Merken

A Gentle Introduction to Data Science

Zitierlink des Filmsegments
Embed Code

Automatisierte Medienanalyse

Beta
Erkannte Entitäten
Sprachtranskript
hello everyone things for coming to my talk and in new and it was not in that so I guess it was a bit challenging for you to know what was going on here In this as the title says it's this 1 story Argentine inversion tool for data scientists believe an intuition there's some form of some stuff I'm not sure cogently
is is being gentle the goal is not that you really understand everything on that there really all the concept is more to get an intuition our 1st where data science from from or where artificial intelligence come from that it's quite related it's it's my area and 2nd justices some applications for which can be used it's not really really a technical but there is some parts of
the Russian parts of the beats I read more about formalizing understanding some some concepts so why on a time working currently not dating site
will I be the mandate assigned is the I have amassed 300 so intelligence to leave some context of disease data science but also the quality of the human brain and things that are more consider that efficient in than than the the science and most the day and focus ambassador cured Europe from the properties of foundation that feeds a sponsoring many of the baby the projects or all the price of the product in the by date of Romney's aunts and focused in in for what is also organizing the baby the events so you can do find me most of the time and then for both later on if you are interested in some of the projects so little data science and by phone or even within by from some of them are heroes and other technologies like Julia Austin so it's soccer with that of the human brain and I would like to lower the Canova as more not experiment but but just try through it'll make you
visualize have a tool to really be able to show what they want all the it's kind of a silly example it's probably and this was something you more as as a meditation exercise than than as an antigen intelligent but what what they want you to think is that that's your own eyes I kind of like programmers that imagine that they're like 20 where 20 pixel commerce and this image on the screen just like the number 5 there is nothing magical meat and it's not like anything that's that fancy wages like at fame was made from a famous dataset name means is quite common used as a half as a means a different there isn't in in in the learning and other techniques on they went a little is just to imagine that that your eyes are just come and that users perceive this this 20 by 20 pixels so when you arrive with the deriving you're writing your readiness issues like these 2008 and information that can have values let's say that 0 4 1 and as for white and 1 for black
they give you imagine these than full information it goes into your brain like it that it feels arrived in the eyes these days fears you can you can see that there are these channels of neurons will talk about me that what Neuron-2 now in a 2nd but there are these channels that would just always provide a this 20 by 20 pixels imagine 20 like 20 neurons just taking this bullion of the brain where the visual cortex cortex is the visual cortex is defined as the name visual say that it's it's mostly focused on their on the understanding and the recognition of of images so much to example we put this number 5 just like zeros and ones to say that get to your eyes and then at the end you get to the visual cortex so some years ago this article thing this this brand I think it's something quite obvious whether can distort it still worth mentioning then that the the human brain is just like a network of neurons and and neurons are basically I just like the cells that transmit electing information from 1 side to the other triggered by the whole human brain the whole of human mind everything we know or we think we know it comes from from these neurons just inviting enough interacting among them in a massive scale and all the interactions are what really making a lower order lower perceptions everything OK it's growing up very intuitive when you think of a single and neuron life having all these powerful but when you combine thousands of them it's really July that is like if you mightiness leach switch is actually not nothing intelligent but if you have our microprocessor of a computer DeLex Witiges like zeros and ones is activating from signal symmetrical signals on and at the end it's it's a awful awful when you can do without the computer so it's it's exactly the same thing so all Hubble
cost and where we're tool researchers quite a bond for for the period that they did an experiment with with this book that he did I'm using contribution but I think he wasn't really happy about that they basically connected some sensor for his brain people there will be initial part of the visual cortex as I was saying my it's not my example you're getting these information years unwanted like perceived by your eyes as if they were a coming and then they will to the visual cortex and what they were staying is this 1st lady of the visual cortex so what happens when these neurons instead of being just the channel that provided information the estoppel makes 1 with another so here start build these networks all their ideas lyotropic early discussed in the visual cortex is recognizing circles like the 1 you can see in the is broadly recognizing squares recognizing different shapes on the the book describing how the got looking at stage for many hours looking for Citizen waiting for neurons 28 and the funny part of the story of a know if you can see in these thing it can be seen in the in the screen is those obvious that alters lights and where to start transferring paper and they have like the shapes on that the funny part is that they in the neurons of the count of the visual of the visual cortex of the cat differently actually activated
