrates how subjects initially focus on good predictors but eventually automate responses as they become familiar with the associations.

John Pierce's Contributions

John Pierce introduced a perspective emphasizing learning about configurations rather than individual stimuli, highlighting the complexity of associative learning.

He proposed that the net associative strength of a stimulus derives from its direct pairings with outcomes and its similarity to other associated stimuli.

Complexity in Associative Learning Theories

Pierce's theory indicates that responses depend on similarities between stimuli associated with outcomes; learning halts when net associative strength reaches a threshold.

Overall, no single theory fully explains associative learning; each has strengths and weaknesses. The ongoing debate reflects the intricate nature of these theories.

Conclusion: Ongoing Debate in Learning Theories

There is no definitive answer regarding which associative learning theory prevails; all have limitations. Each explanation reveals different aspects while leaving others unaddressed.