Behaviour and Reasoning Description Language (BRDL) Antonio Cerone - PowerPoint PPT Presentation

Behaviour and Reasoning Description Language (BRDL) Antonio Cerone Department of Computer Science School of Science and Technology Nazarbayev University Nur-Sultan, Kazakhstan email: antonio.cerone@nu.edu.kz A. Cerone, Nazarbayev University – p.1/26

CIFMA 2019, 17 September 2019 BRDL Behaviour and Reasoning Description Language • knolwledge representation • knowledge retrieval • reasoning • human behaviour • problem solving A. Cerone, Nazarbayev University – p.2/26

CIFMA 2019, 17 September 2019 Human Memory Environment ✻ Output Channel Input Channel ❄ Human Memory A. Cerone, Nazarbayev University – p.3/26

CIFMA 2019, 17 September 2019 Human Memory Environment ✻ Output Channel Input Channel ❄ Long-Term Sensory ✛ Memory Memory (LTM) ❄ Short-Term Memory (STM) A. Cerone, Nazarbayev University – p.3/26

CIFMA 2019, 17 September 2019 Long-Term Memory (LTM) Environment ✻ Output Channel Input Channel ❄ Sensory LTM Memory Declarative knowledge ✛ of the world “knowing what” ❄ Short-Term Memory (STM) A. Cerone, Nazarbayev University – p.4/26

CIFMA 2019, 17 September 2019 Long-Term Memory (LTM) Environment ✻ Output Channel Input Channel ❄ Sensory LTM Memory Declarative Procedural knowledge ✛ refers to of the world human skills “knowing what” “knowing how” ❄ Short-Term Memory (STM) A. Cerone, Nazarbayev University – p.4/26

CIFMA 2019, 17 September 2019 LTM Components Environment ✻ Output Channel Input Channel ❄ Sensory LTM Memory Declarative Procedural knowledge ✛ refers to of the world human skills “knowing what” “knowing how” Episodic ❄ events Short-Term Memory (STM) A. Cerone, Nazarbayev University – p.5/26

CIFMA 2019, 17 September 2019 LTM Components Environment ✻ Output Channel Input Channel ❄ Sensory LTM Memory Declarative Procedural knowledge ✛ refers to of the world human skills “knowing what” “knowing how” Episodic ❄ events Short-Term Semantic Memory facts (STM) A. Cerone, Nazarbayev University – p.5/26

CIFMA 2019, 17 September 2019 Semantic Memory Environment ✻ Output Channel Input Channel ❄ Sensory LTM Episodic Memory events ✛ Procedural human skills ❄ Short-Term Semantic Memory facts (STM) A. Cerone, Nazarbayev University – p.6/26

CIFMA 2019, 17 September 2019 Semantic Memory Environment ✻ Output Channel Input Channel ❄ Sensory LTM Episodic Memory events ✛ Semantic Procedural facts human skills semantic network ❄ of associations Short-Term deliberate control Memory processing (STM) A. Cerone, Nazarbayev University – p.6/26

CIFMA 2019, 17 September 2019 Semantic Memory Environment ✻ Output Channel Input Channel ❄ Sensory LTM Episodic Memory events ✛ Semantic Procedural facts human skills semantic network ❄ of associations Short-Term deliberate control Memory processing (STM) A. Cerone, Nazarbayev University – p.6/26

Example of Semantic Network four_legs ANIMAL bark ❍ ❨ ✟✟✟✟ ✯ � ❍ ✻ is_a does does has ❍ does ✠ � ❄ ❍ move breath DOG ❍❍❍❍ has ❥ tail A. Cerone, Nazarbayev University – p.7/26

Example of Semantic Network four_legs ANIMAL bark SHEEP ❍ ❨ ✟✟✟✟ ✯ � ❍ ✻ ✻ is_a does does has ❍ does works � ✠ ❄ ❍ is_a ✛ move breath DOG SHEEPDOG ❍❍❍❍ ✻ ✻ has is_a is_a ❥ is_a ✲ tail BASENJI HOUND COLLIE ❍❍❍❍ ✛ doesnt ✻ ✻ is_a bark does is_a ❥ is_a ✲ SNOOPY BEAGLE track LASSIE A. Cerone, Nazarbayev University – p.7/26

Example of Knowledge Domain four_legs ANIMAL bark SHEEP ❍ ❨ ✟✟✟✟ ✯ � ❍ ✻ ✻ is_a does does has ❍ does works � ✠ ❄ ❍ is_a ✛ move breath DOG SHEEPDOG ❍❍❍❍ ✻ ✻ has is_a is_a ❥ is_a ✲ tail BASENJI HOUND COLLIE ❍❍❍❍ ✛ doesnt ✻ ✻ bark does is_a is_a ❥ is_a ✲ SNOOPY BEAGLE track LASSIE dogs (knowledge domain) A. Cerone, Nazarbayev University – p.8/26

Example of Knowledge Domain four_legs animals ANIMAL bark SHEEP ❍ ❨ ✟✟✟✟ ✯ � ❍ ✻ ✻ is_a does does has ❍ does works � ✠ ❄ ❍ is_a ✛ move breath DOG SHEEPDOG ❍❍❍❍ ✻ ✻ has is_a is_a ❥ is_a ✲ tail BASENJI HOUND COLLIE ❍❍❍❍ ✛ doesnt ✻ ✻ bark does is_a is_a ❥ is_a ✲ SNOOPY BEAGLE track LASSIE dogs (knowledge domain) A. Cerone, Nazarbayev University – p.8/26

