ARTIFICIAL INTELLIGENCE
Lecture 2: Introduction to Agents
Asst. Prof. Syed Faisal Ali
[email protected]
Outcome Based Education
Learning Objectives:
Learning Outcomes:
1. To understand different types of agents  1.
Students will demonstrate an
and its mechanism.
understanding
of
agent-based
AI
architectures.
2. To understand the types of agents and
how agents can work.
 2. Students will get familiar with the
environments of agents.
3. Applications and performances of
agents.
 3. Students will know the performance
parameters based on PAGE and PEAS.
3/13/2017
© Asst. Prof. Syed Faisal Ali
2
What are Agents?
 An intelligent agent is a software that assists people and act on their behalf.
 Agents are used to provide a consistent viewpoint on various topics in the field AI.
 Agents require essential skills to perform tasks that require intelligence.
 Intelligent agents use methods and techniques from the field of AI.
3/13/2017
© Asst. Prof. Syed Faisal Ali
3
An agent and its Environment
sensors
percepts
?
environment
agent
actions
effectors
In general, an entity that interacts with its environment, perception through sensors and
actions through effectors or actuators
3/13/2017
© Asst. Prof. Syed Faisal Ali
4
How to design an intelligent agent?
 Definition: An agent perceives its environment via sensors and acts in that
environment with its effectors.
Hence, an agent gets percepts one at a time, and maps this percept sequence to
actions (one action at a time).
 Properties:
 Autonomous ‫خود مختار‬
 Interacts with other agents plus the environment
 Reactive to the environment
 Pro-active (goal-directed)
3/13/2017
© Asst. Prof. Syed Faisal Ali
5
Agents & Environments
 an agent perceives its environment through sensors
 the complete set of inputs at a given time is called a percept
 the current percept, or a sequence of percepts may influence the actions of an
agent
 it can change the environment through actuators
 an operation involving an actuator is called an action
 actions can be grouped into action sequences
3/13/2017
© Asst. Prof. Syed Faisal Ali
6
Examples of agents
 human agent
 eyes, ears, skin, taste buds, etc. for sensors
 hands, fingers, legs, mouth, etc. for actuators
 powered by muscles
 robot
 camera, infrared, bumper, etc. for sensors
 grippers, wheels, lights, speakers, etc. for actuators
 often powered by motors
 software agent
 functions as sensors
 information provided as input to functions in the form of encoded bit strings
or symbols
 functions as actuators
 results deliver the output © Asst. Prof. Syed Faisal Ali
3/13/2017
7
Examples of Agents in different Types of Applications
Agent type
Percept
Action
Goals
Environment
Medical diagnosis
system
Symptoms, findings,
patient’s answers
Questions, tests,
treatments
Healthy patients,
minimize costs
Patient, hospital
Satellite image
analysis system
Pixels of varying
intensity, color
Print a categorization
of scene
Correct
categorization
Images from orbiting
satellite
Part-Picking Robot
Pixels of varying
intensity
Pick up parts and sort Place parts in correct
into bins
bins
Refinery Controller
Temperature, Pressure
readings
Open, close valves,
adjust temperatures
Maximize purity, Yield
safety
Refinery
Interactive English
tutor
Typed Words
Print exercises,
suggestions,
corrections
Maximize student’s
score on test
Set of students
3/13/2017
© Asst. Prof. Syed Faisal Ali
Conveyor belts with
parts
8
Definition of Ideal Rational Agent
Ideal Rational Agent: For each possible percept sequence, such an agent does
whatever action is expected to maximize its performance measure, on the basis of
the evidence provided by the percept sequence and whatever built-in knowledge the
agent has.
What do you think? Is this an acceptable definition?
Not looking left when crossing the street: If I don’t see a car coming from the left, it is rational
to cross the street?
No. Should also consider taking information gathering actions.
Limited/costly computation time
Bounded rationality
Limited/costly memory
…
3/13/2017
© Asst. Prof. Syed Faisal Ali
9
Agents and their actions
• a rational agent does “the right thing”
• the action that leads to the best outcome under the given circumstances
• an agent function maps percept sequences to actions
• abstract mathematical description
• an agent program is a concrete implementation of the respective function
• it runs on a specific agent architecture (“platform”)
problems:
• what is “ the right thing”
• how do you measure the “best outcome”
3/13/2017
© Asst. Prof. Syed Faisal Ali
10
Performance of Agents
• criteria for measuring the outcome and the expenses of the agent
• often subjective, but should be objective
• task dependent
• time may be important
Example: vacuum agent
• number of tiles cleaned during a certain period
• based on the agent’s report, or validated by an objective
authority
• doesn’t consider expenses of the agent, side effects
• energy, noise, loss of useful objects, damaged
furniture, scratched floor
• might lead to unwanted activities
• agent re-cleans clean tiles, covers only part of the
room, drops dirt on tiles to have more tiles to clean,
etc.
