Merken

Deep Shit: Paradigms, Paranoia and Politics of Machine Intelligence

Zitierlink des Filmsegments
Embed Code

Automatisierte Medienanalyse

Beta
Erkannte Entitäten
Sprachtranskript
work at it if it if it if it if he will want to the at that
to the but in the way everyone
so it's good to see that quite a few actually weren't completely paranoid and
disenchanted after terrorists talking left and I wanna talk about something that has is very closely connected to what travel pattern was talking about when he spoke about images but go a little bit deeper into the more general ramifications and infrastructures and that underlie the phenomenon of artificial intelligence machine learning and also distributed computing blockchain and some other things to do so and I will start with the and a letter of from China that reached Norbert Wiener the inventor of Cybernetics in 1940 8 I think um the it came to him from 1 of his colleagues and friends called man and childbearing interesting type of who worked as the interpreter of Bertrand Russell and Albert Einstein during their travels in the 19 twenties in China and and nobody met him when he was also to being a guest professor at Ching Y. University in China in the 19 thirties and in this letter so the Enron child who was a linguist who also Funny Side Story of it the 1st transition of Alice in Wonderland by Lewis Carroll into Chinese which became such a vacate but it's on the Chinese sequel Alice's Adventures in China because what's even weirder than Wonderland and In this letter in he spoke about the game of go always being as you can see something that might be of interest for computational processes and questions of complexities
some you might remember that last year the game of go up appeared in the media when Google's software alpha goal of beats the Korean great master um
of co called the seed or this is how he looked after words and in Korea this really resulted in a cultural shock the Korean government the next they put a billion dollars into Artificial Intelligence research and there were newspaper reports about mass of the
events of people getting completely hammered out of frustration and so you could see at this set some enlightening impact and you remembered me chess is something that has been and done by machines for quite a while now but with just people said OK it's a it's a bruise number again basically and it's the once you have a computer that strong enough we don't really have a problem somewhere its goal was something that people always connected with human intelligence with the intuition is something that was genuinely human and people were really puzzled by watching the game between the signal doll and offer go because after good did something that
was on unforeseen that was unexpected from and therefore that people to believe that this software and had developed something like intuition or even intelligence this of course is not really true but I come to that later on alpha goal was also something that wasn't really new but it's interesting to note that on deep learning came to us in the form of of bread games if you will you might remember these sort of pictures that came out a couple of months earlier who will deep dream um which also work with and enhancement of patterns that it finds it sort of we thought that it something like Machine Inc elucidation which holds
true if you've seen hallucination is something that finds patterns somewhere and just enhances then enhances them enhances them from as Chavez pointed out these images are not images in the way that we are known to deal with images these images are basically the computer doesn't really care if it's a dog or whatever to image is a very interesting data set to work with but for
us in form of images and gains in these are images that don't show something but that obscure infrastructure behind it some which is called cancer floor which is the global um tool set for machine learning that at the same time alpha go and cool the dream happened was released as open source this and sounds like a very nice move by Google but will
always try to pretend that they're the good guys and but what it did in reality was significantly enhanced the user base and the experience that Google could gain from other people working with the tools
from this went so far article was 1 of the 1st companies that completely and restructured their infrastructure from software to hardware this is to keep the travel spoke of GP use graphical processing units this is the so-called tensor processing
unit the 1st single-purpose deep learning and processes that was built and so they completely restructured the whole infrastructure to feed
into the deep learning process so every global search you do every image you uploaded every communications um you this fire through Google mail from is being used to train the software and hardware to get better at certain tasks like natural language processing image recognition or whatever