Example of Knowledge Domain four_legs animals ANIMAL bark SHEEP ❍ ❨ ✟✟✟✟ ✯ � ❍ ✻ ✻ is_a does does has ❍ does works � ✠ ❄ ❍ is_a ✛ move breath DOG SHEEPDOG ❍❍❍❍ ✻ ✻ has is_a is_a ❥ is_a ✲ tail BASENJI HOUND COLLIE ❍❍❍❍ ✛ doesnt ✻ ✻ bark does is_a is_a ❥ is_a ✲ SNOOPY BEAGLE track LASSIE ✻ ✻ dogs character character (knowledge domain) cartoons films A. Cerone, Nazarbayev University – p.8/26

CIFMA 2019, 17 September 2019 Knowledge Representation d 1 animals : animal | = ⇒ | does ( breath ) ANIMAL � does does � ✠ ❄ move breath A. Cerone, Nazarbayev University – p.9/26

CIFMA 2019, 17 September 2019 Knowledge Representation d 1 animals : animal | = ⇒ | does ( breath ) d 2 ⇒ | is _ a ( animal ) dogs : dog | = four_legs ANIMAL bark ❨ ❍ ✟✟✟✟ ✯ � ❍ ✻ is_a does does has ❍ does ✠ � ❄ ❍ move breath DOG ❍❍❍❍ ❥ tail has A. Cerone, Nazarbayev University – p.9/26

CIFMA 2019, 17 September 2019 Knowledge Representation d 1 animals : animal | = ⇒ | does ( breath ) d 2 ⇒ | is _ a ( animal ) dogs : dog | = d 3 dogs : dog | = ⇒ | does ( bark ) d 4 dogs : basenji | = ⇒ | doesnt ( bark ) A. Cerone, Nazarbayev University – p.9/26

CIFMA 2019, 17 September 2019 Knowledge Representation d 1 animals : animal | = ⇒ | does ( breath ) d 2 ⇒ | is _ a ( animal ) dogs : dog | = d 3 dogs : dog | = ⇒ | does ( bark ) d 4 dogs : basenji | = ⇒ | doesnt ( bark ) d 5 ⇒ | is _ a ( dog ) dogs : sheepdog | = d 6 dogs : sheepdog | = ⇒ | works ( sheep ) d 7 ⇒ | is _ a ( sheepdog ) dogs : collie | = d 8 ⇒ | is _ a ( collie ) dogs : lassie | = A. Cerone, Nazarbayev University – p.9/26

CIFMA 2019, 17 September 2019 Knowledge Representation d 1 animals : animal | = ⇒ | does ( breath ) d 2 ⇒ | is _ a ( animal ) dogs : dog | = d 3 dogs : dog | = ⇒ | does ( bark ) d 4 dogs : basenji | = ⇒ | doesnt ( bark ) d 5 ⇒ | is _ a ( dog ) dogs : sheepdog | = d 6 dogs : sheepdog | = ⇒ | works ( sheep ) d 7 ⇒ | is _ a ( sheepdog ) dogs : collie | = d 8 ⇒ | is _ a ( collie ) dogs : lassie | = d 9 dogs : hound | = ⇒ | does ( track ) A. Cerone, Nazarbayev University – p.9/26

CIFMA 2019, 17 September 2019 Semantic Network in BRDL d domain : category | ⇒ | type ( attribute ) = Environment d Example: dogs : dog | ⇒ | is _ a ( animal ) = ✻ Output Channel Input Channel ❄ Sensory LTM Episodic Memory events ✛ Semantic Procedural facts human skills semantic network ❄ of associations Short-Term deliberate control Memory processing (STM) A. Cerone, Nazarbayev University – p.10/26

CIFMA 2019, 17 September 2019 Semantic Network in BRDL d domain : category | ⇒ | type ( attribute ) = Environment d Example: dogs : dog | ⇒ | is _ a ( animal ) = ✻ Output Channel Input Channel ❄ Sensory LTM Episodic Memory events ✛ Semantic Procedural facts human skills ❄ d dogs : dog | ⇒ | = is _ a ( animal ) Short-Term Memory (STM) A. Cerone, Nazarbayev University – p.10/26

CIFMA 2019, 17 September 2019 Knowledge Retrieval d domain : category | ⇒ | type ( attribute ) = Environment goal ( domain, type ?( category, attribute )) ✻ Output Channel Input Channel ❄ Sensory LTM Episodic Memory events ✛ Semantic Procedural facts human skills ❄ d dogs : dog | ⇒ | = is _ a ( animal ) Short-Term Memory (STM) A. Cerone, Nazarbayev University – p.11/26

CIFMA 2019, 17 September 2019 Knowledge Retrieval d domain : category | ⇒ | type ( attribute ) = Environment goal ( domain, type ?( category, attribute )) ✻ Output Channel Input Channel ❄ Sensory LTM Episodic Memory events ✛ Semantic Procedural facts human skills ❄ d dogs : dog | ⇒ | = is _ a ( animal ) STM goal ( dogs, is _ a ?( dog, animal ) A. Cerone, Nazarbayev University – p.11/26

CIFMA 2019, 17 September 2019 Knowledge Retrieval d domain : category | ⇒ | type ( attribute ) = Environment goal ( domain, type ?( category, attribute )) ✻ Output Channel Input Channel ❄ Sensory LTM Episodic Memory events ✛ Semantic Procedural facts human skills ❄ d dogs : dog | ⇒ | = ✛ is _ a ( animal ) STM goal ( dogs, pattern matching ✲ is _ a ?( dog, animal ) A. Cerone, Nazarbayev University – p.11/26