PSYCH 10 Lecture Notes - Lecture 14: Jennifer Aniston, Big Data, Striatum
22 Sep 2019
School
Department
Course
Professor
Document Summary
Arti cial neural networks: mimic what we know about the brain, they are. Units occupy separate connected layers (similar to multiple brain regions in sensory pathways) Require learning to perform tasks e ciently (similar to learning in the brain and changes in synaptic connections: operant learning and reinforcement and so on, training. Give examples to the ann so that it builds connections: applications. Lower-level layers: general characteristics like light, lines, and dots. Higher-level layers: object as a whole tend to emerge. Like jennifer aniston cells, or the hierarchy in human brains: feedback projections. Signals not only go forward, but also has connection in the opposite direction. Might be useful for directing attention and hold memory template. Comparing performance of ann and that of human: challenges for ann. Learning from a small number of examples and less practice (requires lots of samples/trainings) Solving multiple tasks simultaneously (anns are highly specialized)