3/13/2017
© Asst. Prof. Syed Faisal Ali
Vacc.
Agent
11
Rational Agent and OMNISCIENCE
• a rational agent is not omniscient
• it doesn’t know the actual outcome of its actions
• it may not know certain aspects of its environment
• rationality takes into account the limitations of the agent
• percept sequence, background knowledge, feasible actions
• it deals with the expected outcome of actions
3/13/2017
© Asst. Prof. Syed Faisal Ali
12
Environment Properties
1.
2.
3.
4.
5.
6.
7.
fully observable vs. partially observable
• sensors capture all relevant information from the environment
deterministic vs. stochastic (non-deterministic)
• changes in the environment are predictable
episodic vs. sequential (non-episodic)
• independent perceiving-acting episodes
static vs. dynamic
• no changes while the agent is “thinking”
discrete vs. continuous
• limited number of distinct percepts/actions
single vs. multiple agents
• interaction and collaboration among agents
competitive, cooperative
3/13/2017
© Asst. Prof. Syed Faisal Ali
13
PEAS Description of Task Environment
used for high-level characterization of agents
Performance
Measures
used to evaluate how well an agent solves the task at hand
Environment
surroundings beyond the control of the agent
Actuators
determine the actions the agent can perform
Sensors
provide information about the current state of
the environment
3/13/2017
© Asst. Prof. Syed Faisal Ali
14
PAGE Description
used for high-level characterization of agents
Percepts
information acquired through the agent’s sensory system
Actions
operations performed by the agent on the environment
through its actuators
Goals
desired outcome of the task with a measurable
performance
Environment
surroundings beyond the control of the agent
3/13/2017
© Asst. Prof. Syed Faisal Ali
15
Vacuum Bot PEAS Description
cleanliness of the floor
time needed
energy consumed
Performance Measures
Environment
grid of tiles
dirt on tiles
possibly obstacles, varying amounts of dirt
Actuators
movement (wheels, tracks, legs, ...)
dirt removal (nozzle, gripper, ...)
Sensors
3/13/2017
position (tile ID reader, camera, GPS, ...)
dirtiness (camera, sniffer, touch, ...)
possibly movement (camera, wheel movement)
© Asst. Prof. Syed Faisal Ali
16
Vacuum Bot PAGE Description
Percepts
Actions
Goals
tile properties like clean/dirty, empty/occupied
movement and orientation
pick up dirt, move
desired outcome of the task with a
measurable performance
Environment
surroundings beyond the control of the agent
3/13/2017
© Asst. Prof. Syed Faisal Ali
17
State Space for the Intelligent Vacuum Machine
3/13/2017
© Asst. Prof. Syed Faisal Ali
18
Agent’s Strategy
 Agent’s strategy is a mapping from percept sequence to action
 How to encode an agent’s strategy?
 Long list of what should be done for each possible percept sequence vs. shorter
specification (e.g. algorithm)
Example:
 Plan intelligent car route from home to UIT
3/13/2017
© Asst. Prof. Syed Faisal Ali
19
WARNING: Might
not get what you ask for in the
Performance Measure
 Cleaning robot
 Pick up as much trash as possible
 Vehicle route optimization
 Maximize utilizations => Driving fully loaded
 Capitalizing on oddities in tariff list => Renegotiation
 Don’t include solution method in the criterion
25% of all transportation in Dubai will be smart and
driverless by 2030: Mohammad Bin Rashid
3/13/2017
© Asst. Prof. Syed Faisal Ali
20
agent = architecture + program
In this course we concentrates on the program rather on Hardware
Physical Agents
Soft Agents
Physical agents vs. software agents
(software agents = softbots)
3/13/2017
© Asst. Prof. Syed Faisal Ali
21
Skeleton Agent
function SKELETON-AGENT (percept) returns action
static: memory, the agent’s memory of the world
memory  UPDATE-MEMORY(memory, percept)
action  CHOOSE-BEST-ACTION(memory)
memory  UPDATE-MEMORY(memory, action)
return action
On each invocation, the agent’s memory is updated to reflect the new percept,
the best action is chosen, and the fact that the action was taken is also stored in
the memory. The memory persists from one invocation to the next.