not because of the circle not because of the triangle not because of any of the shapes but because of the edge of their of their and like the more the of the of the state but you can see these line here and there were no that was really what they may make the decapitated knowledge so all in the 1st layer they take away of this is that in the 1st layer of these no rows of these neurons connecting 1 with another actually what it's going on is that is that they reactivate for a very very small small items the say models like ages in different in different rotations on and that's another important
an important experiment and then I'll head is based in there the In theory is that these connections in the brain actually our Arab besides some of them that are of course when you're born you already get some some connections in in your ranks but some of them operate that just because they get activated so if you gave people watching the same which we've seen before for many many times with these theories say that there is a complication i'd a spectacle to even read that is so you have them in the slides but the whole point is that if you keep activating you you have like an impulse you'll see like the space these phase phases Phase 1 time another time and and one's and once more then at some point the neurons that are activating that when the 1st time you see them they become stronger and stronger and their there they're kind of recording then the information in the brain so it's like something that you've never seen 1 activate much now runs on something that you see often they will much more orange that something very interesting part of the and like memory the learning actually visited everything is start from from these everything is based on on seeing something and then having it call it was some connections from the neurons are there and then being able to reproduce the so we'll see this kind of like a bit of the history of the human brain what's in the human brain novel gold weakened the party like cold computers I've kind of cloning going all those things I start this minus sign to get this dishes like a linear regression I assume that most of you know a linear regression issues just like I watch from import let's say that i you told me your age and only you or your ways of say and I want to break your 5 so usually digital divide each of the values and then I that I had them to wear that are linear regression financial 1 you find there is a correlation between them and this is what the thoughts on that you have that so this something quite simple but at the end it's just like the way you you thing has these as neurons is a way that they bring your content have always try every every input you will have will have more or less importance the annual back altogether will have like are a a single value and it what's very important in this is that deceptive it function what it does is penalizes the day and the end what you have is that the output of this is uh 0 and 1 this something similar is lost and logistic regression if you ever heritable public that indeed the science but is still quite powerful if you start spherical you the gold always it be not binarize I want them bond much of that but it's not usually taking alive like positive plus negative even making positive 0 or positive 1 negative 0 which usually used in in the sigmoid functions does as a as a reference that would be the way it is is do those kind of the same exact thing but you can take the derivative of that and it is very key because in in optimization problems when you want to build these networks and these and you need to see the which weights are the would ones being able to the rate that it it really really useful useful for you really optimise Markowitz what hopefully it works and what covered this briefly Hopfield words are just like networks of this artificial neurons that everything is connected with everything and a nice property it's it's kind of something super super simple it's nothing really really complicated was in on just a linear regression you can implement in a single line of courts and then you just connect them among them the nice thing is that this has the property that its name and as such as
