Lecture 1: Introduction to AI

Question 1

Hierarchy of artificial intelligence

Answer 1

AI > Machine learning > Deep learning

Question 2

Engineering Approach vs Cognitive Approach

Answer 2

Engineering:
- Tires to find optimal solutions.
- No matter how
- Act intelligentyly, think intelligently

Cognitive
- Tries to understand the process
- Tries to reproduce human beavior (even if wrong result)
- Act like humans, think like humans

Question 3

Weak vs Strong AI

Answer 3

Weak:
- Capabilities not intended to match or exceed the capabilities of humans
- usually a small application with a single purpose

Strong:
- Matches or exceeds human intelligence
- conciousness, self-awareness
- General purpose application of several tasks.

Question 4

What is intelligence

Answer 4

• Intellectual vs physical capabilities
• reflex vs planned/reasoned action
• awareness of existence

Question 5

What is the turing test?

Answer 5

• “If a human interrogator cannot tell the computer and human apart, then the computer is intelligent”`

Question 6

What year was the turing test created?

Answer 6

1950

Question 7

What are some of the capabilities required to pass the Turing Test?

Answer 7

• Natural Language Processing (NLP) to
  communicate
• Knowledge Representation to store knowledge
• Automated Reasoning to infer new knowledge
• Machine Learning

Question 8

Arguments in favor and against the Turing Test?

Answer 8

Pros:
- Objective notion of intelligence
- Prevents us from arguments about the computer’s consciousness
- Eliminates bias in favor of humans
Cons:
- Not reproducible
- Not constructive
- Machine intelligence designed w.r.t. humans
- test is anthropomorphic. It only tests if the subject resembles a human being.

Question 9

Limitations of the Turing Test?

Answer 9

 Superficiality: It focuses on deception, not intelligence.
 Lack of Scalability: Not practical for evaluating the vast
array of AI capabilities.
 Human Imitation vs. AI Innovation: Prioritizes mimicking
human behavior over unique AI strengths.
 Not Comprehensive: Fails to assess the wide range of
abilities that modern AI possesses.
 Advanced Interactions: AI now handles nuanced and
context-rich interactions, which the Turing Test doesn’t
fully capture.
 Ethical AI Considerations: Ensuring AI systems are fair,
unbiased, and safe.
 Automated Testing: Evaluate AI on a much larger scale and
more efficiently than the Turing Test

Question 10

What is being evaluated in modern chat bots?

Answer 10

 Natural Language Understanding (NLU): Understanding and processing human language (e.g., sentiment analysis, question answering).
 Creative Tasks: AI’s ability to generate creative content (e.g., writing poems, creating artwork).
 Decision-Making Skills: AI’s effectiveness in scenarios requiring complex decision-making (e.g., strategic game playing, business forecasting)

Question 11

HOW are chatbots are being evaluated today?

Answer 11

 Accuracy and Precision: Measure of correctness in AI’s outputs (e.g., percentage of correct answers in NLU tasks).
 Response Time: Speed at which AI provides responses, important in real-time applications.
 Robustness and Generalization: AI’s ability to handle unexpected inputs or scenarios

Question 12

What are some of the things we do with AI?

Answer 12

 Knowledge representation
(including formal logic)
 Search, especially
heuristic search
(puzzles, games)
 Planning
 Reasoning under
uncertainty, including
probabilistic reasoning
 Learning
 Agent architectures
 Robotics and perception
 Natural language
processin

Question 13

When was the early work in neural networks, purely theoretical?

Answer 13

1943

Question 14

Alan Turing Describes the Turing Test

Answer 14

1950

Question 15

Darmouth workshop

Answer 15

• Get-together of the big guys
• The term AI is born
  1956

Question 16

When was the “rise of AI” era?

Answer 16

1956 - 70s

Question 17

What are some notable things on “rise of AI”

Answer 17

• The era of Good old fashioned AI (GOFAI)
• Symbolic computation instead of numeric computation
• Development of AI-specific programming languages

Question 18

What lead to the first AI winter?

Answer 18

Innacurate prediction in regards to the capabilities of machiens in the future.
1966: the ALPAC report kills work in machine translation (and
NLP in general)
 People realized that scaling up from micro-worlds (toyworlds) to reality is not just a manner of faster machines and
larger memories…
 Minsky & Papert’s paper on the limits of perceptrons (cannot
learn just any function…) kills work in neural networks
 in 1971, the British government stops funding research in AI
due to no significant results
 it’s the first major AI Winter…

Question 19

What are Expert Systems?

Answer 19

1970-1980s
- Knowledge-intensive, rule-based techniques
- Decision support systems

Humans need to write the rules by hand

Question 20

Context of second AI Winter?

Answer 20

Mid-80s to mid-90s,
End of expert systems, too tedious to write rules by hand and too expensive to maintiain

Question 21

When is the rise of machien learning?

Answer 21

1980-2010
- More powerful CPUs - usable implementation of neural networks.
- Big Data
- Rules are now learned automatically!

Question 22

Deep Learning

Answer 22

2010 - Today
“Deep neural networks”
- Generic networks are used for many applications, such as image recogniiton, self-driving cars, etc.

Question 23

What is Eliza?

Answer 23

 Developed by Joseph Weizenbaum
 In the 1960’s
 Simulation of a dialogue with a psychotherapis

“Input sentences are analyzed on the basis of decomposition rules which are triggered by key words appearing in the input text. Responses are generated by reassembly rules associated with selected decomposition rules.”

Question 24

What are the 4 Steps of Eliza’s algorithm?

Answer 24

1. Identify Keywords:
2. Get Decomposition Patterns
3. Apply Reassembly Pattern
4. If there is no keywords in the input sentence, then generate a canned
   respons

Example:
 Keyword:
I am
 Decomposition pattern:
I am <whatever>
 Reassembly Pattern:
How long have you been <whatever>?</whatever></whatever>

 Input sentence:
“It seems that you hate me”.
 Keywords: “you” and “me“
 Decomposition pattern:
 <whatever1> you <whatever2> me
 It seems that you hate me
 Reassembly pattern
 What makes you think I hate you?</whatever2></whatever1>

Question 25

Eliza - Steps of “Identify Keywords”

Answer 25

1. Scan the input sentence to look for words in a dictionary of keywords.
2. Keywords are given a rank, and the highest ranking keyword is considered
   first

Question 26

Eliza - Steps of “Get decomposition patterns”

Answer 26

1. For each keyword, there is an associated list of decomposition patterns
2. The first decomposition pattern that matches the input sentence is
   selected.
3. If the decomposition rule does not match, then the next best ranking
   keyword is selected (and go back to step 2.1).

Question 27

Eliza - Steps of “Apply Reassembly pattern”

Answer 27

1. For each decomposition rule, there is a set of associated reassembly
   patterns to generate a response.
2. If a subsequent sentence selects the same decomposition pattern, the
   next reassembly pattern is used (so the output is not repetitive)

Question 28

What are the 6 Things we learned from Eliza?

Answer 28

• User experience Matters
• The Illusion of understanding
• Ethical concerns
• Simplicity Can be effective
  -Interdisciplinary Insignts
• Data Security and Privacy