Action
AI
Mix1
Mix2
Mix3
100

When two hands move together at high speed, the system naturally prefers this stable coordination pattern.

in phase movement

100

This early approach to AI assumed intelligence came symbolic and rule-based processing in the mind.

symbolic AI


100

This early conversational program simulated a psychotherapist using pattern matching.

ELIZA

100

This type of fallacy occurred when someone concluded “I am a unicorn because unicorns are awesome and I am awesome.”  

incorrect logical inference (fallacy of invalid reasoning)

100

Marr’s levels include computational, algorithmic, and this level that focuses on physical mechanisms.  

implementational level

200

This stage is labeled as the bottleneck because only one action can be chosen at a time

(eg. You can drive while talking until something unexpected happens, because at that moment this stage becomes overloaded. )

response selection

200

This classic AI system by Winograd demonstrated language understanding in a restricted “blocks world.”

SHRDLU

200

The infinite-regress argument that “a little person in your head makes decisions” is known as this problem.

homunculus problem

200

This long-term change strengthens synapses when they are repeatedly activated together.

long-term potentiation (LTP)

200

In hypothesis testing, this type of error is a “false positive.”

Type I error

300

This neural coding scheme represents movement direction using patterns of activity across many neurons.

population codes


300

This AI critique argues that symbol manipulation alone cannot produce genuine understanding, even if behavior looks intelligent.

Chinese Room

300

EEG/ERP has excellent temporal resolution but poor ______ resolution.

temporal

300

In the symbolic AI slides, this test evaluates whether a computer can convincingly imitate human conversation.

Turing Test

300

Tanenhaus et al. (1995) demonstrated this property of language comprehension (e.g. word presentation).

incremental

400

This concept proposes that movement sequences are stored as high-level action codes before they are executed.

motor programs

400

This criticism of symbolic AI highlights its difficulty handling real-time, uncertain environments.

brittleness (other synonyms)


400

These two visual pathways are often described as the “what” and the “how” systems.

ventral = what; dorsal = how


400

Putnam’s thought experiment distinguishing Earth’s “water” from Twin Earth’s “XYZ” shows this debate.

internalism vs. externalism 

400

Ballard, Hayhoe, Pook, & Rao (1997) showed that eye movements can replace internal memory, especially when this resource is not maxed out.

working memory

500

These are specific neurons fire when you perform an action, and when you watch someone else perform it.

mirror neurons


500

This approach to AI emphasizes perception–action coupling over abstract symbol manipulation.

embodied and embedded cognition

500

Dynamical systems describe cognition as movement toward these stable states.

attractors


500

This system uses cortex–basal ganglia loops to plan actions and cortex–cerebellum–spinal loops to adjust them during execution.

feedback loops (for planning and execution)

500

In neural networks, concepts such as “lion” or “car” are not stored in single nodes but in overlapping patterns across many units. This encoding enables robustness and generalization to novel examples.

distributed representations