General Task
(Redirected from General Task Type)
Jump to navigation
Jump to search
A General Task is a task that exists across several domains.
- Context:
- It can have a distinctive set of Task Properties.
- It has:
- It can be:
- a Processing Task, that focuses on the ability to process an Input (e.g. Language Comprehension).
- a Production Task, that focuses on the ability to produce an Output (e.g. Language Production)
- It can have an Output Requirement that makes it a:
- a Deterministic Task, such as an Exact Matching Task.
- an Approximation Task, such as a Fuzzy Matching Task.
- It can have an Output Requirement of:
- a single Data Record, such as an Optimization Task.
- a Data Set, such as a Matching Task.
- a Data Sequence, such as a Sorting Task.
- It can require:
- all Valid Results, such as an Optimization Task, or a Matching Task.
- any Valid Result, such as a Scheduling Constraint Satisfaction Task.
- It can require a Maximum Record Read Operations, with the special case of Only One Read for Tasks with Data Streams.
- It can require a Minimum Time Performance Requirement, with the special case of Real Time Task.
- It can require a Distance Function (an Item Comparison Function ?)
- It can require a Cost Function (a Background Rule ?)
- It can require a Constraint Rule Set (a Background Rule ?)
- It can be a combination of General Task Types, such as a Satisfaction Task and an Optimization Task (to find an optimal point within a constrained space).
- Example(s):
- Deterministic Tasks.
- Data Sequence:
- Sorting Task: given a set and an Ordering Relation, create an Ordered Sequence.
- Data Set:
- A Searching Task: given a Space (Dataset) and a Query identify all Matching Items.
- A Matching Task: given a Dataset and a Pattern identify all Matching Items.
- A Similarity Search Task: given a Space and a point Pattern.
- Data Record:
- A Constraint Satisfaction Task: given a Constraint Set find any Solution.
- An Optimization Task: given a Cost Function find Optimal Solution.
- A Detection Task: given a Dataset and a Pattern Type predict Matching Items.
- Data Sequence:
- A Learning Task: given a Dataset improve Performance on a Testset.
- A Classification Task: given a Data Record predict its Classes.
- A Parsing Task: given a Formal Language and a string create a Parse Tree.
- A Combined Task:
- A Recognition Task combines a Detection Task and a Classification Task.
- …
- Deterministic Tasks.
- Counter-Example(s):
- a Domain Specific Task, such as a Movie Recommendation Task.
- See: General Algorithm Strategy.
References
1999
- (Gómez-Pérez & Benjamins, 1999) ⇒ Asunción Gómez-Pérez, and Richard Benjamins. (1999). “Overview of knowledge sharing and reuse components: Ontologies and problem-solving methods.” In: CEUR Workshop Proceedings.
- QUOTE: … The most illustrative linguistic ontologies are the Generalized Upper Model [BMF95], WordNet [Mil90] and Sensus [SPKR97]. The Generalized Upper Model (GUM) is a general task and domain-independent linguistic ontology. …