and other companies followed suit Intel acquired a company aptly called nirvana and also last year all these companies came together to form a cartel copartnership on I what could possibly go wrong and and other infrastructural moved that happened in the last couple of years was this campaign called Avila some of you might remember there was a big Facebook idea um with them and the project Internet don't work with they were planning to build huge unmanned drones with a wingspan of 42 meters that would on the skies above developing countries beaming wireless internet access down to the people living there which of course also again seems like a very nice gesture but from what it does I think there very a conic imagery in this image this sort of some really reminded me of this so you can
see the number here is a billion people are getting access to the internet which also means a billion people feeding more data into Facebook
or whatever other platform on and then show you little video about Facebook's self-perception of um what the AI research is good for
and uh uh for goal is
to build intelligent machines to help people in
the Unites and what makes it intelligent started to and this is what we're trying
to institutions uh
it's funny to note that the voice really sounds like 1 of those early on machine voices that you can use when you factory of more documents it's not just an
exam we get inspiration from biology what we can do with my technology used to build networks of stability on the order of
magnitude of the brain of the analysis
in the so the way we try to
machines we showed a bunch of images of the use of dogs and for each have to tell it what it is you can go and measure the term
of dust small children example of Italy and what's very interesting is that
after what can I do this with your exams before is able to exhibit severe performance on the test and so I can we
help those was passed take half
a test that hatch on this work
would not talk not not not to this lecture seeks
from time closer hot
from what all
what talk slide that
lots of trade that's what now I'm in the village in my hand is like the new knows
that it might have been made
you go you did this through a battle same the houses where they I think it's
true that unlike the approximated
privately-owned even having like 3 words just helps flesh out problems so
that I can see it and continue the is
more than page and you have just
come online tool I think initially like 1 humanities he
did not know whether it is new and build an unknown men in heads
so there and usually
what do you do this division of unit interval
having this thing has a lot of this kind of
common human my resistance and of our people to do this and you will be
important to use and additional intelligent systems of the interview extension of words this word cars or extensional modes but other places so should you have the fire and
you want your memory you
knowledge and they're going to allow us to
concentrations of all the new and school to believe it it yeah OK there are a couple of things that I find interesting about this video 1st of
all the incredibly Biblical symbolism and language of making the blind see on which of course again is a nice thing to do and if you think it a little bit further and if you combine it with what Facebook is now working
on our along mask is working on with neural based on the creation of images on the neural cortex is something that is in the very graspable near future and 2nd of all they're all women and who have to be helped and 3rd of all most of them are black and so on that's something to think about I guess and the 4th thing that it I found interesting is the utter neglect for a any kind of critical potential and guarantees so he literally uses of the word extensions of our bodies like in the in the very middle make learning in sense and without even thinking about the political arm and critical implications that come with young the cool and has also been mentioned before and Chavez talk it's been around for ages so this is a paper he wrote in 1990 on optimal brain damage which is basically the approach now being used for computing in convolutional neural networks by pruning away redundant connections in the brain but not just on a side note when of course always
reminds me of chocolate and 2 on the mirror stage so the mirror neurons and their stage I another kind of infrastructure I
wanna talk about is that an actual older infrastructure for all the critical infrastructures that surround us and 1 of them being for example the sum Iranian nuclear infrastructure that a couple of years ago was as you probably know attacked by a very
sophisticated cyber weapon the most sophisticated by far and that was ever discovered in it 2 years it was called Stuxnet and it took years to find out that it was actually there
and then it took another couple of years to find out who did it so it was probably the Americans and the Israelis and this was a cyber weapon is was the result of a billion dollar project for many years that that was a huge project it was focused on the very tiny thing like the little switch in a centrifuge that could be sabotage this system cyberwarfare since then as you I have heard recently was a lot in the media and has completely
