You will be answering parts (a)-(c) for a smart home assistant like Google Home
ID: 3812950 • Letter: Y
Question
You will be answering parts (a)-(c) for a smart home assistant like Google Home or Amazon's Alexa. These agents listen for keywords (e.g. "Okay, Google"), and when addressed, respond to questions and perform tasks (e.g. adding an item to a list). a) Develop a description of the task environment using the PEAS description i.e. Performance Environment Actuators Sensors b) Then describe the environment according to the following properties: fully vs partially observable deterministic vs stochastic episodic vs sequential static vs dynamic discrete vs continuous single vs multi-agent Note that in some cases, both answers might be correct. Justify each answer to the task environment properties with a one sentence explanation. c) Suggest the most appropriate agent design by choosing the most appropriate of the following agent types: simple reflex agent model-based reflex agent goal-based agents utility-based agent Justify your answer with a one sentence explanation.Explanation / Answer
PEAS
Performance - An action agent which can perform any given sequence of percepts. ( consider designing a truck- safe, maximize profits, obeying laws, reach destination etc are the percepts.)
Environment - Shouldbe able to move on roads, free ways, traffic, should notice pedestrians, weather, load material etc.
Actuators - wheel, accelerator, brake, horn etc.
Sensors - GPS tracking, guages, engine, sensors, keyborad, vidoe, accelorometers.
2. Fully observable Vs Partially observable : agents sensors gives its access to the complete environment at each point of time.
Deterministic Vs Stochatic : Next state of the environment is completely determined by action executed by the agent and the current state. If environment is deterministic except for actions then the enviroment is stratergic.
Episodic Vs Sequential : Agents actions divided into atomic episodes. Previous decisions are not useful for the current decisions. or actions.
Static Vs Dynamic : if the agent is deliberating the environment is unchanged.its is semidynamic if the environment itself doesnt change with the passage of time.
Discrete Vs Continous : Limited number of distinct, clearly defined percepts and actions. represent or abstratc or model the world.
Single Vs Multi : An agent operatin gitself is single. If the performance measure is done based on othe ragent then it is multi.
Appropriate agent designs
1. simple reflex agent : It is fast but too simple, No memory, fails if environment is partially observable. Eg vaccum cleaner.
2. Model based reflex agents : Current world state, Modeled on how the world chances and actions change in word.. It can work even with partial information, It is unclear what to do without a clear goal.
3. Goal base agents : They provide reason to prefer one action over the other. Predicting the future is important, need to plan and search.
4. Utility based agent : Some solutions to goal state are better than others, Which is best and which combination of goal is preffered.
When all the different type of agents are considered the best one utility based agent. It has few reasons like
It is based on goals and utilty of the world. can also choose the best combination.
Related Questions
drjack9650@gmail.com
Navigate
Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.