Short Trip To Artificial Intelligence: Lecturer Adrian Groza

Short Trip to Artificial Intelligence
Lecturer Adrian Groza
Intelligent Systems Group

Department of Computer Science
Technical University of Cluj-Napoca
March 25, 2016

Short Trip Knowledge Representation and Reasoning
Warm-up
A Logician’s View of the World
Let’s start for the very beginning:

John loves Mary

John loves Mary
But let’s view it in a more graphical way:
Something important is missing


John loves Mary
And now we said too much!!!

Who said that John was a man?


Who said that Mary was wearing a skirt?

And what does a heart to do with love?

What we know is that:

There is something called John

There is something called Mary

There is something called Mary
There is something called Love going on between John
and Mary
Knowledge Representation and Reasoning

From cognitive psychology, human problem solving and learning are
processes that involve:
1 the representation and utilization of several types of knowledge
2 the combination of several reasoning methods: rule based reasoning,
causal reasoning, model based reasoning, contraints based reasoning

Outline
1 Short Trip
Semantic Web and Ontologies
Fuzzy expert systems
Reasoning with constraints
Knowledge acquisition
Case-based reasoning
Reasoning about knowledge
Model-based reasoning
2 Knowledge Representation and Reasoning
Reasoning
Logics
Knowledge engineering
Semantic Web Vision

”The Semantic Web is not a separate Web
but an extension of the current one, in which
information is given well-defined meaning,
better enabling computers and people to
work in cooperation.” Tim Barners Lee
Semantic Web
Social Web
Semantic Web Vision

”The Semantic Web is not a separate Web
but an extension of the current one, in which
information is given well-defined meaning,
better enabling computers and people to
work in cooperation.” Tim Barners Lee
Social Web
Semantic Web
Semantic Search
Semantic Search
Semantic Search
Semantic Search
Semantic Search
Rule-based systems
Definition (Expert system)

A computer system that emulates the decision making ability of
a human expert in a restricted domain.
The knowledge of the expert is represented as rules:

Example
If gross cost is greater or equals to 20.00 then add 10%
discount to the shopping cart
If the engine does not turn over and the lights do not come
on then the problem is battery or cables
If blood pressure is likely to be high then risk of heart
failure is high
a ⇒ b, b ⇒ c...
Fuzzy Knowledge
Thinking humanly: ”The cognitive modeling approach.”

Cognitive techniques used in engineering: soft computing :
neural networks, genetic algorithm, fuzzy logic
FuzzyLogic
←→
Fuzzy knowledge can be expressed in a very natural way using

linguistic variables.
Paradoxes?
The world is made up of opposites, for The world is filled with contradictions, that
example male versus female, hot versus almost everything contains some of its
cold, dry versus wet, active versus opposite, or in other words, that things can
passive. Everything has to be A or not-A, be A and not-A at the same time.
it can’t be both.
Law of noncontradiction: A ∧ ¬A 6= 0 Law of excluded middle: A ∨ ¬A 6= 1
Fuzzy Versus Probability

Stochastics and Fuzzy Logic complement each other
A person suffering from hepatitis shows in 60% of all cases a strong fever, in
45% of all cases yellowish colored skin.
Example
A bottle of liquid has a probability of 0.5 of being rat poison
and 0.5 of being pure water.A second bottle’s contents, in the
fuzzy set of liquids containing lots of rat poison, is 0.5. 50%
probability means 50% chance that the water is clean. 50%
fuzzy membership means that the water has poison.
Remark
1 Probability deals with uncertainty, fuzzy logic deals with ambiguity and
vagueness.
2 Probability is only valid for future/unknown events; fuzzy set
membership continues after the event.

Example
Remark
vagueness.

Example
Remark
vagueness.

Example
Remark
vagueness.
Constraints everywhere...
Simplification Combination Contradiction Redundancy

Combination Contradiction Redundancy
Simplification
Contradiction Redundancy
Combination
Simplification
Redundancy
Combination Contradiction
Simplification
Combination Contradiction Redundancy
Simplification
Datamining
”Everything about you is being tracked - get over it.”
Trends leading to dataflood

1 More data is generated:
Business entities:
banks, telecom,
business transactions,
Scientific Data:
astronomy, biology,
physics, ...
Web 2.0: wikis, blogs,
forums, ...
2 More data is captured
Datamining
Definition
Extraction of interesting knowledge (rules, regularities patterns,
constraints ) from data in large datasets.
Example (Cern experiment)

