21
9a. Rules in Production Systems

9a. Rules in Production Systems - Stanford UniversityDirection of Reasoning • Possible to inter-mix goal/data driven with forward/backward chaining • Even though Prolog is a backward

  • Upload
    others

  • View
    3

  • Download
    0

Embed Size (px)

Citation preview

Page 1: 9a. Rules in Production Systems - Stanford UniversityDirection of Reasoning • Possible to inter-mix goal/data driven with forward/backward chaining • Even though Prolog is a backward

9a. Rules in Production Systems

Page 2: 9a. Rules in Production Systems - Stanford UniversityDirection of Reasoning • Possible to inter-mix goal/data driven with forward/backward chaining • Even though Prolog is a backward

Outline

• Direction of Reasoning• Production Systems

– Basics– Examples– Efficiency

• Applications– Early production systems– Some modern production systems

Page 3: 9a. Rules in Production Systems - Stanford UniversityDirection of Reasoning • Possible to inter-mix goal/data driven with forward/backward chaining • Even though Prolog is a backward

Contrast it with P \/ Q

Page 4: 9a. Rules in Production Systems - Stanford UniversityDirection of Reasoning • Possible to inter-mix goal/data driven with forward/backward chaining • Even though Prolog is a backward

Direction of Reasoning

• Possible to inter-mix goal/data driven with forward/backward chaining

• Even though Prolog is a backward chaining system, it can be usedto do data driven reasoning

Page 5: 9a. Rules in Production Systems - Stanford UniversityDirection of Reasoning • Possible to inter-mix goal/data driven with forward/backward chaining • Even though Prolog is a backward

Production Systems

• The production systems emphasize forward chaining over rules as a way of reasoning– Rules are also referred to as ``productions’’– Rules are the primary form of knowledge representation

• Referred to as rule-based systems– Foundation of the expert systems technology

Page 6: 9a. Rules in Production Systems - Stanford UniversityDirection of Reasoning • Possible to inter-mix goal/data driven with forward/backward chaining • Even though Prolog is a backward

``Water is Hot’’

``Turnoff the heat’’

Page 7: 9a. Rules in Production Systems - Stanford UniversityDirection of Reasoning • Possible to inter-mix goal/data driven with forward/backward chaining • Even though Prolog is a backward
Page 8: 9a. Rules in Production Systems - Stanford UniversityDirection of Reasoning • Possible to inter-mix goal/data driven with forward/backward chaining • Even though Prolog is a backward

an evaluable expression in [ ]

Page 9: 9a. Rules in Production Systems - Stanford UniversityDirection of Reasoning • Possible to inter-mix goal/data driven with forward/backward chaining • Even though Prolog is a backward

(not applicable if condition was negative)(not applicable if condition was negative)

forall x student(x) => person(x)

Page 10: 9a. Rules in Production Systems - Stanford UniversityDirection of Reasoning • Possible to inter-mix goal/data driven with forward/backward chaining • Even though Prolog is a backward
Page 11: 9a. Rules in Production Systems - Stanford UniversityDirection of Reasoning • Possible to inter-mix goal/data driven with forward/backward chaining • Even though Prolog is a backward
Page 12: 9a. Rules in Production Systems - Stanford UniversityDirection of Reasoning • Possible to inter-mix goal/data driven with forward/backward chaining • Even though Prolog is a backward
Page 13: 9a. Rules in Production Systems - Stanford UniversityDirection of Reasoning • Possible to inter-mix goal/data driven with forward/backward chaining • Even though Prolog is a backward
Page 14: 9a. Rules in Production Systems - Stanford UniversityDirection of Reasoning • Possible to inter-mix goal/data driven with forward/backward chaining • Even though Prolog is a backward
Page 15: 9a. Rules in Production Systems - Stanford UniversityDirection of Reasoning • Possible to inter-mix goal/data driven with forward/backward chaining • Even though Prolog is a backward
Page 16: 9a. Rules in Production Systems - Stanford UniversityDirection of Reasoning • Possible to inter-mix goal/data driven with forward/backward chaining • Even though Prolog is a backward
Page 17: 9a. Rules in Production Systems - Stanford UniversityDirection of Reasoning • Possible to inter-mix goal/data driven with forward/backward chaining • Even though Prolog is a backward
Page 18: 9a. Rules in Production Systems - Stanford UniversityDirection of Reasoning • Possible to inter-mix goal/data driven with forward/backward chaining • Even though Prolog is a backward
Page 19: 9a. Rules in Production Systems - Stanford UniversityDirection of Reasoning • Possible to inter-mix goal/data driven with forward/backward chaining • Even though Prolog is a backward

Current Applications and Tools

• Currently available tools that use production system technolology– Jess semi-open Java tool, popular among researchers– Drools open source C tool, got popular in last two years

• Production systems are used to invoke external procedures from aprogram (especially in semantic web services)– External queries– Executing external actions that might cause side effects

• SIRI uses production system style rules to implement a cognitiveassistant

Page 20: 9a. Rules in Production Systems - Stanford UniversityDirection of Reasoning • Possible to inter-mix goal/data driven with forward/backward chaining • Even though Prolog is a backward

Business Rules

• Introduction to business rules– http://www.youtube.com/watch?v=2ouhJeH02HU&feature=grec_index

• An example business rule– http://www.youtube.com/watch?v=AK7BSwl2UlY– http://www.youtube.com/watch?v=l_DmDvWR_wM

Page 21: 9a. Rules in Production Systems - Stanford UniversityDirection of Reasoning • Possible to inter-mix goal/data driven with forward/backward chaining • Even though Prolog is a backward

Recommended Reading

• Chapter 7 of Brachman & Levesque