The Inference Process

23
The Inference Process Rules There are two main forms of Inferencing Forward Chaining Backward Chaining

description

The Inference Process. Rules There are two main forms of Inferencing Forward Chaining Backward Chaining. The Inference Process. Rules Rules follow an IF…THEN format English IF the pump’s temperature is hot THEN follow maintenance instructions Aion DS - PowerPoint PPT Presentation

Transcript of The Inference Process

Page 1: The Inference Process

The Inference ProcessThe Inference Process• Rules• There are two main forms of Inferencing

• Forward Chaining

• Backward Chaining

Page 2: The Inference Process

The Inference ProcessThe Inference Process• Rules• Rules follow an IF…THEN format• English

IF the pump’s temperature is hot

THEN follow maintenance instructions

• Aion DSIFMATCH pump with temperature is “hot”

THEN

send (print maintenance to pump)

END

Page 3: The Inference Process

The Inference ProcessThe Inference Process

– IF– the site of the culture is Blood AND– the identity of the organism is Unknown AND– the strain of the organism is Gramneg AND– the morphology of the organism is Rod AND– the patient has been seriously burned– THEN– there is weakly suggestive evidence (0.4) that the– identity of the organism is Pseudomonas

MYCIN (an expert system to diagnose blood disorders)

Page 4: The Inference Process

The Inference ProcessThe Inference Process• Forward vs. Backward Chaining• Do we drive the inference process based on the

known facts (data) or on the questions (goals) we need answers to?

• Data Driven• Starts with the known data and fires rules to infer

new information (forward chaining)

• Goal Driven• Starts with the goal and tries to match facts to the

solution (backward chaining)

Page 5: The Inference Process

The Inference ProcessThe Inference Process• Forward Chaining1 Enter new data

2 Fire forward chaining rules

3 Rule actions infer new data values

4 Go to step 2

5 repeat until no new data can be inferred

6 If no solution, rule base is insufficient

Page 6: The Inference Process

The Inference ProcessThe Inference Process

Get some new data

1

Fire forward chaining rules Infer new data values

from rule actions

2

3

Add new data values to knowledge base

4

Forward Chaining

Page 7: The Inference Process

The Inference ProcessThe Inference Process• Cascading Rules• Only applicable in forward chaining• Rule executes causing inference of new data• New data is added to the knowledge base• New data added causes other rules to fire

• Can be very time consuming and inefficient in large systems

Page 8: The Inference Process

The Inference ProcessThe Inference Process• Forward Chaining• The data is usually entered “up front”• Is usually done on a form by form basis• Relevant questions must be grouped together• Rules will only fire when all information is

available• The inference engine will not try to find out any

unknown information• All possible questions need to be asked at the

start of a consultation

Page 9: The Inference Process

The Inference ProcessThe Inference Process• Why Forward Chain?Reasons to Forward Chain Examples

You want to know everything that can possibly be concluded about a set of data

Monitoring for mechanical problems on a production line

Scanning a new loan application for problem areas

Many conclusions are possible from a single data item

Filtering sensor data

It is important to communicate new conclusions to the user immediately

Advice to shut down faulty machines

Data entry errors

Page 10: The Inference Process

The Inference ProcessThe Inference Process• Backward Chaining1 State a specific goal (question)

2 Find rules which resolve the goal

3 At runtime, answer questions to satisfy the antecedents of the rules as required

4 Obtain a result (goal resolved or not)

Page 11: The Inference Process

The Inference ProcessThe Inference Process

State primary goal to source

1

Fire backward chaining rules

Primary goal sourced

2

4

Source sub goals

3

Backward Chaining

Page 12: The Inference Process

The Inference ProcessThe Inference Process• Why Backward Chain?Reasons to Backward Chain Examples

There is a clear set of statements which must be confirmed or denied

Which machine is causing the quality control problem?

A large number of questions could be asked of the user, but typically only a few are necessary to resolve a situation

Processing of a motor claim for vandalism; not necessary to know about personal injuries

It is desirable to have interactive dialogue with the user

Asking machine operator detailed questions about suspect machinery

Rule execution depends on data gathering which may be expensive or difficult

Real-time observations by the user

Page 13: The Inference Process

The Inference ProcessThe Inference Process• Backward or Forward Chaining?• Criteria to consider -• The logical reasoning process• Design features of the system• What are the inputs and where do they

come from?• What are the outputs and where do they go?• How do these map to forward or backward

chaining?

Page 14: The Inference Process

The Inference ProcessThe Inference Process• ExamplesUse Forward Chaining Use Backward Chaining

Sensor indicates machine failure; want to find out what happens next

Defect observed in product; want to locate faulty machine

User types erroneous input for insurance claim; want to alert user

Suspect an overpayment; want to check form for erroneous input

Stock value suddenly drops; want to predict market response

FTSE industrials drop; want to know which stocks will be affected

Page 15: The Inference Process

The Inference ProcessThe Inference Process• Chaining in Action• A list of known facts - A, B, D, G, P, Q, R, S.

Page 16: The Inference Process

The Inference ProcessThe Inference Process• Chaining in Action• A list of known facts - A, B, D, G, P, Q, R, S.

Raining outside?

Page 17: The Inference Process

The Inference ProcessThe Inference Process• Chaining in Action• A list of known facts - A, B, D, G, P, Q, R, S.

Cold?

Page 18: The Inference Process

The Inference ProcessThe Inference Process• Chaining in - X is the goal• A list of Rules

1 A + I => X

2 A+B => C

3 C+D => E

4 F+G => H

5 E+H => X

6 A+C => F

7 P+Q => R

8 R+S => T

Page 19: The Inference Process

The Inference ProcessThe Inference Process• Chaining in Action (Backward)• A list of Rules

1 A + I => X

2 A+B => C

3 C+D => E

4 F+G => H

5 E+H => X

6 A+C => F

7 P+Q => R

8 R+S => T

Rule 1 never fired because either “I” was not true or the user did not know it was true

Rules 7+8 never fired because they were not relevant to proving our goal of “X”, though they may have been used if we had another consultation with a different goal

Page 20: The Inference Process

The Inference ProcessThe Inference Process• Chaining in Action (Forward)• A list of known facts - “A, B, D, P, Q, R, S”,

are all true (we asked the user) and “I” is false or not known

Page 21: The Inference Process

The Inference ProcessThe Inference Process• Chaining in Action (Forward)• A list of Rules

1 A + I => X

2 A+B => C

3 C+D => E

4 F+G => H

5 E+H => X

6 A+C => F

7 P+Q => R

8 R+S => T

Page 22: The Inference Process

The Inference ProcessThe Inference Process• Chaining in Action (Forward)• A list of Rules

1 A + I => X

2X A+B => C

3X C+D => E

4 F+G => H

5 E+H => X

6X A+C => F

7X P+Q => R

8X R+S => T

Page 23: The Inference Process

The Inference ProcessThe Inference Process• Backward or Forward Chaining?• Backward chaining was more focused but only

answered the question asked• Forward chaining found all possible results but

needed more information “up front”• Backward chaining never used rules 7 and 8