Chapter 4 [ C ] Flashcards
Describe a Learning-based IS (L-IS).
Learning refers to system improving its performance with experience, with respect to some task.
When is a System said to learn from experience E with respect to some class of actions A and performance measure P?
- if its performance at the set of actions A, as measured by P, improves with experience E.
Describe an Adaptive transport scheduling example for L-IS.
- A = “a logistics vehicle picks up goods on route”,
- E = “traveling the route”,
- P = “deviation of actual time from predicted time”.
-> Improvement is measure P reducing to zero.
In L-IS Design of the learning element depends on?
- Which model is learned,
- The type of feedback and the model or knowledge representation.
What are the 3 main types of learning or feedback which can be used?
- supervised learning,
- unsupervised learning
- reinforcement learning
[ since Learning may need model representations that can handle uncertainty ]
What are the commonly used types of knowledge representations?
- Blackboard / EDA systems
- Production / rule-based systems
- Syntactical: RDBMS, XML Web services
- Semantic Type KBs (Ontology-based Systems)
- Classic-Logic-Based KBs
- Soft Computing
In a production system, knowledge is represented as _____ ?
a set of rules or productions
Rules are typically defined as ______ ?
IF-fact | THEN-fact
IF-fact is aka?
- Condition / Antecedent part
THEN-fact is aka?
- Action / Consequence part
Rule-based systems use a set of predefined rules to make decisions or perform actions based on input data.
True
Give some examples of rule-based engines fragments for the other 3 scenarios:
1. Personal memories
2. Foodstuff management
3. Utility Regulation
- Personal memories
Rule: If today’s date matches a stored event date, trigger the memory recall. [Snapchat] - Foodstuff management
Rule: If the current date is within a week of the expiration date of a stored food item, send a notification. - Utility Regulation
Rule: If energy consumption exceeds a certain threshold during peak hours, send an alert and suggest energy-saving measures.
How can Rules get added to the KB Systems?
- Manually [ e.g: enterprise policy-based systems ]
- Automatically [ e.g : machine learning ]
________ enable rule-based systems to incorporated as part of more general distributed systems versus as part of more specialiszed IS?
rule-based engines
Describe a Blackboard (BB) system.
- is a collaborative knowledge-based approach where multiple knowledge sources contribute to solving complex problems.
Knowledge sources of a BB can be?
- Independent & distributed
- Heterogeneous
Give an example of Ontology using Vehicle.
- Concepts:
- Vehicle: Represents the general concept of a vehicle.
- Car: A specific type of vehicle. - Relationships:
- “is-a” Relationship: Car is-a Vehicle
- “has-a” Relationship: Car has-a Engine. - Properties:
- Color Property: Vehicle has Color, Car has Color.
- Fuel Type Property: Vehicle has Fuel Type, Car has Fuel Type. - Constraints:
- If a vehicle is a Car, it must have exactly four wheels.
Give a formal and informal definition of ontology.
- Informally : a collection of descriptions of the world that helps us to define the meaning of their actions on the world.
- Formally : A formal, explicit specifications of a shared conceptualization
What is the r/n ship b/n Ontology and Semantic (IS) : Knowledge Representation (KR)?
- Ontology based models support semantic conceptualization & can directly support reasoning
______ is often used synonymously with the term semantic KR?
Ontology
Depending on how concepts and their relationships are defined and organized what are the ranges of Knowledge Representation (KR) [ or Ontology ]?
- Light-weight
- Medium-weight
- Heavy-weight
Describe what a Light-weight Semantic Representation is.
- have a simple conceptualization with simple structures.
- representations may not be highly machine-readable or easily relatable to other terms.
- Example: Dictionaries
The most widely used light-weight KRs are based on _____ ?
W3C Web XML standards.
Explain W3C Web XML standards?
- Defines an unnamed hierarchy of concepts & properties
- Acts as a basic node labelled graph representation
Example:
<ontology>
<concept>
<property>
<value> Four </value>
</property>
</concept>
</ontology>
_____ is a language designed for exchanging extensible application specific hierarchical data structures?
XML
XML extensions Is used for _____ ?
SOC
XML is an easy data format on which to build automated machine-understandable processing.
False.
- It is a difficult data format
_____ & _____ represent Node Labelled
Graph & Edge Labelled Graph respectivley?
Light-Weight KR & Medium-Weight KR
Why was the Semantic Web (SW) created?
- to evolve the Web from machine-readable to machine-understandable
- to support richer service interoperability
What are the suit of KRs defined by SW and their weights?
- RDF (Resource Description Framework):
Weight: Medium to Heavy - RDFS (RDF Schema):
Weight: Medium to Heavy - OWL (Web Ontology Language):
Weight: Heavy
What are the Semantic KR: Design Issues?
- Open World versus Closed World Semantics
- Knowledge Life-cycle and Knowledge Management
Compare Open World versus Closed World Semantics.
- Open world : knowledge about a domain is incomplete, and new information may be discovered over time.
- Closed world : knowledge about a domain is complete and known. [ If not T then F ]
What aspects are involved in Knowledge Life-cycle and Knowledge Management?
- Creating Knowledge
- Knowledge Deployment
- Maintaining Knowledge
_____ is expressed by data is not present, it is false (negation as failure)?
Closed World Semantics
RDBMS type KB models assume _______ semantics?
closed-world
Semantic type KBs tend to use ______ semantics.
open-world
______ regard absence as unknown (negation as unknown)?
Open World Semantics
Semantics of data is often defined to be declarative.