changed toward something that is probably still out there in the in the the in the
form of very sophisticated large-scale operations by governments but this was a forum posts on the botnet forum and last year but where a user name and sent by published the source code for a botnet called Uribe which turned out to be a and B thus far biggest botnet from the on the internet from which use a lot of the
unsecured IOT Internet of Things devices to do such things as kick the entire country of Liberia off the internet shot on the heating system in Finland or do other things so I think that's a significant change in the weighted infrastructures have been used in attacked from or infrastructures is starting to attack each other and just to give you a short insight
this is a threat not by research company is at the bottom of the text happening just right now so you
have come down here you can see this infections per 2nd and just going to leave it there in the background the little bit on so you can see this is probably half of the devices in this room are currently involved in acts of war sabotage or simply spamming something the the you get the idea the last thing I wanna talk about is and another thing that has been very prominently featured in the media in the last 1 and 2 years after the rise of the scoring people started focusing more and more on the underlying technology called blockchain on which is the idea that you can attach basically a critical ledger to any given entity of which will then make it authentic oval and usable for anything from smart
contracts to voting and other things from the daily quoted in a 19 year old guy from the Ukraine and forms a entity call colored a company called Ethereum from
which at the moment is blooming quite a lot and what I find interesting about this is that it was healed by on the left wing and the right wing but in the same way so this is something where technology again proved to
be absolutely non-neutral it's somehow swung so far out into the left spectrum that is circled back into
the right spectrum you had an oracle capitalists praising it you had from all tried people praising its you
had a any kinds of people present it and then there's also the chance that some of it is being used by the traditional financial industry or political establishment but after
Ethereum came this fierce looking entity called the Dow you can see these trigrams running around the you remember them from the thousand South Korean from the the barbed-wire from the gene and it also looks to the like from the dominant initiative from the TV series lost some of you might remember the Dow was to distributed autonomous organization that aims to
surpass any kind of other organizational infrastructure by basing anything and they could on the blockchain DOS completely taking out any form of human agency which also sounds might at 1st but when you think about it what gives some instances fields of sort of some notion of algorithmic agency that can be traced back to something like the idea of Adam Smith this invisible hand or even back to the idea of the hand of God so something that is sort of super
super idea logic or a super religious or superhuman in a
way and yet this is where
the dollar comes from the Dow also
meaning the way the trigrams themselves they help bring computing
into being actually I'm going back very far out into this early 18th century when guidance based on the knowledge from China on the tri grams of the gene that were sent to him by some adjusted missionary started development binary numbers thus laying the foundation for computing and
computability as is and the Dow in some again in
these slides from the NSA from the southern these are until it access operations and thank you very much it's time for questions it would be there was like here by the way backstage we have questions from the audience at so if you raise your hand seal we have 2 microphones this is 1 of the 1st rule maybe you can introduce yourself and ask the question and my I'm working in different instance so that over you know and
the area to say that there were a lot of false what's the new report from her and I'm not you might not name all of them but you just I don't give a grasp of what the learning is
about but they also made some really make connections like was the connection between tolerant call those 2 completely separate things and also you said set that in the in the Gulu video that the person was talking with a very artificial sounding voice but as really just the voice of an economist who speaks french so you just as the French X and I know this will you around I understand say it has to