its its color associative memory and what it does is like you show as I said before with the number 5 you keep showing images and then did this network is remembering the images in these in the same weights in the weights of a linear regression at each learning there it's lending these these weights in this example would you can see is like that of the original image in in the left is the image that you shown that to the network and then at some point you show image that the it's similar to this 1 and the network enough remembers the optimizes in a way that was uh after several iterations with you that is the original image and tricks can now it can be used for example tool to recover corrupted local means work that the images because whatever you you've shown in were before at some point will win will be back to you that's kind of the
point of the he if you show all these cleaner world to land and urine and of course you need that I have to fulfill networking men and of course you need to have a proportional number of pixels so all of of neurons that are able to restore all that information but if you show all this at the end and you check the weights you represent way amounted to have something like this it's really it's kind of stuff so but you can see that this is like the shape of like the average number and in every of the pixels at the same time we have the average of the numbers that up that that are activated that our that have values of that pixel for for example you check on the side of a here on the on the right and the top right corner you will see the number 7 No. 5 ago going that direction will you see for example in the in the top left you will see the number 3 because the number 3 actually it ends up in that I mentioned this actually it's it's kind of more our CEO thing but but this really shows all the weights of the linear regression the worst of our before actually by keeping information on this images they are able to recover the later after supplement worse the Boltzmann machines that this has to be there on the basis of the bill the learning if you of the 1 that thing is you look modern machines are the same with the difference that In the previous example all the networks are the same as the pixels you're getting so if you're giving like 20 by 20 pixels you have these network these these neurons and this is exactly what you'll want you'll have to store the information in this case is you have like hidden that neurons that they don't cover direct exposure of the dual you won't have like the the initial perception but will have like I'm an indirect perception and it actually much more powerful because the is this your work as a generative network as an active network is something similar what we've seen before but the idea is that you've got
the model that it learns some parameters in this case is like the weights of the neurons but if you think it as hours of ocean imagine that I want all I want pool generate data that it comes from the distribution of the height of the people in the initial or the ages so at some point so I can have like I get that ocean distribution I say OK the mean of the age the people in the room is 13 to say under standard deviation is 5 for a canister generating samples what their samples and getting it would be like there real samples like if it was real people that I knew the age of the people like the same as if I had the data from people entering the from show what is quite important if
it is if you will would be much recognition you gonna start making like these models these in these networks estoppel and patterns it is something that needs men like the dream think it's an experiment by will the bear started to show we might just tool to 1 of these networks and what was happening is that then the said like OK now tell me what will they give me the the that I I was staying with pages and the show me more and more data and I'll show all are those who learn them yet what the data you've already seen is generated new 1 and this is what they do I'm not sure exactly when the used to train this network but this is something generated by a computer that just in some images I guess they they saw some money models for what you see here this Cuchulain than exactly what the yard looks like people people may be are temple in right so what it so what it can actually shoal show in what's based on the matter that that he's is been seen that in just emphasize it the example if you tell me age and I start to assume that the distribution is this and then it's the same numbers like I I think are normal it would be like at an authority door and 24 8 so this is basically what it's doing but with the images which I think is quite quite an our understanding the problem
with artificial intelligence usually is that find these models they are super cool they're quite amazing but at some point you realize that there are and the complete which it means that the computer would take line casts from the same time as the barn door to the present day to compute all the information for that like the optimal solution day in practice what is being reduced is these restricted boltzmann machines that they have some properties that they don't have all the connections for the tongue done really have these properties in the same way it was proved you that for for a vote on machines but it still have mostly everything it's kind of an approximation but with approximation and this 1 can be trained from how there's something mixed optimize them and train them them easily which is quite quite nice to make things in practice in practice you such as 1 of the Internet were would delicious couple ideas of networks and and 1 point to the other and as a