Input = Percept, not history
NOTE: Performance measure is not part of the agent
3/13/2017
© Asst. Prof. Syed Faisal Ali
22
Examples of how the agent function can be implemented
More
sophisticated
3/13/2017
1.
Table-driven agent
2.
Simple reflex agent
3.
Reflex agent with internal state
4.
Agent with explicit goals
5.
Utility-based agent
© Asst. Prof. Syed Faisal Ali
23
BREAK
3/13/2017
© Asst. Prof. Syed Faisal Ali
24
1. Table-driven agent
function TABLE-DRIVEN-AGENT (percept) returns action
static: percepts, a sequence, initially empty
table, a table, indexed by percept sequences, initially fully specified
append percept to the end of percepts
action  LOOKUP(percepts, table)
return action
An agent based on a pre-specified lookup table. It keeps track of percept
sequence and just looks up the best action
• Problems
– Huge number of possible percepts (consider an automated taxi with a
camera as the sensor) => lookup table would be huge
– Takes long time to build the table
– Not adaptive to changes in the environment; requires entire table to be
updated if changes occur
3/13/2017
© Asst. Prof. Syed Faisal Ali
25
2. Simple reflex agent
 Differs from the lookup table based agent is that the condition (that determines the
action) is already higher-level interpretation of the percepts
 Percepts could be e.g. the pixels on the camera of the automated taxi
3/13/2017
© Asst. Prof. Syed Faisal Ali
26
Simple Reflex Agent
sensors
Condition - action rules
What action I
should do now
Environment
What the world
is like now
effectors
function SIMPLE-REFLEX-AGENT(percept) returns action
static: rules, a set of condition-action rules
state  INTERPRET-INPUT (percept)
rule  RULE-MATCH (state,rules)
action  RULE-ACTION [rule]
return action
3/13/2017
First match.
No further matches sought.
Only one level of deduction.
A simple reflex agent works by finding a rule whose condition matches
the current situation (as defined by the percept) and then doing the
action associated with that rule.
© Asst. Prof. Syed Faisal Ali
27
Simple reflex agent…
•
•
3/13/2017
Table lookup of condition-action pairs defining all possible condition-action
rules necessary to interact in an environment
• e.g. if car-in-front-is-breaking then initiate breaking
Problems
– Table is still too big to generate and to store (e.g. taxi)
– Takes long time to build the table
– No knowledge of non-perceptual parts of the current state
– Not adaptive to changes in the environment; requires entire table to be
updated if changes occur
– Looping: Can’t make actions conditional
© Asst. Prof. Syed Faisal Ali
28
3. Reflex agent with internal state …
function REFLEX-AGENT-WITH-STATE (percept) returns action
static: state, a description of the current world state
rules, a set of condition-action rules
state  UPDATE-STATE (state, percept)
rule  RULE-MATCH (state, rules)
action  RULE-ACTION [rule]
state  UPDATE-STATE (state, action)
return action
A reflex agent with internal state works by finding a rule
whose condition matches the current situation (as defined by
the percept and the stored internal state) and then doing the
action associated with that rule.
3/13/2017
© Asst. Prof. Syed Faisal Ali
29
3.Reflex agent with internal state
sensors
State
How the world evolves
What my actions do
Condition - action rules
Environment
What the world
is like now
What action I
should do now
effectors
3/13/2017
© Asst. Prof. Syed Faisal Ali
30
Reflex agent with internal state …
• Encode “internal state of the world to remember the past as contained in
earlier percepts
• Needed because sensors do no usually give the entire state of the world at
each input, so perception of the environment is captured over time. “State”
used to encode different “world states” that generate the same immediate
percept
•
Requires ability to represent change in the world with/without the agent
• one possibility is to represent just the latest state, but then cannot reason
about hypothetical courses of action
• Example: Rodney Brook’s Sub assumption Architecture. Main idea: build
complex intelligent robots by decomposing behaviors into a hierarchy of skills,
each completely defining a complete percept-action cycle for one very specific
task.