like it strengthens the units here that as some could conspiracy theories this OK and I didn't want to in the book the notion that there is some like a conspiracy theory of 1st of all I was talking about not immediate connections between different but things like the Dow and tower was talking about and I come from cultural studies so we're up to make connections between different things that are sometimes even metaphoric and I was looking at different forms of phenomena and technological infrastructures that are happening in the past years I tried to make clear as possible and it has to decide in what what's
the connection between the ball and the of
it's a so-called signifier if you want it's the same words the same work 10
operations is the same thing as the the NGO but this is no connection from you I was talking about different things and was necessarily trying to make a streak connection between them so there's no connection it has nothing to do with each other
OK those of us abbreviations of on the
same from yes I know I was talking about different things I was giving different
examples but try to give you can see them as different chapters in the book the OK I'm sorry if
you dislike group I see a question in the
middle of the audience so all their thank you and and I was just following up on that machine learning deep learning in the eye and so
on and how much have you use all of tried to get into the theories
of of that was my perception is that there's a great misunderstanding on what is actually possible and the speaker before as I mentioned that and I've always thought of seeing that another book so how much you try to get a grasp on what what actually the
technology can deliver cream how much you've actually work with these things
and and and this sounds a bit that I'm having an then you have the mike so how much did you try to
understand what these things can actually do and what they not and as far as my capabilities allow
me to I have a masters in computer science which acquired almost 10 years ago uh and that never worked in the field so maybe we can talk later in the but I do my best to talk with researchers and to talk with people like travel bag when we did the publication together on the work that he is currently doing some glucose a team of programmers on individual aspects of it from and otherwise and and I'm not a deep learning research are not in the field of that and media theorist and I'm trying to shine a critical light on developments that are currently happening usually trying to basically to bring them to light and so as to raise awareness that there is something that we have to have a critical look on thank you do we have more
questions and yes 1 question was so thrown and and idea and maritime book change and the thing
I today is slowly being more about in the structures which are creating all 1 not and and I some of the most well that could be sorry could you speak up for the yes side put in this num wondering and point and learning from you
talk then if a structures of becoming more so that's not necessary but as well we we'll using the touch of what it's going to be and where it's a invention already heading up to time and so the feeling that we are thinking of like early adopters and the selective very very late adopter of those of understanding how ecology works in room and I would like to know how you us well
try to make the set raise awareness OK that that's how these well communicated with makers
governments in understanding what possible indication of the and that this is a
huge problem I think it in in general and and then generalizing broadly of course is that policy makers from our sometimes come from an older generation that tends to be a affirmative in a certain kind of way and people were early adopters of technologies are come from a very young generation were very affirmative of certain developments on without necessarily having the experience of thinking about long term effects of it and there's not much in between so what has to be done and what still isn't being done on a massive scale is public education and is starting kindergarten and goes with this high school system and i and some sort of some understanding that that people have a responsibility as citizens in that too because we still think the all these kinds of things that I talked about even though there were specific some examples like inferior model tower whatever you know something they're going to be around and that these are things that not happen outside the realm of existence but touched on so I mean what I do is I teach I teach at universities and I give lectures and I tried to talk to people and also in in politics but I'm not a politician I think we have time for 1 last question 1 last
question OK that's not the case that thank you the time
to hop from the 1st and of the same
Virtuelle Maschine
ICC-Gruppe
Computeranimation
Wiener, Norbert
Interpretierer
Bit
Gruppenoperation
Kybernetik
Fortsetzung <Mathematik>
Computerunterstütztes Verfahren
Komplex <Algebra>
Abenteuerspiel
Neue Medien
Schwebung
Software
Prozess <Informatik>
Mustersprache
Datentyp
Verzweigungspunkt
Jensen-Maß
Russell, Bertrand
Algorithmische Lerntheorie
Neuronales Netz