said before different layer with the card was used
image you can see these in the in the left you have these very small but utterances for themselves they are not that useful but when you start to combine them in the new and the neuron name in these networks on this on the next layer you're actually estoppel have and I used to have and also thought of things so it's kind of a sequential thing like a cascading thing that you start with the and not by the neurologist operating a small buttons at some point you have an
idea from when you combine 90 of about half a user and you can construct anything you want if you have the norm of in of number of
orphanages out of a very old book very briefly about that just to give you an intuition on also can be on the the whole this coming on well would you need in in the the science is useful from there their so you have some data whenever you have some data you know that let's say that you know that this is an idea you and when you're looking for so you're comparing the the answer with the prediction of your network and what there is mismatching if you're trying to break the caps and you'll have the steady much you compute the difference between this and that this is where growth and it for a linear regression it would be much complex but imagine the Georgia of that you can really compare the images that your modeling generating we've images that you want to see and and based on these over you just the and I was talking about before what is that even if all these functions from the from the you can they've a them because they are just linear regression with the sigmoid function that can be the related and because you know that the derivative derivative is showing you where so you can mortal to the minimum point to the point with less error so you start up to this field and you just say OK you the radio sold a function that you're using was to make and then you see that the the drawing down in that direction so you go a step in that direction and you keep iterating at some point you end up with a place where where that worries me if there is a minimum in the direction that your model actually performing as expected if we're being able to reproduce god or whatever it can it can replace under look is my layers so you always start doing that in the in the last layer you check their around and you'll you avoid the in another
course if we might cost this is actually from a real experiment they talk I think it was 40 thousand and CPU was on the train the sort of know raunchily you previous frames from book reviews for a whole week and to see what was there day of the internal representation of these models were land and what they found were what they say they hominid speech is discussed in 1 of the newer world like in that age what it learned and growing makes sense because in Newport Beach full of cuts show at the end
you thing on on this whole thing on hold the brain is learning the neurons are changing their transformations based on information we see if you can see the brain as a thing that that did really there just a set of patterns that have been learned like these shapes that you have children but you can recognize at some point you have like these agencies triangles these faces and you only able to really really recognize whatever you've seen before your brain has been trainable understand certain patterns and only these patterns and the 1 that you can really understands it exactly what dual wear during is doing in the arms of of
recognition is a very interesting experiment what they is like you have use this picture from from from book that 1 in the middle in a small on the train now the built neural network with with this information where that that information was that wouldn't happen is that the neural network the it happened told to really get there are a representation of the world the neurons were trained in a way that are actually only could see you know about what picture shot after that after all the training after all they did these neurons that has been has been called the blood the rate at certain stages they show this picture of things naturally time so what this place I whenever you the the initial Mideast that I just showed you because in the model the the reality of the modeled internal representation of the model is used to hold on what what it returns actually the same exact picture but just with the with pattern from 1 book it's like about number patterns like there is no way you can think of like a red red ceiling generate group because they're the run has never been trained for that I think it's really interesting works
now the UN that believe like an intuition why or what can be achieved once the direction of all of these for all of the sudden variations on always like that all training by there were no 1 notable but more about like practical applications like for living donors are as data is now they still a the 1st 1 is classification I would say it's the most common ones from something and usually in a much supervised learning that you have some data and I work for these dating site little we have many problems with this respond people trying to abuse the system from people trying to cheat on on a where users and that and when he told Blair take them and lock them with 2 blocks has set like like some accounts so at some point as they descended you better data with many people revealed profiles on this said this is as this is not as prominent and another sperm and you have certain information you can imagine here that the axis is the age of the profile and the the exact is whatever feature you can imagine like there and we can be there the time the user has been resistance on model and understand there are some patents that that the kind of profile that you're looking for the kind of data that you're looking for 4 more in 1 of these sites for users with for broad like a line or a separator so you can identify at and this can be used for like the classical example from a farm whenever you will raised in books and you see some respondents has been classified they're using the same technique to finding something 1st that represent them they plot aligned more or whatever and at some point they they are able to distinguish which 1 they would from that what they did is on and also on cancer prevention for example they analyze images of 2 and they decide based on these models where diseases and cancer or not I'm encounter not for credit payments whether you'll someone is 1 with on your credit you always in the structure of the project that the most used 1 and easier to start they
would say what you need is someone for relabeled or or in some way that you would labels to the data you know all that certain and uses them pay that certain things like that and and would you look just like to replicate the same decisions that were written were we make you your your stuff in the in the same way a regression Eubalaena what exactly this are quite similar thing but let's say I want to know the price of these of these building and and based on the on the square
meters what so you can but all these part of these dogs and and then your
learning a regression we can also lead in the wording of quickly show up on
how video Ch
and it
it so the we know is
what I did is like the state of the art in what they did
so that different need it's usually named reinforcement
learning that it's way of that the what keeps learning based on experience for the robot to do is look at the beginning of the on the training stated it keeps falling and then even if you take it to the bottom 4 it wouldn't surprise me and I think there 1 of the concerns you have same that is where wearable order 1 a controller make their lives at some point I think it will make sense of the same for treaties this poor or what the in Denmark more application
recommendation systems whenever you won't 1 has on their using all these data scientist some techniques to see what 1 customer evolve actually it's the same that that everybody else was buying this product is buying you will find these patterns on these associations then
clustering is that is similar to what I was saying I mentioned before about classification but in this case there the whole point is that you will know the labels you know that you have fully friend bite of users of a book about dating sites in the tail by how the user are interested mostly in sex and the users that are interested mostly in sex but in a more subtle way OK and I would really want tool to identify them but I don't have label I nobody's 1 had just reveal them and say OK you have so long as you have these techniques to train models and these models and the unable to really is their data itself on finding the distance between different groups seem OK there is this group that is quite similar but at the same time and I will that it's quite similar among them but they are very different groups so it's about
finding these patterns network analysis is quite another interesting popping they are able to predict for example song from Berlin with signatures like like the flu waiting it was like the sort some of their financial the floor that they started to that data based on the network or flight connections when it was when I started the 1st cases in every single quantity and they're breaking that we've with great accuracy also for the banking system they check like the transactions of long bonds and they say OK these bonds defaulting these all along is is wonderful because they make so much transitions still there the
much recognition it's quite challenging topics in many cases in this case the idea is this thing which I think this is as while which kind of there's not things and get it learning this is my simple tool it identify which is I'm offering which is a lot in early better than a person because in some cases it becomes quite trick here which I also
celebrating cars is like that the state of the art quality of care about them there already in in the state in several places there already writing out enormously they have these cameras the sensors and the and they're really being able to even see fall when there is a the sample was tuple pros and all that
failed just briefly mention I was mentioned before there are many projects in game by from that you can and that you can use the more famous ones you really know already about minus by which is like the internal representation of of any of these soul were at the end and you have to be there for you already know what these 1 you like and how Web application is like a Python terminal web-based very useful when you are in the the science because usually plot things so long also you have syphilis for he sample listing models quite a one-to-one Jason Johnson is a political 1 and it's it's being used for topic modeling to when relief find the topics of conversation but I was mentioning before which are very popular people is talking about so on and that's it I'm just
our we mentioned were think optimizing something like the pull pull data on on Saturday relief would be partly this principle for wants to contribute to bind but if people is Neil bookbinders on your plate sigh and we can also what pictorial land and things like that so please get involved with me if you're interested uh there on onset of the Jews he just in case anyone wants to continue portal these projects or or just get the tutorial on these because will not yet that's it thank you very much
few only then we had a couple minutes of questions that
we have any choice and because no 1 really understood anything and couldn't bride wasn't as gentle as I read than that what's the positive which the
hi and immigration is really related to most to computers but on the other side of the human race so he said that 1 reason and then he made up of the the the this I mean and you serve as erasing morning you exactly then I mean any leverage so I question is why do we really much more tired when we learn new things at least instead of you know these are my guess he served on neurons that actually once the like more time why is the he and I didn't quite get the question weight
when the number of of neurons but with the tension so my question is he says that the respecting the that you get more tired this is actually where more neurons are activated OK but if they can I I would guess it seems kind where learning new things they much more time think of it that way so the keys in what I think the to the friend Mr. step in this process the say 1 is a boat training I would say that when you feel tired is because if you're seeing something you know your to your new they had job and that of course you feel much more tired because everything is new for your actually kind of training in this training stage so all the connections to be activated and need to be created so you have new connections that some neurons that are not connected among them and they are 1 on that getting their activated for the 1st time so I would say it's related to that and then when you're doing something repetitive is like just your your your neurons arduous are already being activated the ones you you're using so it just like information just float and not uh neuro science is I think it's kind of a question for more for and respect this but but yeah I think that would be brought in that region and over another person that to use
any other languages in Python like most learning like the language is designed for motion and right protocols something I thing all their I use modular
languages in done by finitely I thing everybody is like using scene there because known by his mother programming see there is also a site on that it's quite a standard for for all those things so I would say that they would just find that I use you would use something that is not by phone for for in terms of performance before holding actually I lost by from so I would never really like tool to have to use another language In some cases they for example tential all the and all these things some of them I think the UN it's finally Warwick writing Python but for flow it's not but usually you have findings and anything Python is very with with their interaction with the user with the problematic thing into the most bootiful called you can read in Python more than in C or anything else so yeah I think it's like that in the training is is moving there you have to stand for example for brothers the programming the father of our lives with more than 1 . 4 . 2 things in this world but make it a I'm not using anything else and I think you can really do a lot of data scientist with Python the great-grandson muscles and we have this sort of thing together but if
TUNIS <Programm>
Intel
Software
Bildschirmmaske
Facebook
Umkehrung <Mathematik>
Web Site
Flächeninhalt
Schwebung
Mereologie
Relativitätstheorie
Kartesische Koordinaten
Computeranimation
Programmiergerät
Pixel
Kategorie <Mathematik>
Zahlenbereich
Kontextbezogenes System
Biprodukt
Visuelles System
Fokalpunkt
Ereignishorizont
Computeranimation
Projektive Ebene
Information
Bildgebendes Verfahren
Touchscreen
Decodierung
Zahlenbereich
Zellularer Automat
Interaktives Fernsehen
Computer
Zählen
Symmetrische Matrix
Computeranimation
Eins
Visualisierung
Bildgebendes Verfahren
Touchscreen
Videospiel
Zentrische Streckung
Mikroprozessor
Shape <Informatik>
Pixel
Kreisfläche
Datennetz
Mustererkennung
Frequenz
Quadratzahl
Körper <Physik>
Mereologie
Information
Ordnung <Mathematik>
Computerunterstützte Übersetzung
Bit
Punkt
Mereologie
Drehung
Computerunterstütztes Verfahren
Raum-Zeit
Computeranimation
Eins
Vorzeichen <Mathematik>
Lineare Regression
Gerade
Phasenumwandlung
Korrelationsfunktion
Funktion <Mathematik>
Datennetz
Gruppe <Mathematik>
Prozess <Informatik>
Kategorie <Mathematik>
Optimierungsproblem
Digitale Spaltung
Bitrate
Ein-Ausgabe
Dreieck
Rechenschieber
Funktion <Mathematik>
Datenstruktur
Festspeicher
Physikalische Theorie