15 million gigabytes of data annually most of the data is
uninteresting since it contains already known physics.
all the information about the detector is programmed into
several thousand computers or computing cluster who is the
only guy around with the complete knowledge about the
whole experiment since combine the knowledge of all the
scientists.
data mining challenge: collision between atomic nucleus will
occur 40 million times per second; potential interesting
collisions will occur around 1000 times per second.
Case-based Reasoning
CBR uses specific past experiences to generate

knowledge
Form of problem solving by analogy in which a new
problem is solved by recognizing its similarity to a specific
known problem, then transferring the solution of the known
problem to the new one.
Example
Medicine: Find which diagnosis was made for
similar symptoms in the past, and adapt
treatment appropriately.
Law: Find what judgment was made for
similar cases in the past (’precedents’), and
adapt appropriately
Model-based Checking
Motivation for automatically fault diagnosis
Actuator faults caused airplane to crash Sensor faults caused rollercoaster to

The Intel Pentium bug (1994)
http://www.planecrashinfo.com stop
The Intel Pentium bug (1994), was the ”disaster” that got
model checking on the map in the hardware industry
We examine the applications of computational tree logic to
verify the correctness of computer systems.
Outline
1 Short Trip
Reasoning about knowledge
2 Knowledge Representation and Reasoning
Reasoning
Logics
Pyramid of knowledge
Wisdom ? Using knowledge in a b

way
Meta- danger ⇒ rule r1 is not applicable Rules about knowledge

knowledge
Knowledge r1 : if temperature ≤ 82 ⇒ start engine Rules about information
Information temperature = 82 Potential useful for know
ASCII rules
Data 74 65 6D 70 65 72 61 74 75 72 61 3D 38 32 Potentially useful inform
Rules: r’:group in 2 letters, r”:delete ≤ 30
Signal 1374656D702965726174757261003D383215 No aparent information

Intelectual origins
1 2000 years ago

the syllogism
2 17t h century Aristotel
the algebra of thoughts
human reasoning can be reduced to calculations
Leibniz
3 19t h century
boolean logic
Bool
Intelectual origins
1 2000 years ago

the syllogism Aristotel
2 17t h century
3 19t h century
boolean logic Leibniz
Bool
Intelectual origins
1 2000 years ago Aristotel
the syllogism
2 17t h century
Leibniz
3 19t h century
boolean logic
Bool
Intelectual origins
1 19t h century
analitycial engine (Babbage, 1791-1871)
the first programmer
Babbage
2 20t h century Imitationgame
Turing’s (1912-1954) ideas about thought and
computation McCarthy
McCarthy (1927) reasoning can be reduced to
calculations
Intelectual origins
1 19t h century
Babbage
Imitationgame
2 20t h century McCarthy

computation
calculations
Intelectual origins
1 19t h century
Babbage
Imitationgame
2 20t h century McCarthy

computation
calculations
Intelectual origins
1 19t h century
Babbage
Imitationgame
2 20t h century
computation
calculations McCarthy
Reasoning
Cognitive Processing
Reasoning
The Messy Room1
Mother goes into her teenage daughter’s bedroom.

Her instant impression is that it is a big mess. There
is stuff scattered everywhere. Mother’s impression is
that it is not characteristic of the girl to be like this.
What has happened?
1
Example taken from D. M. Gabbay, Sampling Logic and Argumentation
Networks: A Manifesto, 2010
Reasoning
The Messy Room1

What has happened?
Conjecture: The girl has boyfriend problems.
1
Reasoning
The Messy Room1

What has happened?
Conjecture: The girl has boyfriend problems.
Further Analysis: Mother noticed a collapsed shelf. Did the girl

smash it? Upon further observation, mother notices that the
pattern of chaos shows that a shelf has collapsed because of
excessive weight and scattered everything around, giving the
impression of a big mess. But, actually, it is not a mess, it does
make some (gravitational) sense.
1
Reasoning
The Messy Room
There are several modes of reasoning:

1
Remarks
Reasoning
The Messy Room

1 Neural nets type of reasoning: She recognises the mess
instantly, like we recognise a face.
2
Remarks
Reasoning
The Messy Room

2 Nonmonotonic deduction: Mother reasons from context
and her knowledge of her daughter that the girl is not
disorganised like this. She asks ”what happened?”.
3
Remarks
Reasoning
The Messy Room

3 Abduction: She offers a reasonable explanation that the
girl has boyfriend problems. This is common to that age.
4
Remarks
Reasoning
The Messy Room

4 Deduction: She then applies deduction and recognises
that the mess is due to gravity. This deduction is no longer
a neural net impression. It is a careful calculation.
Remarks
Reasoning
The Messy Room

Remarks
Item (4) can be a neural net impression: one who sees
many shelf collapsing mess cases recognises the pattern.
Reasoning
The Messy Room

Remarks
Item (4) can be a neural net impression: one who sees
many shelf collapsing mess cases recognises the pattern.
Item (2) could have been a Bayesian network.
Reasoning
Logical Inferences
Making inferences involves the derivation of

new facts from a set of true facts. Given a set of
premises known and a reasoning method
certain conclusions can be inferred to be also
true.
Reasoning
Deduction, Abduction, Induction
Deduction:
apply a general principle to infer some fact.
r1 : A ⇒ B
B is true.
A is true.
Reasoning
Deduction:
r1 : A ⇒ B
B is true.
A is true.
Abduction
: allows inffering A as an explanation of B
r1 : A ⇒ B
A is true.
B is true.
r2 : C ⇒ B → nondeterminism
Example (E. Ionescu’s play Rhinoceros)

All cats die.
Socrates is a cat.
Socrates is dead.
Reasoning

Deduction:
r1 : A ⇒ B
B is true.
A is true.
Abduction
r1 : A ⇒ B
A is true.
B is true.

