Artificial Intelligence - Question and Answer

Define intelligence.

Intelligence is a rather hard to define term.
            Intelligence is often defined in terms of what we understand as intelligence in humans.
             Allen Newell defines intelligence as the ability to bring all the knowledge a system has at its disposal to bear in the solution of a problem.
                   A more practical definition that has been used in the context of building artificial systems with intelligence is to perform better on tasks that humans currently do better.

What are the different approaches in defining artificial intelligence ?

Thinking rationally
Acting rationally
Thinking like a human
Acting like a human

Suppose you design a machine to pass the Turing test. What are the capabilities such a machine must have ?

Natural Language Processing
Knowledge representation
Automated reasoning
Machine Learning

Define an agent.
An agent is anything that can be viewed as perceiving its environment through sensors and executing actions using actuators.

What is a rational agent ?
A rational agent always selects an action based on the precept sequence it has received so as to maximize its (expected) performance measure give the precepts it has received and the knowledge possessed by it.

What is bounded rationality ?
A rational agent that can use only bounded resources cannot exhibit the optimal behaviour. A bounded rational agent does the best possible job of selecting good actions give its goal, and give its bounded resources.

What is an autonomous agent ?
Autonomous agents are software entities that are capable of independent action in dynamic, unpredictable environments, An autonomous agent can learn and adapt to a new environment.

Describe the salient features of an agent.

An agent perceives its environment using sensors
An agent takes actions in the environment using actuators
A rational agent acts so as to reach its goal. or to maximize its utility
Reactive agents decide their action on the basis of their current state and the percepts. Deliberative agents reason about their goals to decide their action.

Find out about the Mars rover.

a) What are the percepts for this agent ?
Spirit's sensor include
* panoramic and microscopic cameras,
* a radio receiver,
* spectrometers for studying rock samples including an alpha particle x-ray spectrometer, M'' ossbauer spectrometer, and miniature thermal emission spectrometer.

b) Characterize the operating environment
The environment (the Martian surface)

partially observable
non-deterministic
sequential
dynamic
continuous and
may be single-agent.

If a rover must cooperate with its mother ship or other rovers, or if mischievous Martians tamper with its progress, then the environment gains additional agents.

c) What are the actions the agent can take ?
The rover Spirit has

motor-driven wheels for locomotion
along with a robotic arm to bring sensors close to interesting rocks and a
rock abrasion tool (RAT) capable of efficiently drilling 45mm holes in hard volcanic rock.
Spirit also has a radio transmitter for communication.

d) How can one evaluate the performance of the agent ?
Performance measure : A Mars rover may be tasked with

maximizing the distance or variety of terrain it traverses,
or with collecting as many samples as possible,
or with finding life (for which it receives 1 point if it succeeds, and 0 points if it fails).

Criteria such as maximizing lifetime or minimizing power consumption are (at best) derived from more fundamental goals; e.g., if it crashes or runs out of power in the field, then it can't explore.

e) What sort of agent architecture do you think is most suitable for this agent ?
model-based reflex agent for low level navigation;
for route planning, experimentation etc. some combination of goal-based, and utility-based would be needed.