Mailbox
Information
Aggregatzustand
Subtraktion
Mathematische Logik
Gewicht <Mathematik>
Kanalkapazität
Derivation <Algebra>
ROM <Informatik>
Physikalische Theorie
Systemprogrammierung
Informationsmodellierung
Datensatz
Selbst organisierendes System
Rangstatistik
Datennetz
Vererbungshierarchie
Kalkül
Inhalt <Mathematik>
Ideal <Mathematik>
Schreib-Lese-Kopf
Einfach zusammenhängender Raum
Kreisfläche
Antwortfunktion
Einfache Genauigkeit
Mathematisierung
Mereologie
Wort <Informatik>
Sigmoide Funktion
Subtraktion
Gewicht <Mathematik>
Punkt
Minimierung
Zahlenbereich
Iteration
Computer
Computeranimation
Richtung
Überlagerung <Mathematik>
Virtuelle Maschine
Mittelwert
Lineare Regression
Datennetz
Schlussfolgern
Boltzmann-Gleichung
Bildgebendes Verfahren
Assoziativgesetz
Pixel
Datennetz
Globale Optimierung
Arithmetisches Mittel
Mustersprache
Datenstruktur
Rechter Winkel
Festspeicher
Basisvektor
Kantenfärbung
Information
Benutzerführung
Boltzmann-Gleichung
Autorisierung
Distributionstheorie
Parametersystem
Gewicht <Mathematik>
Punkt
Datennetz
Zahlenbereich
Computer
Mustererkennung
Computeranimation
Homepage
Arithmetisches Mittel
Informationsmodellierung
Generator <Informatik>
Datenstruktur
Mustersprache
Stichprobenumfang
Boltzmann-Gleichung
Bildgebendes Verfahren
Standardabweichung
Abstimmung <Frequenz>
Punkt
Computer
Information
Computeranimation
Internetworking
Virtuelle Maschine
Systemprogrammierung
Informationsmodellierung
Datennetz
Diskretes System
Gerade
Parallele Schnittstelle
Boltzmann-Gleichung
Einfach zusammenhängender Raum
Approximation
Datennetz
Prozess <Informatik>
Kategorie <Mathematik>
Optimierungsproblem
Chipkarte
Datenstruktur
Einheit <Mathematik>
Information
Boltzmann-Gleichung
Punkt
Datenstruktur
Datennetz
Datennetz
Zahlenbereich
Normalvektor
Bildgebendes Verfahren
Computeranimation
Subtraktion
Punkt
Wellenpaket
Rahmenproblem
Extrempunkt
Selbstrepräsentation
Gebäude <Mathematik>
Derivation <Algebra>
Sprachsynthese
Zentraleinheit
Computeranimation
Richtung
Informationsmodellierung
Prognoseverfahren
Lineare Regression
Schnitt <Graphentheorie>
Bildgebendes Verfahren
Lineares Funktional
Datennetz
Prognostik
Quick-Sort
Kugelkappe
Datenfeld
Funktion <Mathematik>
Grundsätze ordnungsmäßiger Datenverarbeitung
Gradientenverfahren
Sigmoide Funktion
Fehlermeldung
Shape <Informatik>
Punkt
Wellenpaket
Selbstrepräsentation
Gruppenkeim
Zahlenbereich
Transformation <Mathematik>
Mustererkennung
Bitrate
Dreieck
Computeranimation
Generator <Informatik>
Informationsmodellierung
Menge
Mustersprache
Information
Neuronales Netz
Trennungsaxiom
TVD-Verfahren
Web Site
Wellenpaket
Punkt
Profil <Aerodynamik>
Ähnlichkeitsgeometrie
Kartesische Koordinaten
p-Block
Physikalisches System
Computeranimation
Entscheidungstheorie
Richtung
Eins
Überwachtes Lernen
Informationsmodellierung
Quadratzahl
Lineare Regression
Endogene Variable
Projektive Ebene
Information
Datenstruktur
Bildgebendes Verfahren
Gerade
Systemprogrammierung
Lineare Regression
Lineare Regression
Mereologie
Besprechung/Interview
Meter
Wort <Informatik>
Computeranimation
Subtraktion
YouTube
Besprechung/Interview
Aggregatzustand
Assoziativgesetz
Systemprogrammierung
Punkt
Wellenpaket
Mustersprache
Gamecontroller
Kartesische Koordinaten
Physikalisches System
Biprodukt
Ordnung <Mathematik>
Computeranimation
Roboter
Einfach zusammenhängender Raum
Subtraktion
Web Site
Punkt
Datennetz
Benutzerfreundlichkeit
Gruppenoperation
Gruppenkeim
Einfache Genauigkeit
Physikalisches System
Elektronische Unterschrift
Analysis
Quick-Sort
Computeranimation
Informationsmodellierung
Transaktionsverwaltung
Datennetz
Mustersprache
Kontrollstruktur
Abstand
Steuerwerk
Spezialrechner
Stichprobenumfang
Ablöseblase
Mustererkennung
Computeranimation
Aggregatzustand
Schnelltaste
Umsetzung <Informatik>
Betragsfläche
Web-Applikation
Plot <Graphische Darstellung>
Mailing-Liste
Computeranimation
Eins
Informationsmodellierung
Spieltheorie
Julia-Menge
Stichprobenumfang
Radikal <Mathematik>
Projektive Ebene
Bijektion
Virtuelle Adresse
Gewicht <Mathematik>
Einfach zusammenhängender Raum
Prozess <Physik>
Wellenpaket
Protokoll <Datenverarbeitungssystem>
Rechter Winkel
Prozess <Informatik>
Formale Sprache
Besprechung/Interview
Zahlenbereich
Information
Schlüsselverwaltung
Eins
Demoszene <Programmierung>
Web Site
Wellenpaket
Formale Sprache
Familie <Mathematik>
Interaktives Fernsehen
Optimierung
Term
Datenfluss
Quick-Sort

Metadaten

Formale Metadaten

Titel A Gentle Introduction to Data Science
Serientitel EuroPython 2017
Autor Garcia, Marc
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 10.5446/33749
Herausgeber EuroPython
Erscheinungsjahr 2017
Sprache Englisch

Inhaltliche Metadaten

Fachgebiet Informatik
Abstract A Gentle Introduction to Data Science [EuroPython 2017 - Talk - 2017-07-12 - Anfiteatro 1] [Rimini, Italy] This introductory talk, will cover the basics of datascience. From the incluence of artificial intelligence, and the quest to replicate a human mind, to a practical demo on how to build a hello world machine learning in Python. The talk will try to answer questions such as: What do we understand by data science? What do we know about the human mind, that can be an inspiration for our programs? Which problems can we solve with data science? What tools are available to do data science in Python

Ähnliche Filme

Loading...