Lessons:
3. Constructivist epistemology
Home » 6. Cognitive architecture » 63. Demonstration
Video 63-1 demonstrates the ECA architecture presented in Figure 62. It uses a similar experiment as the one used in the E-puck Experiment in Video 53, and in the Little Loop Experiment in Video 55.
Video 63-1: Demonstration of the enactive cognitive architecture in the Little Loop Problem.
Video 63-2 (below) demonstrates the ECA architecture when the agent has a visual system similar to that in Video 41. In this video, the two gray sharks implement ECA, whereas the blue fish are simply moving on a straight ligne to illustrate a dynamic environment. In addition to preys (blue fish), sharks can also see other salient objects: flowers, colored bricks of walls, and other sharks (but not dark green walls because they blend into the background).
Due to their interactional motivation, the sharks tend to move towards salient objects (as in Video 41, interactions consisting of getting closer to salient objects have a predefined positive valence).
Once they reach a salient object, the sharks can experience specific interactions: they can eat a fish, bump into a wall, or cuddle with another shark (flowers afford the move forward interaction as empty cells, the sharks just swim through them).
As explained in Page 62, the ontology mechanism associates interactions when they overlap in space. As a result, the sharks learn bundles of interactions that represent fish, brick walls, other sharks, and flowers. Once these bundles are learned, the behavior selection mechanism favors moving towards fish or other sharks because they afford interactions that have a positive valence (eating and cuddling).
Video 63-2: Demonstration of the enactive cognitive architecture in a continuous, open, and dynamic environment.
See public discussions about this page or start a new discussion by clicking on the Google+ Share button. Please type the #IDEALMOOC063 hashtag in your post:
