Merken

Game Theory to the Rescue When Hard Decisions Are to Be Made

Zitierlink des Filmsegments
Embed Code

Automatisierte Medienanalyse

Beta
Erkannte Entitäten
Sprachtranskript
next class would be the URI to the rescue when father she's in ought to be made abilities and gentlemen and welcome our next speaker I extend the hand of few few probably wait until the top students would that be ok on and going to talk about the theory of the so wall let me introduce yourself myself off so my day job I developed and I can solve for what will logically coughing and I review technical solutions and work on I'm reading high-tech start-ups together with the companies if it's something that helps both companies so that's what I'm doing to time that the other challenging tech sort of up to me and maybe we can help you on of also of 1 of the organizers of the conference here and a profit-sharing and active and the mom would you be community like to talk to me if you have any comments questions you need to know where later following in 3 areas on we will only be able
to stretch the surface of game theory on today so there will
be no according to last chance that we now and I hope I didn't give you an idea of what the because it's it's a bit of a life beyond the else else spin on it's it's it's a glance into a different life a little bit I stumbled across using it just like my coincidence that the meaning of the you will actually go on call for proposals they're running on and yeah people think collective on unfortunately so and game theory is very human concept now it's not check this has a background of in economics philosophy and and I think you can imagine yourself like falls in much more active complicated so many options positions and required to make decisions all the time and sometimes that can be really difficult on was stressful and the very often worried that we do right to make the right choice the following didn't pick the right Pokémon to win the next cell and that of that's so little knowledge of game theory can become really handy on my only coding life solving of some problems on the schedule of making the best decisions you can but probably not only the best decisions for yourself also the best decisions for you and the people around will to that later so to think about how do you make decisions how to make judgments that they really know don't like the pros and cons that wake them up and some people on some people won't be very much along the lines of the stomach as the same Germany common so you think about the user decisions affect others and do you care about on and and and and this the only game theory is going on you have to anticipate the decisions that others do while you're interacting with them and how they impact you and how you decisions impact the others on the yeah this is basically like the vicinity of the infinity and but actually I stumbled across that are so when trying to over all beginning messy accounting data from lots of very popular on online platform for videos and make can the check what the what's in there but quite sometimes they just make match some words stuff we don't really know and we tried to come to the people who gave the video material audio licenses and all so if we could do do it manually but it's like really lot of data on it would be like many people on and not effective because we're so it's really a penny accounting business and soul we also use like systems like for example if the elastic search and really ElasticSearch who knows ElasticSearch yet most of you so you know you can also get a probability back on the quality of the results he all you get home indicator whole uh ElasticSearch things how accurate results and so if we have something to work with and so on but we have like select customized systems become directly the database we have some some some funny search stuff I don't want to go to the printer but just like I was restart like in a combination of all OK always expert systems I can ask from and probably 1 expert system will improve my input and it'll put it into the the other expert-system system I can even get their input and so I was getting very confused like and so I thought about what you and inferior because I can't really use is else cells there will only so we use all and so on I remembered Game theory and give it a try on solid means to use you want to game theory and I hope it will help you to make decisions and you coding or whatever the problems facing for like basically maybe just in your life who may be better of so them freely very often it is see
picture of game theory you will see checks and yes but it's actually a very bad example because it's game tree of all playing games that again you can leverage false in talk but not 10 games of
interaction between humans which we can also like that's it becomes basically on imagine it's it's also like again like simplified version of interactions we can regard this as a is and we know things like chance of gold what tic-tac-toe but what determines the on every player almost every stage of the game everything you see can see
somewhere like so if basically it
would be like playing poker with all cards on the table the time on and basically it's just like all of that you can calculate all possible results and of course we don't because our brains are capable of using computers a lot more capable of calculating all the results imaginable of all of them for each and every move in the in time so it's imagine like in theory having all the cards on the table on if the informant information is not not sort of not having all the cards on the table some information might be here on and is there a way almost to determine on how you were still can evaluate chances can make built up basically build up a strategy and basically this is all I was doing on then there as a as a dual I believe I will chess against the Department of my mother if I just chose the right color in the beginning on after I lost many many games they came to the conclusion that still the case very unlikely but I think that the average chess player is also trying to find a strategy on and they're anticipating what the next move of your own and like the old volume basically calculate each and every possible result you anticipating what you play it will do you're anticipating his strategy and those this is getting us a little bit closer to inferior and game theory was not just like invented on overnight on the level of plants and 2
of the interviewer history and it also like 1st mention this goes back to the 17 hundreds on will break letters in the guys from Britain on there is a French don't I wrote the paper on the 3rd leading in the direction on and then in the beginning of the 27th century things start to pick up with the answers to the level and we can talk and similar anybody under minimum mainly ability to lead shot use of which actually concerns slate of also whole applying our that's that's your feet
set theory in chess game on that and then in the 19 the late 19 twenties on the same starts to really pick up the paper form Hungarian all all scientists shown from my and
I think he moved to the the US later and he wrote another to a 2 week because that's excuse which visible like theory in the book random In this paper we noticed a lot and in 1944 he published a book called the move turn our theory of games and economic behavior and so we see already on the game the of thing is very close to economics and
usually we think economic that people care about economics audits just of business it's not like programming is much more better so on we don't probably look into Fiori's their and the the the challenges this can be built the mathematical model can be basically you analyze calculate results of of people's actions on the decisions the players might do and basically that's probably most famous being in the 19 fifties it's like prisoner's dilemma dilemma and so much came up with the national equilibrium work and and gets very famous because like John Nash life also beautiful mind moving so he's very popular and that's the way this is a picture of
Don Norman Hungarian scientists so background in
game theory is an economics political science and psychology of use well and logic computer science but do it with biology and it's mathematical approach to relapse simulations they situations that involve 2 or more at each decision maker you for example has a number of different actions that might take on and the ultimate outcome was not just can depend on your actions it also depends on what the other people that and so on you're forced to think and come up with a strategy on and this is the basic so let's start with a really simple game that's something easy and practical called zero-sum game and a zero-sum game is the winner takes it all and the other is going to be of lost and I'm really happy so many people and so crowded because it would like to play to finger more so please partners in the 1 on your right so it sorry sorry inferior the to people who have had so to the and is a very simple gave it a partner and the you basically of your choice have 2 choices show 1 thing or or to finger and you have to decide who is that even player who still player and then just like as references you are that play with your partner you show your fingers at the same time and is the result of even even player wins and the result was a lot of things with so you basically have a 50 50 chance soul and the 5 minutes I will be back until results and this is a simulation of the of the right 5 minutes so all right so in this and I don't think about it and to so OK we continue now there is material upon with I didn't mention this simple game would make this so lucky what OK but basically what you'd be like the mathematical representation of what we were just like doing all that we call this like a strategic form and basically and called you know you
are the plot and the have but 1 the I don't know why I so day of like what we call a strategic form and so and this is
like basis you 1 can calculate almost like matrices so on 0 sum games 1 when 1 user so very competitive and this another mostly on 1 0 sum games on more of symmetry and everyone games on the for example like doing the same game and a little bit different forms so to non-zero same game but please don't get as excited as I have only 15 minutes left and I have to get you to present dilemma Michael so that's great but the positive weight on let's look the the same game books on
a similar game so if you still the same kind of thing during each of you get 2 points and if it differs you only get 1 point and this is basically a boat world of cooperation you wanna give it try
not to get and so long yeah yeah so you can like interactions and this is a very simple convection put into a matrix and we can work on there is there are games with perfect information as I mentioned chest goal
tic-tac-toe so everybody knows about everything all the options and outcomes for the other players and their are games with incomplete information basically so that's like the 2
and is next to the selected classical game theory is the prisoners dilemma so what is the development of the prisoner's dilemma imagine there like to criminals on there already with the police the police has the problem and
imagine having them into rooms and the police the strength of of police have war and the baby interrelated volunteer
and do the dilemma is each of the 2 criminals doesn't know how the other criminal decide so if you know if you call operator I mean you and if you ever have seen our model of US all all time series it cooperates we can make a deal with the DA and here you can guide you will get free so all the thing is that you know if people operate on you might get free you don't get any sentence let's see what it is you're calling you criminal calling you incorporated on the conspirator also is cooperating on because then both cooperating and it's not that it'll all criminals cooperating they're just like saying yes we do so sorry and everybody will get free of course they will might be able to get a sentence to the reduced sentence now on and the other outcome of courses just stay quiet and probably the police will be able to prove anything more probably overnight only like a minor crime on and then you might only end up with like the 1 year sentence and this is the prisoner's dilemma we have been
time this my and to group and now let's get into all of our strategic form itself so that
they might to wish to what is on the social because you're really you really screwed you go to jail for 20 years and you don't want that and if you
both people quite we assume yeah OK you will get some sentence but not to the full crime you from might be and and in prison for 1 year and how can you actually decide because this tool is a dominant strategy is what we call them in game theory its dominant so what what always the best outcome you should choose and actually confessing as and because you you would give go to jail for 1 year for 5 years but never for 20 years so the best choice and also confessing is only the the the the only strong Nash equilibrium on the game and the and the equilibrium is basically on when there are no more gains from no 1 in being in the games so basically nobody wants to take any action because he can no longer gain from and this is of course in this simple game just like
1 1 round but just like imagine there's more criminals involved like gang of flies
and you have to think about what to the on off people decide on and this is completely different game because all of there couple they're
very close and they will never really give each other but if it's just like some random guys who met on the street on and you form the gang to break in you might tend to decide on its better confessed because I don't trust those guys and also I don't want to be a part of that it can be apart from them but you don't want to puke bought you're all
girlfriend or boyfriend for so this is where things can get complicated on and how can we calculate stuff like that on Wikinews algebra artist many approaches Bayesian Nashequilibrium so we can do just like want to constitute a simulations on 2 to to get some results on the yeah and there's some of material you kind of look into the there's a of pattern to gamble on it and just play
around with and I think that this is realized from both from a biologist market in terms of evolving strategies and for iterated prisoner's dilemma tone and I think it's very descriptive of of columns and everything so had really but a lot of fun reading all the doing research for for this talk on and there's also an from central because there are a game theory on causal from method Jackson which I think is really going on very often talked stuff and and that's a helpful way all was able to give you a glance into the game theory you own a little bit different approaches thinking on that and then you had a good
time and if there's time left we can play some more to two-finger moral yes you want you if extent yes OK so 1st of all I think very much for your point and ask you did you go back to this library where the matrix on that so on here it's very clear and think a little bit out Ali and and and effects yeah so here I think it's very clear that the optimal for bolt the door and will use that they bought a surf minimum sentence is that the vote keep quiet but
you said that the should leave room is basically that if they would compress it basically like the the like that that best strategy is to contest what was there's not bad strategy not but the thing is you
can calculate like I mean I I I don't understand so whatever learn why would you would you would prefer on me being imprisoned for 5 years instead of 1 I don't know I don't understand why you would prefer being can prison for 5 years instead of 1 I mean I know they have good food but the thing is you don't know what the other 1 does that the problem you don't know what part of the 2nd that's the whole article that's thank you for the question that's the key you don't know what the other particles that are and so you make you need to make a decision on basically what's the best movie but you don't know now that's why that's why this is so that's why this is the dominant strategy if it minimizes the loss you will never go to 2 prison for 20 years if you confess that select because I think the minimum of of the maximum you can 1 of things that you so basically if you would keep quiet you would be secure use of very high agreement and yet I think a few provided by not like having you also so 1st of all a side note I think it's interesting how people standing up and being very loud it kind of is the tragedy of the commons another got the concept but my question really is we are able to solve your data problems using game theory I was say that I was able to improve it on because like the thing is when when when lions like like putting like information all in into the next system to improve its so basically on the head and they mention topical supposedly brief can also on defined on the 1st floor and if you know like in the game will be minimal I I think I have already remember it was like 5 per cent so so say we we're we think the information of 95 % accurate and and after like for all selected threshold of iterations so we will go on forever communities just have 1 letter we might never we will never get the results of that the that the 2 things they can do with federal usually on say like OK or with the percentage on probability or like iterations and then just like take the results of this was very good for improving and interaction what systems we have a few more questions we don't have much time so try to be shot up but again this is not we have 1 question you have you and use the Python axle roads uh knowledge and our view but I've seen it online doing everything that we use that he was developed by a friend of mine and unacknowledged yes it's been very interesting in kind of raising interest in Python programmers in the in the standard that is 1 of the few and had a strategy that are not actually like actually I was quite surprised because I thought when I when I always did research on and when I was trying to solve the the problem on I was going to follow the quite game theory I expected like 10 libraries and the results were hardly any on that surprised because usually the library for everything in life and so yes but it it it it be great if we could pick up on the like helping and building and stuff like that and also like on just like that game theory to all tools that thinking when we're trying to solve problems because and also played the cycle of being like programming a lot on if you tend to think very deterministic here this is the truth this is the end you all opens of you might be a little bit more flexible finding that that practical solution that that's my well we're out of time and so let's think have began the
Flächeninhalt
Selbst organisierendes System
Prozess <Informatik>
Klasse <Mathematik>
t-Test
NP-hartes Problem
Quick-Sort
Physikalische Theorie
Computeranimation
Lesen <Datenverarbeitung>
Resultante
Stereometrie
Bit
Ortsoperator
Schaltnetz
Zellularer Automat
Systemplattform
Computeranimation
Videokonferenz
Spieltheorie
Flächentheorie
Indexberechnung
Auswahlaxiom
Gerade
Videospiel
Expertensystem
Matching <Graphentheorie>
Datenhaltung
Spieltheorie
Systemaufruf
Physikalisches System
Konfiguration <Informatik>
Entscheidungstheorie
Arithmetisches Mittel
Scheduling
Codierung
Wort <Informatik>
Netzwerktopologie
Spieltheorie
Versionsverwaltung
Interaktives Fernsehen
Computeranimation
Resultante
Bit
Computerunterstütztes Verfahren
Physikalische Theorie
Quick-Sort
Computeranimation
Übergang
Chipkarte
Computerschach
Spieltheorie
Mittelwert
Strategisches Spiel
Vorlesung/Konferenz
Kantenfärbung
Information
Spezifisches Volumen
Tabelle <Informatik>
Bildschirmmaske
Computerschach
Spieltheorie
Tablet PC
Von Neumann, John
Physikalische Theorie
Computeranimation
Übergang
Resultante
Videospiel
Gruppenoperation
Spieltheorie
Thermodynamisches Gleichgewicht
Von Neumann, John
Rechnen
Physikalische Theorie
Computeranimation
Entscheidungstheorie
Gefangenendilemma
Spieltheorie
Gefangenendilemma
Mathematische Modellierung
Randomisierung
Optimierung
Chi-Quadrat-Verteilung
Resultante
Subtraktion
Mathematik
Selbstrepräsentation
Gruppenoperation
Zahlenbereich
Von Neumann, John
Mathematische Logik
Computeranimation
Entscheidungstheorie
Normalform
Rechter Winkel
Spieltheorie
Diskrete Simulation
Simulation
Informatik
Auswahlaxiom
Bit
Subtraktion
Gewichtete Summe
Gewicht <Mathematik>
Matrizenring
Ortsoperator
Gewichtete Summe
Spieltheorie
Plot <Graphische Darstellung>
Kombinatorische Gruppentheorie
Computeranimation
Gefangenendilemma
Bildschirmmaske
Symmetrie
Normalform
Spieltheorie
Basisvektor
Matrizenrechnung
Perfekte Gruppe
Punkt
Spieltheorie
Gewichtete Summe
Spieltheorie
Interaktives Fernsehen
Information
Computeranimation
Gefangenendilemma
Vollkommene Information
Spieltheorie
Klassische Physik
Spieltheorie
Information
Softwareentwickler
Computeranimation
Konfiguration <Informatik>
Gefangenendilemma
Nichtlinearer Operator
Informationsmodellierung
Zeitreihenanalyse
Gefangenendilemma
Vorlesung/Konferenz
CAM
Computeranimation
Kappa-Koeffizient
Normalform
Gefangenendilemma
Gruppenkeim
Computeranimation
Gefangenendilemma
Kappa-Koeffizient
Spieltheorie
Gefangenendilemma
Gruppenoperation
Strategisches Spiel
Unrundheit
Thermodynamisches Gleichgewicht
Auswahlaxiom
Computeranimation
Kappa-Koeffizient
Spieltheorie
Gefangenendilemma
Computeranimation
Algebraisches Modell
Resultante
Subtraktion
Bit
Physikalischer Effekt
Spieltheorie
Term
Computeranimation
Gefangenendilemma
Spieltheorie
Gefangenendilemma
Diskrete Simulation
Mustersprache
Strategisches Spiel
Bayes-Netz
Soundverarbeitung
Matrizenrechnung
Kappa-Koeffizient
Bit
Extrempunkt
Gefangenendilemma
Minimierung
Strategisches Spiel
Programmbibliothek
Maßerweiterung
Computeranimation
Resultante
Programmiergerät
Bit
Einfügungsdämpfung
Extrempunkt
Iteration
Interaktives Fernsehen
Computeranimation
Data Mining
Spieltheorie
Programmbibliothek
Schreib-Lese-Kopf
Videospiel
Schwellwertverfahren
Sichtenkonzept
Physikalisches System
Entscheidungstheorie
Gefangenendilemma
Mereologie
Dreiecksfreier Graph
Strategisches Spiel
Partikelsystem
Information
Schlüsselverwaltung
Standardabweichung

Metadaten

Formale Metadaten

Titel Game Theory to the Rescue When Hard Decisions Are to Be Made
Serientitel EuroPython 2016
Teil 47
Anzahl der Teile 169
Autor Hendorf, Alexander
Lizenz CC-Namensnennung - keine kommerzielle Nutzung - Weitergabe unter gleichen Bedingungen 3.0 Unported:
Sie dürfen das Werk bzw. den Inhalt zu jedem legalen und nicht-kommerziellen Zweck nutzen, verändern und in unveränderter oder veränderter Form vervielfältigen, verbreiten und öffentlich zugänglich machen, sofern Sie den Namen des Autors/Rechteinhabers in der von ihm festgelegten Weise nennen und das Werk bzw. diesen Inhalt auch in veränderter Form nur unter den Bedingungen dieser Lizenz weitergeben
DOI 10.5446/21088
Herausgeber EuroPython
Erscheinungsjahr 2016
Sprache Englisch

Inhaltliche Metadaten

Fachgebiet Informatik
Abstract Alexander Hendorf - Game Theory to the Rescue When Hard Decisions Are to Be Made Sometimes it's hard to decide when a something is really done or cannot be improved further. **Game theory** can help you to make complicated decisions whenever you encounter flow problems. ----- Game theory is "the study of mathematical models of conflict and cooperation between intelligent rational decision-makers." In our use case we had to match data for accounting: - the data was not always clean but we had some extra tools at hand and a complex system to make good guesses. Nevertheless it was hard to decide when to give up, some records were just not processable. Finally we used Game theory to make the decision.

Zugehöriges Material

Ähnliche Filme

Loading...