LESSSON 16 - Computational modelling in Psychology Flashcards
What are the two initial branches of the taxonomy for models mentioned in the text?
The two initial branches of the taxonomy for models are pre-quantitative models and quantitative models.
How are pre-quantitative models described, and what is a sophisticated form of these models?
Pre-quantitative models are verbal descriptions, and in a sophisticated form, they can be analyzed as box and arrows or flow charts.
What distinguishes computational models from other types of models mentioned?
Computational models not only describe phenomena but also depict the mechanisms that generate those phenomena. They are based on Artificial Neural Networks (ANN) and are employed to investigate specific questions in neuroscience.
Why is it mentioned that a model should account for a wide range of experimental data?
A model should be able to account for a wide range of experimental data to ensure that it explains macro studies and different phenomena, making it broader in scope.
What purpose do alternative models serve, and why are they important?
Alternative models serve the purpose of providing multiple theories or explanations for a phenomenon. They are important because they allow for systematic testing and comparison of different theories or models.
In what way does the text emphasize the need for explicitness in models compared to verbal theories?
The text emphasizes the need for models to be more explicit compared to verbal theories. Making theories more complete and specified is crucial for model development.
How does the text suggest comparing model behavior to human behavior?
The text suggests comparing model behavior to human behavior as a means of evaluating the model. This involves assessing how well the model aligns with empirical observations of human behavior.
What is highlighted as the most critical part for understanding the quality of a model?
Model evaluation is highlighted as the most critical part for understanding the quality of a model.
Why does the text caution that being good at exploring phenomena doesn’t guarantee a model is correct?
The text cautions that being good at exploring phenomena does not guarantee a model is correct because there may be multiple models representing the same results. Additional criteria are needed for model comparison.
What is mentioned as a preferable characteristic for a model that can be used for different tasks?
If a model can be used for different tasks, it is considered better, although more difficult. The versatility of a model across tasks is highlighted.
What is the major problem mentioned in psychology concerning models?
The major problem mentioned in psychology concerning models is the preference for simpler models when they are similar. The text notes that if models are similar, the simpler one should be preferred.
What characterizes localist models, and what are the two important features mentioned?
Localist models are characterized by having all levels of representation pre-defined. The two important features are that each neuron/node has a specific pre-specified role, and there is usually no learning algorithm.
What distinguishes PDP models from localist models?
PDP (Parallel Distributed Processing) models have a learning algorithm, which is a key feature that distinguishes them from localist models. These models learn from data, and there is a greater interest in investigating representation during learning.