All cats die.
Socrates is a cat.
Socrates is dead.
Induction: reasoning from particular facts to a general
conclusion
P(a) is true.
For all X , P(X ) is true.
P(b) is true.
Reasoning

Deduction:
r1 : A ⇒ B
B is true.
A is true.
Abduction
r1 : A ⇒ B
A is true.
B is true.

All cats die.
Socrates is a cat.
Socrates is dead.
conclusion
P(a) is true.
P(b) is true.
Reasoning

Deduction:
r1 : A ⇒ B
B is true.
A is true.
Abduction
r1 : A ⇒ B
A is true.
B is true.

All cats die.
Socrates is a cat.
Socrates is dead.
conclusion
P(a) is true.
P(b) is true.
Reasoning

Deduction:
r1 : A ⇒ B
B is true.
A is true.
Abduction
r1 : A ⇒ B
A is true.
B is true.

All cats die.
Socrates is a cat.
Socrates is dead.
conclusion
P(a) is true.
P(b) is true.
Logics
Logics for Computer Science
Expressive power 1 Boolean Logic

Able to represent the problem 2 Predicate Logic
3 First Order Logic
Correctness of entailment procedure 4 Second Order Logic
No false conclusion are drawn
5 Modal Logics
6 Temporal Logic
Completeness of entailment procedure
7 Epistemic Logic
8 Fuzzy Logic
All correct conclusions are drawn
9 Nonmonotonic Logics
10 Default Logic
Decidability of entailment problem
11 Defeasible Logic
There exists a terminating algorithm
12 Intuitionistic Logic
13 Linear Logic
Complexity 14 Description Logics
Resources needed for computing the solution 15 ...
Logics
FOL Warm-up
Everyone in this room is smart

Logics
FOL Warm-up
∀x In(x, F ) ⇒ Smart(x)
Logics
FOL Warm-up
Someone in the other room is smart
Logics
FOL Warm-up
∃x In(x, 45) ∧ Smart(x)

Logics
FOL Warm-up

Not all students take KBS
Logics
FOL Warm-up

∃x, Student(x) ⇒ ¬Take(KBS, x)

Logics
FOL Warm-up


Only one student failed KBS
Logics
FOL Warm-up


∃x, [Student(x)∧Fail(KBS, x)∧∀y, [¬(x = y )∧Student(y) ⇒ ¬Fail(KBS, y )]]

Logics
FOL Warm-up


There is a woman who likes all men who are not vegetarians
Logics
FOL Warm-up



There is a woman who likes all men who are not vegetarians
∃x Woman(x) ∀y Man(y ) ∧ ¬Vegetarian(y ) ⇒ Likes(x, y)

.
Logics
Warm-Up
There is a barber who shaves all men in town who do not

shave themselves.
Order of mixed quantifiers does matter

∀x∃y (Loves(x, y):
∃y∀x(Loves(x, y):
∀y∃x(Loves(x, y):
∃x∀y (Loves(x, y):
Logics
Warm-Up

shave themselves.
∃x, Barber (x)∧∀y , Man(y )∧¬Shaves(y, y ) ⇒ Shaves(x, y )

Logics
Warm-Up

shave themselves.
Politicians can fool all of the people some of the time, but
they can’t fool all of the people all of the time.

Logics
Warm-Up
shave themselves.
∀x, Politician(x) ⇒ (∀y ∃t Person(y) ⇒ Fools(x, y , t))
∧¬(∀y∀tPerson(y ) ⇒ Fool(x, y, t))

Logics
Warm-Up
shave themselves.

∀x∃y (Loves(x, y): everybody loves somebody
Logics
Warm-Up
shave themselves.

∃y∀x(Loves(x, y): there is someone who is loved by everyone
Logics
Warm-Up
shave themselves.

∀y∃x(Loves(x, y): everyone has someone who loves them
Logics
Warm-Up
shave themselves.

∀y∃x(Loves(x, y): everyone has someone who loves them
∃x∀y (Loves(x, y): there is someone who loves everyone
The Complexity of Knowledge Soup
Overgeneralizations. Birds fly.

