Embed code
Note: this content has been automatically generated.
and we now have the tools i mean to filename to answer that uh that need and
that's what we should embrace uh that you can do that now and we should do
it seems that there's quite an optimistic uh outlook on all the so if whether there's optimism
or not there's a there's a it seems to be an acknowledgement that it's essential
to do this it's the only way to go to in order to cope with demand and cope with the way that the the market is going so um i
want to thank our our panelists here um if there are any more questions we
do have one more that we we do have time to fit in
i think um my question addresses to both nicholas federations
uh the is a i also to got uh some data for your sponsors
good question have uh not yet on you if you wanna maybe you're more than something that's like i mean yes and no i mean we
we are gathering information i mean but the gathering of information doesn't imply
that use a and it's it's the tool again but uh one
it's an additional two let's say i were i don't know if it's called a or
not it's again a tool that is able to underlies moving images or or contents
uh just like some of the tools measure you know exporter on t. v. so
we're starting to use this to measure it ought to try to to to
get a new the value of all the digital activities that were going
'cause one of the downside if i can say of creating all this content is that content is
scaling like like crazy and it's true that we have a little bit of trouble of controlling where it's going and
understanding exactly were things or or ending up and how they're being consumes alright these liberals or well known
thank you very much to this panel uh and i hope you'll join me
for the limited we've rather want people to concentrate on what can you do once you have a calibrated a camera
then you can add graphics uh i mean for thing with a nice little sideline for soccer
'cause you're just sitting next to me that there are other ideas uh that we have
and other actions i mean activity that we're working on if
the um what's old with the automatic cameras selection
what that is is we have an automatic system that can say liked automatically which is
the best camera up to look at so you could say
okay wait then we can automated production well yes
maybe for other work here or production but not for hire your productions there we want to use this
two supports either active to support operators to make suggestions would operate
without a look that camera angle there's something to see
went to consider that or using that information to prepare replace or to prepare highlights
just to be able and that's the value of artificial intelligence
and machine learning is to be able to do more
artificial intelligence will all slowed things and is capable of dealing with more contents more personalised content
rather than the say the operator or the director that's generating one mace when feeds
to be on their using a on you could create many more personalised fee
uh in soccer you would say okay i want to feed focusing all my method for instance
or i want to feed actually a bit emphasising my thing because i like my team
and that's the value that it brings and by using all these yeah elements um that's what we want
to give obviously your customers to make them produce more and better to better serve their audience
and there was the idea parallel uh a parallel production as you mentioned before like a lot
uh when we were talking that sometimes on the same uh event on the playing field
different camera angles might be suitable to different platforms absolutely yep that we can
do that in the say in two ways that either we say okay
um the cameras on the page and the official when got director for gays
you feed to be able to put on there we can t. and it's a bit and say okay i want to see a feat
this is in favour of my thing 'cause i like my team or i want to see a feat
that uh favourites um in attacks or that's that favour some other things and then
you can really or the favour specific player and then you get really generate
it's much more uh of these fees and ultimately that's uh what
today's view worse want really want more personalise the content
and that's what a i can't win without exploding the number of uh people
and not because we don't want to have this people but there's
no place to put them and they simply do not all uh exist
coming back to what you said is it is also not just
personalising feats but also targeting feats to specific a device
and therefore we're working on the tool that what we call automatic framing basically what
it does is it cuts out a part of the full h. d.