For example, avoiding contact, wandering, exploring, recognizing
doorways, etc. Each behavior
is Prof.
modeled
3/13/2017
© Asst.
Syed Faisalby
Ali a finite-state machine with a few
31
4. Agent with explicit goals …
• Choose actions so as to achieve a (given or computed) goal =
a description of desirable situations. e.g. where the taxi wants
to go
• Keeping track of the current state is often not enough – need to add
goals to decide which situations are good
• Deliberative instead of reactive
• May have to consider long sequences of possible actions before
deciding if goal is achieved – involves considerations of the future, “what
will happen if I do…?” (search and planning)
• More flexible than reflex agent. (e.g. rain / new destination) In the
reflex agent, the entire database of rules would have to be rewritten
3/13/2017
© Asst. Prof. Syed Faisal Ali
32
4. Agent with explicit goals
sensors
State
How the world evolves
What my actions do
What it will be like
if I do action A
Goals
Environment
What the world
is like now
What action I
should do now
effectors
3/13/2017
© Asst. Prof. Syed Faisal Ali
33
5. Utility-based agent …
• When there are multiple possible alternatives, how to decide which one
is best?
• A goal specifies a crude destination between a happy and unhappy
state, but often need a more general performance measure that describes
“degree of happiness”
• Utility function U: State  Reals indicating a measure of success or
happiness when at a given state
• Allows decisions comparing choice between conflicting goals, and choice
between likelihood of success and importance of goal (if achievement is
uncertain)
3/13/2017
© Asst. Prof. Syed Faisal Ali
34
5. Utility-based agent
sensors
State
How the world evolves
What the world
is like now
What my actions do
Utility
How happy I will be
in such as a state
Environment
What it will be like
if I do action A
What action I
should do now
effectors
3/13/2017
© Asst. Prof. Syed Faisal Ali
35
Properties of environments
Accessible (observable) : The agent’s sensory apparatus gives it access to
the complete state of the environment
Deterministic: The next state of the environment is completely determined
by the current state and the actions selected by the agent
Subjective non-determinism
- Limited memory (poker)
- Too complex environment to model directly (weather, dice)
- Inaccessibility
Episodic: The agent’s experience is divided into independent “episodes,”
each episode consisting of agent perceiving and then acting. Quality of
action depends just on the episode itself, because subsequent episodes do
not depend on what actions occur in previous episodes.  Do not need to
think ahead
3/13/2017
© Asst. Prof. Syed Faisal Ali
36
Properties of environments …
Static: If the environment can change while the agent is deliberating,
then the environment is dynamic; otherwise it’s static.
Need to worry about time
Dynamic
Need to observe while deliberating
Discrete: There are a limited number of distinct, clearly defined
percepts and actions we say that environment is discrete.
3/13/2017
© Asst. Prof. Syed Faisal Ali
37
PROPERTIES OF ENVIRONMENTS …
Environment
Accessible
Deterministic
Episodic
Static
Discrete
Chess with a clock
Yes
Yes
No
Semi
Yes
Chess without a clock
Yes
Yes
No
Yes
Yes
Poker
No
No
No
Yes
Yes
Backgammon
Yes
No
No
Yes
Yes
Taxi Driving
No
No
No
No
No
Medical Diagnosis System
No
No
No
No
No
Image-analysis system
Yes
Yes
Yes
Semi
No
Part Picking robot
No
No
Yes
No
No
Refinery Controller
No
No
No
No
No
Interactive English Tutor
No
No
No
No
Yes
3/13/2017
© Asst. Prof. Syed Faisal Ali
38
Running the agents and the environment
procedure RUN-ENVIRONMENT (state, UPDATE-FN, agents, termination)
inputs: state, the initial state of the environment
UPDATE-FN, function to modify the environment
agents, a set of agents
termination, a predicate to test when we are done
repeat
for each agent in agents do
PERCEPT[agent]  GET-PERCEPT(agent,state)
end
for each agent in agents do
ACTION[agent]  PROGRAM[agent] (PERCEPT[agent])
end
state  UPDATE-FN(actions, agents, state)
until termination (state)
3/13/2017
© Asst. Prof. Syed Faisal Ali
39
Question & Answer
3/13/2017
© Asst. Prof. Syed Faisal Ali
40
END OF LECTURE 1
3/13/2017
© Asst. Prof. Syed Faisal Ali
41