Metropolitan area network
Virtuelle Maschine
Computerschach
Spieltheorie
Geometrische Frustration
Besprechung/Interview
Zahlenbereich
Ruhmasse
Wort <Informatik>
Computer
Ereignishorizont
Verkehrsinformation
Sichtbarkeitsverfahren
Bildschirmmaske
Spieltheorie
Software
Mustersprache
Jensen-Maß
Computer
Bildgebendes Verfahren
Quick-Sort
Virtuelle Maschine
Bildschirmmaske
Open Source
Jensen-Maß
Bildgebendes Verfahren
Computeranimation
Einheit <Mathematik>
Prozess <Physik>
Hardware
Tensor
Software
Besprechung/Interview
Vorlesung/Konferenz
Graphikprozessor
Computeranimation
Telekommunikation
Suite <Programmpaket>
Facebook
Prozess <Physik>
Hardware
Besprechung/Interview
Mustererkennung
Natürliche Sprache
Quick-Sort
Internetworking
Task
Software
Gmail
Meter
ICC-Gruppe
Projektive Ebene
Drahtloses lokales Netz
Bildgebendes Verfahren
Facebook
Zahlenbereich
Große Vereinheitlichung
Videokonferenz
Internetworking
Virtuelle Maschine
Einheit <Mathematik>
Besprechung/Interview
Virtuelle Maschine
Stabilitätstheorie <Logik>
Datennetz
Gebäude <Mathematik>
Besprechung/Interview
Vorlesung/Konferenz
Faktor <Algebra>
Größenordnung
Analysis
Sichtbarkeitsverfahren
Virtuelle Maschine
Verzeichnisdienst
Besprechung/Interview
Facebook
Term
Bildgebendes Verfahren
Softwaretest
Besprechung/Interview
Hilfesystem
Rechenschieber
Besprechung/Interview
Besprechung/Interview
Wort <Informatik>
Homepage
ATM
Verzeichnisdienst
Einheit <Mathematik>
Besprechung/Interview
Vorlesung/Konferenz
Facebook
Wort <Informatik>
Physikalisches System
Maßerweiterung
Division
Schreib-Lese-Kopf
Konzentrizität
Festspeicher
Besprechung/Interview
Videokonferenz
Einfach zusammenhängender Raum
Facebook
Vektorpotenzial
Minimierung
Formale Sprache
Kolmogorov-Komplexität
Symboltabelle
Computerunterstütztes Verfahren
Computeranimation
Verdeckungsrechnung
Datennetz
Faltungsoperator
Wort <Informatik>
Maßerweiterung
Bildgebendes Verfahren
Neuronales Netz
Gewichtete Summe
Besprechung/Interview
Nuklearer Raum
Computeranimation
Resultante
Neue Medien
Cybersex
Vorlesung/Konferenz
Projektive Ebene
Physikalisches System
Nichtlinearer Operator
Bildschirmmaske
Webforum
Quellcode
Computeranimation
Internetworking
Minimum
Mathematisierung
Besprechung/Interview
Physikalisches System
Internet der Dinge
Ganze Funktion
Internetworking
Neue Medien
Bildschirmmaske
Abstimmung <Frequenz>
Bit
Oval
Benutzerfreundlichkeit
Besprechung/Interview
Systemaufruf
Authentifikation
Computeranimation
Design by Contract
Chipkarte
Momentenproblem
Vorlesung/Konferenz
Punktspektrum
Computeranimation
Selbst organisierendes System
Reihe
Punktspektrum
Computeranimation
Orakel <Informatik>
Bildschirmmaske
Datenfeld
Algorithmus
Grundsätze ordnungsmäßiger Datenverarbeitung
Vererbungshierarchie
Mathematische Logik
Quick-Sort
Computeranimation
Instantiierung
Zahlenbereich
Computerunterstütztes Verfahren
Softwareentwickler
Binärcode
Computeranimation
Rechenschieber
Nichtlinearer Operator
Schlussregel
Computerunterstütztes Verfahren
Computeranimation
Einfach zusammenhängender Raum
Flächeninhalt
Besprechung/Interview
Systemaufruf
Vorlesung/Konferenz
Verkehrsinformation
Videokonferenz
Einfach zusammenhängender Raum
Beobachtungsstudie
Bildschirmmaske
Subtraktion
Einheit <Mathematik>
Besprechung/Interview
Turm <Mathematik>
Physikalische Theorie
Computeranimation
Einfach zusammenhängender Raum
Nichtlinearer Operator
Vorzeichen <Mathematik>
Wort <Informatik>
Computeranimation
Besprechung/Interview
Gruppenkeim
Computeranimation
Virtuelle Maschine
Besprechung/Interview
Vorlesung/Konferenz
Physikalische Theorie
Besprechung/Interview
Neue Medien
Programmiergerät
Bit
Datenfeld
Besprechung/Interview
Softwareentwickler
Kontextbezogenes System
Informatik
Physikalische Theorie
Computeranimation
Punkt
Mathematisierung
Besprechung/Interview
Zahlenbereich
Datenstruktur
Maschinenschreiben
Besprechung/Interview
Kontextbezogenes System
Datenstruktur
Umwandlungsenthalpie
Soundverarbeitung
Zentrische Streckung
Informationsmodellierung
Generator <Informatik>
Existenzsatz
Endogene Variable
Turm <Mathematik>
Physikalisches System
Indexberechnung
Grundraum
Term
Quick-Sort
Besprechung/Interview

Metadaten

Formale Metadaten

Titel Deep Shit: Paradigms, Paranoia and Politics of Machine Intelligence
Serientitel re:publica 2017
Teil 105
Anzahl der Teile 235
Autor Feigelfeld, Paul
Lizenz CC-Namensnennung - Weitergabe unter gleichen Bedingungen 3.0 Deutschland:
Sie dürfen das Werk bzw. den Inhalt zu jedem legalen 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 10.5446/32966
Herausgeber re:publica
Erscheinungsjahr 2017
Sprache Englisch

Inhaltliche Metadaten

Fachgebiet Informatik
Abstract The lecture explores the infrastructuralisation of artificial intelligence techniques and technologies including deep learning, convolutional neural networks, robotics and IoT along with the autonomisation of capitalist processes in tools and entities like blockchain, DAO and Ethereum, approaching them in the context of their cultural, philosophical, political, social, economic, and ecologic entanglements.

Ähnliche Filme

Loading...