w1 what is an agent, utility, rationality also w2 normal form games Flashcards
this is the introductino to ai. we learn about what makes an ai agent 'intelligent' and learn about what an agent seeks out, rationality, preferences, utilitiy and risk. we learn about the basic components that goes into artiifical intelligence trainning, week 2 is about games and introduction to nash equilibrium
what is an agent
computer system that can make independent actions on behalf of user/owner. user does not need to tell them everything that needs to be done
it has its own goals and behaviours and it is autonommous, adaptable and modiyable
what are the four axioms four that involve outcome and preferences
completeness
outcome 1 is preferred over outcome 2 (vise versa) or about the same
transitivity
if a < b and b < c then a < c
continutuy
there is a probability p that the outcome b is equally prefered to outcome a with probablity p and outcome c with probabiltiy 1 - p
independence
if o1 is less prefereed than o2 then a gamble involving o1 is less prefered than gamble of o2 IF the chances of getting o3 are the same in both gambles
what is utility
the amount of happiness you get from consuming goods and services
what is rational choice therory and how does it relate to utility
Rational Choice Theory is the idea that people make decisions by logically choosing the option that gives them the greatest benefit or satisfaction, based on their preferences and available information.
It relates to utility because utility represents the measure of satisfaction or benefit someone gets from a choice. In simple terms:
People are assumed to act rationally by choosing the option that maximizes their utility (happiness, profit, or preference).
For example, if choosing between two snacks, a rational person picks the one they enjoy more (higher utility).
what are the three diff ways agents tolerate risk
they can be risk adverse, neutral, seeking risk
seeking prefers higher expected utiltiy even if they come with less certainty
what do we consider for agent design
build agents capable of independent, autonomous action, so that they can successfully carry out tasks we delegate to them
what do we consider about agent society design
build agents that are capable of interacting (cooperating, coordinating, negotiating) with other agents in order to successfully carry out those delegated tasks, especially when the other agents cannot be assumed to share the same interests/goals
what is a multiagent system
a system that consists of many agents that interact with each other,
agents can interact with different goals and motivations
what types of games are we dealing with in ai 2
simultanous action (ie rock paper scissors)/normal form and turn based (like chess)/ extensive form
what things do you need in a game
set of players, strategies, outcomes and game rule and payoff
how to solve a game
dominant strategy equilibirum,
nash equilibrium
what is the differenbce between better and best response
better response is favourable but not the best but best is best
what is dominance
stratagy a dominates b if a always results in a better payoff than by playing b no matter what other players do
dominate stratgey doesnt always exist in a game
what is a pure strategy vs mixed strategy
pure strategy is deterministic action of a player
A pure strategy is a deterministic approach where a player chooses a specific action with certainty every time a particular situation arises.
mixed strategy is probablity distrubution over pure strategies
mixed Adds an element of randomness to choices, making the player’s actions less predictable.
Useful in games with incomplete information or when trying to counter opponents.
Example:
In Rock-Paper-Scissors, a mixed strategy could involve choosing:
Rock with 50% probability,
Paper with 30%, and
Scissors with 20%.
what is pure nash equilibrium vs mixed nash equilibrium
each player has chosen a strategy and no player can benefit more by changing their strategy while the other doesnt change theirs
mixed nash equilibrium is when palyers could used mixed stratigies
A Nash equilibrium where at least one player uses a mixed strategy, choosing between actions probabilistically. The outcome is stable when considering the expected payoffs.
In Rock-Paper-Scissors, the mixed Nash equilibrium is for each player to choose Rock, Paper, and Scissors with equal probabilities (⅓ each). This ensures no player can gain an advantage by deviating.