What is Intelligence (in its broadest sense)?
The capability to act purposefully towards the achievement of goals
What are the four key capabilities needed for intelligent systems?
- Perceiving 2. Reasoning 3. Acting 4. Adaptivity
What is Perceiving in AI systems?
The capability to perceive and understand what happens in environment, often through processing sensory data
What two types of reasoning exist in reasoning systems?
- Reactive reasoning (based on pre-defined rules) 2. Strategic reasoning (comparing current vs desired situation)
What is an autonomous agent?
An intelligent entity that can act in autonomy (without external control or supervision)
What are the three types of AI systems based on capabilities?
- Perceptive Systems 2. Reasoning Systems 3. Autonomous Agents
How does distributed perception differ from single AI perception?
In DAI, different entities perceive different portions of environment, while single AI has unified perception
What are the two types of interactions between agents?
- Indirect Interactions (behavior affects others) 2. Direct Interactions (explicit communication)
What are the five key reasons to use DAI?
- Environmental Coexistence 2. Complex, Inherently Distributed Problems 3. Reliability and Fault Tolerance 4. Scalability 5. Natural Modeling of Domains
What is a Distributed System?
A system of multiple computational processes, each with own local memory and capable of interacting
What is a Distributed Architecture?
Set of connected computational elements that don’t share memory/clock/space, with communication system for interaction
What is Distributed Computing?
Solving problems using distributed entities where each has partial knowledge/capability of solving the problem
What are the four factors that distinguish DAI from distributed computing?
- Degree of Adaptivity 2. Degree of Autonomy 3. Flexibility of Interactions 4. Learning Abilities
What are Distributed Algorithms?
Algorithms executed by concurrent distributed processes to solve distributed problems and coordinate actions
What are five main challenges in DAI?
- Communication Overhead 2. Consistency and Coherence 3. Scalability 4. Security and Trust 5. Unpredictability
What is Middleware?
Software layer between operating system and application software that enables components to interact effectively
What are the four key challenges middleware addresses?
- Transparency and Scalability 2. Heterogeneity 3. Application-level Adaptivity 4. Environment-level Adaptivity
What are the characteristics of Client-Server Models?
- Asymmetric roles 2. Centralized resource management 3. Scalability challenges
What are the characteristics of P2P Models?
- Decentralized 2. Scalable 3. Resilient
What is a Multiagent Model?
Autonomous independent software modules acting independently with own goals, interacting with environment and each other
