| Description |
-
Sequential, routine tasks are part of our every day life _�_ putting on clothes, driving to work, making coffee, making dinner, etc. They are the definition of ordinary and trivial goals. However, the amount of computation required by the brain to complete these actions is quite large, and consequently, the ability to simulate these actions on a computer has been limited up to now. One model, ... read moreContention Scheduling, is considered a successful computational model of routine task implementation, demonstrating ordinary action selection, as well as errors in implementation we observe in humans. However, the model has its own deficiencies with regards to learning and action correction. The purpose of this paper is to explore the use of Hebbian learning and bottom-up connections in modifying Contention Scheduling to resolve these issues. In order to do this, we implemented our own version of the original model, as well as versions including learning and action correction, and compared the results. Our results indicate that Hebbian learning complements the model well to simulate the cognition of procedural learning, and that bottom-up connections mitigate a large portion of the observed errors in the model.read less
|
|---|
This object is in collection
