blackboard architecture.pptx

blackboard architecture

1. Blackboard Architecture It is used for problem solving especially for a complex problem which a single expert is not able to handle. Example: multilingual translation. In general, problem-solving in blackboard architecture begins with the announcement of a problem and writing initial data onto the blackboard . The specialists are watching the blackboard looking for an opportunity in order to make a contribution to solution development. When a specialist finds this opportunity, they record the contribution on the blackboard, in the hope that others will use his contribution for final problem-solving. This process continues until the problem is solved.

2. Components of Blackboard architecture: Knowledge sources: Knowledge sources (KSs) are "independent modules that contain the knowledge needed for problem-solving" . KSs need to focus on the progress of solving the current problem However, it is not necessary for a KS to communicate with or even know the existence of other KSs. Blackboard: Blackboard is the place where the specialists work together to find the solution . The formal definition of the blackboard is "a globally accessible database which is used for intermediate, partial results of problem-solving ". Each KS can read and write the shared blackboard. Controller: The controller decides " which knowledge source to execute next in runtime for optimal problem solution“.

3. Blackboard Architecture Controller Knowledge Sources Blackboard

4. Applicable problems: Complex problems where information is ambiguous A large problem consisting of smaller sub-problems A problem with many different solutions Resilience to change: When new knowledge sources are added, the other components of the system are not affected. If new information is added to the problem, new knowledge sources can easily be added. There are no dependencies between knowledge sources, so they can easily be put in, taken out, or swapped. Negative behavior: The difficulty of testing: Because of the non-deterministic nature of the blackboard architecture, the solution of each execution may not be consistent. Single point of failure: If the blackboard crashes, the whole system will stop working because none of KSs are able to access the memory to solve the problem.

5. The blackboard approach to organizing a large program was first presented in the HEARSAY-II research. HEARSAY-II was a speech understanding program; it was initially designed as the front end for a library database of computer science articles. The user of the library would address the computer in spoken English with queries such as, “Are any by Feigenbaum and Feldman?” and the computer would answer the question with information from the library database. Speech understanding requires that we integrate a number of different processes like Signal processing; recognition of phonemes, syllables, and words; syntactic parsing; and semantic analysis. The blackboard architecture allowed HEARSAY-II to coordinate the several different knowledge sources required for this complex task.

6. KS 1 – the waveform of the acoustic signal KS 2 – the phonemes or possible sound segment of the acoustic signal KS 3 - The syllables that the phonemes could produce. KS 4 – try to generate possible word sequences KS 5 – puts the word sequence into possible phrases