INTRO TO COG SCI FINAL EXAM
QUESTIONS AND ANSWERS 100% PASS
What is the basic idea of the Physical Symbol System Hypothesis? - ✔✔the brain
contains word-like symbols
What does it mean to say that physical symbols in the PSSH are "multiply realizable"
within the brain? - ✔✔the same mental property, state, or event can be implemented by
different physical properties, states, or events
What is thinking according to the Physical Symbol System Hypothesis? - ✔✔the
manipulation and transformation of word-like symbols
What is a "search space" and what does it have to do with intelligence? - ✔✔A search
space starts with initial conditions and a list of permissible transformations (options).
The subject tries to reach the end state (goal) in a finite series of transformations/moves.
According to the PSSH, intelligence describes ease in navigating a search space.
How does Newell and Simon's General Problem Solver work? - ✔✔represents problems
as states, even breaking them down into subproblems; uses means-end analysis to
figure out difference between current state and goal state, actions req'd to reach goal st;
learns from previous searches
EMILY CHARLENE © 2025, ALL RIGHTS RESERVED 1
,What is a "heuristic search?" - ✔✔an option for navigating a search space -- the subject
considered all options immediately available and chooses the one that brings them
closest to their goal
How can we use river-crossing problems to illustrate a heuristic search? - ✔✔the best
options make meaningful progress to bring you closer to your goal
Who is "Shakey?" - ✔✔first mobile robot with the ability to perceive and reason about
its surroundings
Is Shakey a Physical Symbol System? - ✔✔yes
What can Shakey do? - ✔✔move around without running into things
How does the basic unit of a Connectionist network work? - ✔✔***each input and
output is connected to each hidden unit***
unit receives input, applies weight to input, sums up inputs, passes result thru
activation function, transmits output to other units in network
What do the weighted connections that obtain between Connectionist units do? -
✔✔decides whether to amplify/diminish signals
What is a feed-forward network? - ✔✔activation (signals) move in only one direction:
from input unit to output unit
What are hidden units? - ✔✔units that receive input from input units
EMILY CHARLENE © 2025, ALL RIGHTS RESERVED 2
, To what extent are Connectionist networks typically pre-programmed? - ✔✔weighted
connections are initially set randomly, then modified over a large number of trials using
a back propagation algorithm
To what extent do Connectionist networks learn? - ✔✔can adjust strengths of weighted
connections based on input data
What is the back propagation algorithm? - ✔✔trains AI networks by comparing target
values to output values, works backward through the system to calculate required
weight updates
What are some similarities and differences between a standard Connectionist network
and the human brain? - ✔✔similarities: individual units in network act like human
neurons
difference: sz. of systems; largest connectionist networks contain only 4000/5000 units
What's the distinction between a local optimal solution and global optimal solution? -
✔✔local optimal solution - best solution w/I specific region of solution space
global optimal solution - best solution w/I whole solution space, considers all possible
solutions
What are some programming techniques for achieving a global optimal solution? - ✔✔
What is the phenomenon of a Connectionist network overfitting data to its training set?
- ✔✔network learns data too well, captures details that may not apply to new data
EMILY CHARLENE © 2025, ALL RIGHTS RESERVED 3
QUESTIONS AND ANSWERS 100% PASS
What is the basic idea of the Physical Symbol System Hypothesis? - ✔✔the brain
contains word-like symbols
What does it mean to say that physical symbols in the PSSH are "multiply realizable"
within the brain? - ✔✔the same mental property, state, or event can be implemented by
different physical properties, states, or events
What is thinking according to the Physical Symbol System Hypothesis? - ✔✔the
manipulation and transformation of word-like symbols
What is a "search space" and what does it have to do with intelligence? - ✔✔A search
space starts with initial conditions and a list of permissible transformations (options).
The subject tries to reach the end state (goal) in a finite series of transformations/moves.
According to the PSSH, intelligence describes ease in navigating a search space.
How does Newell and Simon's General Problem Solver work? - ✔✔represents problems
as states, even breaking them down into subproblems; uses means-end analysis to
figure out difference between current state and goal state, actions req'd to reach goal st;
learns from previous searches
EMILY CHARLENE © 2025, ALL RIGHTS RESERVED 1
,What is a "heuristic search?" - ✔✔an option for navigating a search space -- the subject
considered all options immediately available and chooses the one that brings them
closest to their goal
How can we use river-crossing problems to illustrate a heuristic search? - ✔✔the best
options make meaningful progress to bring you closer to your goal
Who is "Shakey?" - ✔✔first mobile robot with the ability to perceive and reason about
its surroundings
Is Shakey a Physical Symbol System? - ✔✔yes
What can Shakey do? - ✔✔move around without running into things
How does the basic unit of a Connectionist network work? - ✔✔***each input and
output is connected to each hidden unit***
unit receives input, applies weight to input, sums up inputs, passes result thru
activation function, transmits output to other units in network
What do the weighted connections that obtain between Connectionist units do? -
✔✔decides whether to amplify/diminish signals
What is a feed-forward network? - ✔✔activation (signals) move in only one direction:
from input unit to output unit
What are hidden units? - ✔✔units that receive input from input units
EMILY CHARLENE © 2025, ALL RIGHTS RESERVED 2
, To what extent are Connectionist networks typically pre-programmed? - ✔✔weighted
connections are initially set randomly, then modified over a large number of trials using
a back propagation algorithm
To what extent do Connectionist networks learn? - ✔✔can adjust strengths of weighted
connections based on input data
What is the back propagation algorithm? - ✔✔trains AI networks by comparing target
values to output values, works backward through the system to calculate required
weight updates
What are some similarities and differences between a standard Connectionist network
and the human brain? - ✔✔similarities: individual units in network act like human
neurons
difference: sz. of systems; largest connectionist networks contain only 4000/5000 units
What's the distinction between a local optimal solution and global optimal solution? -
✔✔local optimal solution - best solution w/I specific region of solution space
global optimal solution - best solution w/I whole solution space, considers all possible
solutions
What are some programming techniques for achieving a global optimal solution? - ✔✔
What is the phenomenon of a Connectionist network overfitting data to its training set?
- ✔✔network learns data too well, captures details that may not apply to new data
EMILY CHARLENE © 2025, ALL RIGHTS RESERVED 3
