Artificial Intelligence

Daheane, Lau, and Kouider (2017) suggest that AI distinguish between two uses of the term "consciousness" that describe two different types computations that humans do. Match the kind of consciousness with the best description.

Being aware of something having some information in mind in order to act on it. - C1 (The first kind of "consciousness") Reflexive and self-referential - like self-awareness or self-monitoring or introspection and subjective experience. - C2 (The second kind of "conscousness")

Match the particular cognitive biases with the type of bias they best exemplifies.

Confirmation Anchoring Availability - Decisional Authority Conformity Association - Social Bizarreness Egocentric Fading - Memory

Which statement best describes "the hard problem of AI"?

Consciousness and self awareness

Match the description with the name of the theory of consciousness from the field of Neuroscience

Consciousness is something we do. It is active cause-effect relationships in physical systems rather than passive reception like a movie on a screen. - Integrated Information theory Consciousness is analogous to memory. The streams of information from memory in the brain is consciousness. Consciousness is like present memory. - Global Work Space theory

Match the type of AI with the best description

Focuses on cognitive representation - Analytical AI Focuses on cognitive and emotional intelligence - Human-inspired AI Focuses on the broadest range including cognitive and emotional and social intelligence - Humanized AI

Match the description of the method to optimize for non-discrimination to the name that relates to it.

Irrelevant to the outcome - Demographic parity Conditional on the outcome - Equalized odds Conditional on the predictor - Well-calibrated system

Match the description with the best category or label of the type of machine learning algorithm

Maps an input to a known output for a data set - Supervised algorithms Data is modeled according to inherent clusters or associations - Unsupervised algorithms Some data is labeled giving descriptive and predictive outcomes - Semi-supervised algorithms Uses positive and negative reward signals - Reinforcement algorithms

Match the term with the best description of it.

The ability to reflect on and examine one's own thinking and biases. - Reflexivity A process that calls itself or a similar process. When something is defined in terms of itself. - Recursion