chapter 1

Question 1

singularity

Answer 1

a time in the near future when computers will become smarter than humans

empowered by its ability to improve itself and learn on its own, will quickly reach, and then exceed, human-level intelligence

Question 2

Usual worries about AI

Answer 2

that increasingly sophisticated AI will replace humans in some jobs, that AI applied to big data sets could subvert privacy and enable subtle discrimination, and that ill-understood AI systems allowed to make autonomous decisions have the potential to cause havoc

Question 3

Hofstadter’s worries about AI

Answer 3

terrified that intelligence, creativity, emotions, and maybe even consciousness itself would be too easy to produce—that what he valued most in humanity would end up being nothing more than a “bag of tricks,” that a superficial set of brute-force algorithms could explain the human spirit.

Question 4

McCarthy

Answer 4

persuaded Minsky, Shannon, and Rochester to help him organize “a 2 month, 10 man study of artificial intelligence

The proposed study was based on “the conjecture that every aspect of learning or any other feature of intelligence can be in principle so precisely described that a machine can be made to simulate it.”

Question 5

obstacles

Answer 5

1. The Rockefeller Foundation came through with only half the requested amount of funding.
2. It turned out to be harder than McCarthy had thought to persuade the participants to actually come and then stay/agree on anything.
3. not a lot of coherence

(funding, participants, coherence)

Question 6

suitcase word

Answer 6

terms that are packed like a suitcase with a jumble of different meanings

> like intelligence, thinking, cognition, consciousness, and emotion.

Question 7

AI focused on two efforts

Answer 7

1. Scientific:
   investigate mechanism of natural (biological) intelligence by trying to embed it in computers
2. Practical:
   create computer programs that do tasks as well or better than humans
   (no worries abt if programs actually think like we think)

Question 8

An anarchy of methods

different opinions about the correct approach to take to develop AI.

Answer 8

1. mathematical logic and deductive reasoning as the language of rational thought
2. Inductive methods: extract statistics from data and use probabilities to deal with uncertainty
3. Biology and psychology: create brain like programs

current dominant paradigm = deep learning

Question 9

machine learning

Answer 9

a subfield of AI in which machines “learn” from data or from their own “experiences.”

AI > machine learning > deep learning

Question 10

AI

Answer 10

field that includes broad set of approaches, with the goal of creating machines with intelligence

Question 11

symbolic AI

Answer 11

A symbolic AI program’s knowledge consists of

1. words or phrases (the “symbols”), typically understandable to a human
2. rules by which the program can combine and process these symbols in order to perform its assigned task.

Question 12

General Problem Solver

Answer 12

 Simon and Newell designed their program to mimic what they believed were the students’ thought processes.

 At each step in its procedure, GPS attempts to change its current state to make it more similar to the desired state.

Question 13

GPS

what are operators, rules, and arguments

Answer 13

operators (in the form of subprograms) can transform the current state into a new state

rules encode the constraints of the task.

Arguments are the words inside the parentheses

Question 14

Advocates of the symbolic approach to AI argued that to attain intelligence in computers …

Answer 14

it would not be necessary to build programs that mimic the brain.

Instead, general intelligence can be captured entirely by the right kind of symbol-processing program.

Question 15

most notable form of symbolic AI

Answer 15

expert systems

in expert systems, human experts devised rules for computer programs to use in tasks

 such as medical diagnosis and legal decision-making

Question 16

Symbolic AI was originally inspired by …

Answer 16

1. mathematical logic
2. the way people described their conscious thought processes

Question 17

subsymbolic AI took inspiration from…

Answer 17

neuroscience:
sought to capture the sometimes-unconscious thought processes underlying what some have called fast perception

> such as recognizing faces or identifying spoken words

Question 18

a subsymbolic program is…

Answer 18

essentially a stack of equations—a thicket of often hard-to-interpret operations on numbers

such systems are designed to learn from data how to perform a task

Question 19

perceptron

Answer 19

invented by Frank Rosenblatt.

inspired by the way in which neurons process information.

Question 20

perceptron neuron analogy

Answer 20

the perceptron adds up its inputs

if the resulting sum is equal to or greater than the perceptron’s threshold, the perceptron outputs the value 1 (it “fires”); otherwise it outputs the value 0 (it “does not fire”).

A perceptron’s threshold is simply a number set by the programmer

Question 21

To simulate the different strengths of connections to a neuron

Answer 21

a numerical weight is assigned to each of a perceptron’s inputs; each input is multiplied by its weight before being added to the sum.

Question 22

how do perceptrons learn

Answer 22

supervised learning

 perceptrons should learn via conditioning
 should be trained on examples: it should be rewarded when it fires correctly and punished when it errs.

Question 23

supervised learning

Answer 23

the learning system is given an example

it produces an output

it is then given a “supervision signal,” which tells how much the system’s output differs from the correct output.

The system then uses this signal to adjust its weights and threshold.

Question 24

training set

Answer 24

part of a set of labeled positive and negative examples to train the system with

Question 25

test set

Answer 25

The remainder is used to evaluate the system’s performance after it has been trained, to see how well it has learned to answer correctly in general

Question 26

perceptron-learning algorithm

Answer 26

algorithm by which a perceptron could be trained from examples to determine the weights and threshold that would produce correct answers. 1. the weights and threshold are set to random values between −1 and 1. 2. The first training example is given to the perceptron; each input is multiplied by its weight, all the results are summed and compared with the threshold, and outputs either 1 or 0. 3. the training process compares the perceptron’s output with the correct answer given by the human-provided label 4. If the perceptron is correct, the weights and threshold don’t change. if the perceptron is wrong, the weights and threshold are changed a little bit, making the perceptron’s sum on this training example closer to producing the right answer. 5. The whole process is repeated for the next training example. 6. After many repetitions on each training example, the system eventually ettles on a set of weights and a threshold that result in correct answers for all the training examples. At that point, we can evaluate the perceptron on the test examples to see how it performs on images it hasn’t been trained on.

Question 27

major difference between symbolic and subsymbolic AI

Answer 27

The fact that a perceptron’s “knowledge” consists of a set of numbers —namely, the weights and threshold it has learned—means that it is hard to uncover the rules the perceptron is using in performing its recognition task. The perceptron’s rules are not symbolic

Question 28

Minsky on perceptrons

Answer 28

1. the types of problems a perceptron could solve perfectly were very limited 2. the perceptron-learning algorithm would not do well in scaling up to tasks requiring a large number of weights and thresholds. 3. if a perceptron is augmented by adding a “layer” of simulated neurons, the types of problems that the device can solve is, in principle, much broader. There is no reason to suppose that any of the perceptron's virtues carry over to the many-layered version.

Question 29

multilayer neural network

Answer 29

A perceptron with an added layer Such networks form the foundations of much of modern AI

Question 30

The two-part cycle in AI

Answer 30

1. AI spring: New ideas create a lot of optimism in the research community. Results of imminent AI breakthroughs are promised, and often hyped in the news media. Money pours in from government funders and venture capitalists for both academic research and commercial start-ups. 2. AI winter: The promised breakthroughs don’t occur, or are much less impressive than promised. Government funding and venture capital dry up. Start-up companies fold, and AI research slows.