Overgeneralizations. Birds fly. But what about penguins?

Overgeneralizations. Birds fly. But what about penguins? A

day-old chick?

day-old chick? A bird with a broken wing?

day-old chick? A bird with a broken wing? A stuffed bird?

day-old chick? A bird with a broken wing? A stuffed bird? A
sleeping bird?

sleeping bird? A bird in a cage?

Abnormal conditions.

Abnormal conditions. If you have a car, you can drive from
New York to Boston.

New York to Boston. But what if the battery is dead?

New York to Boston. But what if the battery is dead? Your
license has expired?

license has expired? There is a major snowstorm?

Conflicting defaults.

Conflicting defaults.Quakers are pacifists, and Republicans
are not.

Conflicting defaults.Quakers are pacifists, and Republicans
are not. But what about Richard Nixon, who was both a
Quaker and a Republican? Was he or was he not a
pacifist?
Unanticipated applications. The parts of the human body

are described in anatomy books.

are described in anatomy books. But is hair a part of the
body?

body? Hair implants?

body? Hair implants? A wig?

body? Hair implants? A wig? A wig made from a person’s
own hair?

own hair? Fingernails?

own hair? Fingernails? Plastic fingernail extender?

own hair? Fingernails? Plastic fingernail extender? A skin
graft?

graft? Artificial skin used for emergency patches?

graft? Artificial skin used for emergency patches? A
band-aid?

band-aid? A bone implant?

band-aid? A bone implant? An artificial implant in a bone?

A heart transplant?

A heart transplant? An artificial heart?

A heart transplant? An artificial heart? An artificial leg?

Teeth?

Teeth? Fillings in the teeth?

Teeth? Fillings in the teeth? A porcelain crown?

Teeth? Fillings in the teeth? A porcelain crown? False
teeth?

teeth? A corneal transplant?

teeth? A corneal transplant? Contact lenses?

teeth? A corneal transplant? Contact lenses? Eyeglasses?

A tattoo?

A tattoo? Make-up?

A tattoo? Make-up? Clothes?
Ethical and Legal Considerations
Using expert systems implies that knowledge

engineers and domain experts will indirectly
participate in the medical decision process.
Who is responsible if the advice is wrong?

1 The user?
2 The domain expert?
3 The knowledge engineer?
4 The programmer?
5 The company selling the software?
Google’s Autonomous Vehicle
Who’s liable when an autonomous car crashes? The driver?

Google? The programmer?
Do you make autonomous cars conform to current laws and
vehicle codes or establish new ones?
What about cybersecurity in vehicular networks?
Do you still allow manually operated cars when most cars are
autonomous?
Obtaining driver licence

Short Trip To Artificial Intelligence: Lecturer Adrian Groza

Uploaded by

Copyright:

Available Formats

Short Trip To Artificial Intelligence: Lecturer Adrian Groza

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Short Trip To Artificial Intelligence: Lecturer Adrian Groza

Uploaded by

Copyright:

Available Formats

Short Trip to Artificial Intelligence

Lecturer Adrian Groza

Intelligent Systems Group

March 25, 2016

A Logician’s View of the World

Let’s start for the very beginning:

A Logician’s View of the World

Let’s start for the very beginning:

Something important is missing

A Logician’s View of the World

Let’s start for the very beginning:

And now we said too much!!!

And now we said too much!!!

Who said that John was a man?

And now we said too much!!!

Who said that John was a man?

And now we said too much!!!

Who said that John was a man?

And now we said too much!!!

Who said that John was a man?

And now we said too much!!!

Who said that John was a man?

And now we said too much!!!

Who said that John was a man?

And now we said too much!!!

Who said that John was a man?

Knowledge Representation and Reasoning

2 the combination of several reasoning methods: rule based reasoning,

causal reasoning, model based reasoning, contraints based reasoning

Semantic Web and Ontologies

Semantic Web Vision

Semantic Web and Ontologies

Semantic Web Vision

Semantic Web and Ontologies

Semantic Web and Ontologies

Semantic Web and Ontologies

Semantic Web and Ontologies

Semantic Web and Ontologies

Fuzzy expert systems

Definition (Expert system)

The knowledge of the expert is represented as rules:

Fuzzy expert systems

Thinking humanly: ”The cognitive modeling approach.”

Fuzzy knowledge can be expressed in a very natural way using

Fuzzy expert systems

Fuzzy expert systems

Fuzzy Versus Probability

Fuzzy expert systems

Fuzzy Versus Probability

Fuzzy expert systems

Fuzzy Versus Probability

Fuzzy expert systems

Fuzzy Versus Probability

Reasoning with constraints

Simplification Combination Contradiction Redundancy

Reasoning with constraints

Combination Contradiction Redundancy

Reasoning with constraints

Reasoning with constraints