False.
- is dynamic, heterogeneous and depends on context
Semantics is undefined or variable.
True
_______ enable less expert developers & users to create knowledge models.
interactive knowledge creation tools
What is a key feature of KB IS: Creation Tools?
the ability to export knowledge models in a format that can be easily interpreted, parsed, and invoked by computer applications, often through APIs
Give examples of KB IS: Creation Tools.
- Protégé
- Jena
Devices embedded into a local environment often have _______ of their environment.
partial view [ rather than a global view]
________ is positioned at the core of information systems that need to support reasoning about the system?
Classic Logic, based on first-order predicate logic
_____ systems involve the manipulation of
symbols in the form of logic formulae?
Classic Logic Systems
[ logic reasoning systems
[ deliberative IS
[ symbolic Al
What is Propositional logic?
knowledge represented in the form of relations which are either true or false.
What are The standard logic operators?
- And, or, not, equals, implies
Describe a Predicate logic.
supports more expressive sentences than propositions by allowing properties to be related to objects or values.
Give examples for Predicate logic and Propositional Logic.
- P: It is raining.
Q: I am staying indoors.
P⇒Q: If it is raining, then I am staying indoors. - Sentence: “Device A is in hibernate mode.”
Predicate: mode ( Device A , Hibernate )
- Here, the predicate relates the object “Device A” to the property “Hibernate mode.”
_____ is the Most c o m m o n form of Predicate logic?
First-Order Predicate Logic [ FOPL ]
What is Description Logic (DL)?
- Logic based on combining First-Order Logic (FOL) with a conceptualization on graphs [ i.e: RDF Schema (RDF-S) … ]
______ is Used extensively in Heavy-weight Ontology languages?
Description Logics (DL)
Explain Reasoning?
- Aka inferencing, involves logical operations on logical statements n order to draw conclusions.
E.g: A entails B, A |= B.
Inferencing is used to search for entailments.
True
______ is used to check that the entailments of sentences are valid in all possible worlds or models.
Model Checking
What are Tautologies?
Sentences that are valid in all possible worlds
What are the design issues of Reasoning?
Reasoning needs to be :
* scalable
* selectively used
* computationally efficient
What are the Challenges of Classic Logic Systems?
- Difficulty in expressing exceptions
- Imprecision
- Uncertainty
- High computation is needed to establish truth
- Logical inconsistencies
- Existence of Different sub-types and extensions to classical logic
Many decisions which involve interaction with humans and the physical world are soft.
True
- rather than being expressed as either true or false
What is Soft computing?
- is a computing paradigm that deals with problems that are inherently imprecise and uncertain, often involving human judgment
_____ is also called a Belief network or Bayesian Network (BN)?
Probabilistic network
What are Probabilistic networks?
- networks used to model the likelihood of events that are uncertain in a proposition or predicate
_______ consider a prior or unconditional probability and conditional or posterior probabilities.
Probabilistic networks
What is the Product Law?
- Expresses a conditional probability in terms of another conditional probability and two unconditional probabilities.
What is Fuzzy Logic?
is a mathematical framework that deals with imprecision and approximation
Explain Fuzzy Logic Application in the Adaptive
Transport Scheduling Scenario.
- If the bus is traveling slowly away from the pickup point
- and a passenger is moving quickly towards the pickup point
- then slow down the vehicle to stop near the pickup point.
- Terms : slowly, quickly, and near, act as fuzzy
descriptors.
List the IS System Operations.
- Searching
- Planning
- Reasoning
- Learning
_________ Is a problem-solving technique that systematically explores a space of problem statments.
Searching
When searching, alternative solutions is searched to find an answer.
True
How do we express a search problem?
- Start state
- Goal state
- Goal test function [ is it goal state or nah ]
- Utility function [ is it desired or nah ]
What is Uninformed search?
- Also known as blind search, is a type of search algorithm that does not have any information about the structure or likelihood of success in the search space.
- Checks every node
What Traversal Strategies does Uninformed search algorithms typically use?
- Breadth-first search [ BFS ]: Explores the search space level by level, starting from the root.
- Depth-first search [ DFS ]: Explores as far as possible along one branch before backtracking.
Uninformed problem space searches tend to operate in the ______ direction?
forward
- Forward-chaining:
What are the Challenges in Uninformed Search Algorithm Design?
- Handling Unknown Start State: the start state maybe unknown
- Non-Uniform Search Space: different nodes may have varying numbers of branches
- Multiple-Valued Problem Spaces: multiple possible solutions or paths
- Large Search Spaces
_____ is a general solution to reduce the
computation of an uninformed search?
Informed search
_______ Uses problem specific information to limit the problem space?
Informed search
What is the core component of an informed search?
- A heuristic function
What is A heuristic function?
- Is a function that depends on the current node in the problem space.
- Guides the search algorithm by prioritizing paths that seem more promising based on the estimated cost.
What is A* search?
a popular informed search algorithm
What is Planning?
Planning involves searching for a plan and then executing the plan.
What is represented by the Planning model?
- States
- Goals
- Actions : which transition states towards goals
- Chains of actions : between non-adjacent states
- Heuristic cost function : to allow the choice of multiple paths
When planning is modeled as a graph what do the nodes and the link between nodes represent?
- nodes : represent states
- link between nodes : represent actions
Explain The integration of planning and Geographic Information Systems (G-IS) design.
- It involves using planning to enable the selection of a chain of actions that will lead to the achievement of a goal within a geographic context /Environment
