Distributed Artificial Intelligence (DAI) is an approach to solving complex learning, planning, and decision making problems. It is embarrassingly parallel, thus able to exploit large scale computation and spatial distribution of computing resources.
This question is very generic however, it depends upon the application of the DAI, for example, it's a progressing filed in 5G Networks but hasn't seen any applications in computer vision.
In any meaningful comparison of distributed and centralized systems, the distributed system advantages will win out. Centralized systems are vulnerable to attack, to unexpected scenarios, and simple failure. DAI is more robust since the distributed elements are typically not interdependent. The best example in the natural world is the distribution of the entire set of genetic instructions in every cell of a complex organism. Artificial intelligence systems that emulate natural ones will tend towards distributed strategies.
Hi,
Distributed Computing becomes feasible when:
1. The cost/ Quantity of data transfer to server exceed its computational value.
2. Data privacy is not an issue.
3. Lots of unused CPU time, in several machines are at your disposal.
AI is a perfect Candidate to leverage from Distributed computing. With the rising amount of computation power in a CPU, training the data remotely or distributed dose makes a lot of sense. In future the server might just keep the final trained models whereas the training and the data can be kept distributed.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Artificial Intelligence