view image of relates the video of um the reason for that is if you want to publish let's say on the mobile phone
for soccer of full wide view of a soccer page i mean
your players will be ones millimetre higher so on your
i mean mobile phones plane that's not very pleasant to watch so we cut down
it's really the part of the video where the heat of the action is
and that's some interesting information but also can be used for stuff
the one oh sorry nickel out there that sorry uh for example a lot of people
watching sports on on their mobile phones uh if the data that you get in
uh a square picture maybe more interesting than a rectangular pitcher for example in this kind of solution of low to
create a split peter so people can watch on them away from that you could for example so that's
something which is better and it doesn't change the the the the the whole view
of the content but with leaves actually evolving that it's improve what a
i brings there is that it interprets the image and so it can really
show this is where human people this what human person we've watched
and that's different that's it to compare to more traditional approaches that would get parts uh left
in light of uh of the image really show there is where the action a seven
there's an where you commit all this is a little bit different uh we're talking here back camera automated
let's say a a assisting production uh using the
tools um your a company is about using
a machine learning in a slightly different way could you tell us how that works out
yeah him because machine learning a i slightly in a different way
and b. b. n. a. i platform that watches the gains
and you think you know beyond a broadcast the eight or any i their visual information and it can but
somehow talking gains eight process that begins in real time
and it generates tracking data and and action
well it turns out that everyone a is and what everyone is doing at every
say can again so we end up with a massive pile of data
many many pains beach are extremely difficult for human being to interpret
then be use machine learning a high and bitch because game models and they can tell us
and for example how much an action of the play yeah contributes to the outcome again i think didn't talk
this way we can add quantitative late compared tactics the strategy is
effecting play yeah a place up my aunt not and
then we can find interesting that's the way we can find and things that are
important for fact we can generate personalise this story is looking at the data
back thing with the data but not necessarily talking about data
we get the numbers we find interesting it's though is the
pair them with videos and they show those videos
be create content we have comments and a fan but
at the game differently so basically the big
we try to change the bay that at people are watching the game like it's going to be a new experience air
getting more information cater what's your create friends about the things that you hear and back
to it that you know the us to question backstage that i just uh
thinking about which is how good or you're predicting the outcome of matches use your technology you're saying
you're quite good so actually you could get to a point relate were you could say
yep they're gonna win this one here think it hypothetically i can predict the outcome again so i'll
no one gets to play even know who is going to maine but honestly us
forces not about predicting the gay if there's always onset think the and
i can major things that our national but there are many other factors and
again that can't be midget can't be quantified and we watched again
we didn't see it good looking for a surprise might be
we know that this tactic is not going to work against sets in fact they
played by another team and be half is sort of quite they tell but
was it works and it is chords and that's the moment that you want
to make and story and that's the moment that you're predictive model face
yes we have air their predictive capabilities we can't predict the outcome of the games but
we are not using things to predict the outcome getting big take those
predictions let's say we can say but with the expected escort
up again or what would be the expected a score of one
is this big play yeah in that page and maybe
we can see the results and we can find discrepancy and we can generate s. stories based on that thought that
it's going to be the neil kind of experience of watching again
it's all about emotion ultimately and and and the human element
um and what what we can what i just like to ask everyone here is
to be clear for the audience also that that the human eye the human element remains in all
this at least for now like this what we were debating a little bit behind the scenes
because they all you were saying it's a tool it's important recognise that it is that you all the question for
you or someone saying the idea that this is gonna put people out of work is just not the case
oh it's i don't it's not the case i mean in our in our situation or when we
brought in that the the this machine to create content uh smart content and automated content
our it'll team started to get a grey hair and you know heart attacks and and and shaking but there's only
relies uh for a couple of weeks that the machine was actually doing all the liberties work for them
the boring work if i can say do you know the content that goes out and that
is just consume very quickly and then from the wave where's the could actually shift
their focus on the more premium qualitative type of content and seventy
that's totally change the perspective of the of the job
and they're actually you know with the machine no because it frees up so much time for
them to give them even more budget to work on on bigger and nicer products or
i i i know at the moment i i don't i'm not afraid
of abuse i see this as a huge facilitator as a curator
uh and we just have to uh that's maybe the way we
wanna producing the way we wanna distribute brings to what
the machine is able to give us a bit like when computers came out everyone was scared like we're gonna be
our work plan and we see today that it's giving us more work
uh norwegians and uh and and we are more productive so
and i don't see a without computers these days yeah that computers and cell phones directly
likely you're quite optimistic about this because you were saying the human element
is so fundamental i mean everything is and sports about fashion
it's live it's it's happening and that's what we wanted to believe and
want to communicate these these fashion to the to the viewers
the machine is just here to help us in it but uh i i i've
i don't see any kind of coffee between the machine and human uh_huh
the two on the doing the same thing and the machines are here to do a lot of
automatic things all boring things because they are so fast a regular yeah they are so but consistent
but the passion is not coming from consistent and you use the use the point that would sit down
head butting somewhat of a football match the machine would've would've ignored that because it wasn't relevant
to the attorney for the football that that it's not it's not the official event
yeah so yeah that's the one clip all of all the continent can be created so
so and and the the mission can bring a lot of things but different in that the special
sessions not going from formation get fashion coming from human beings doing an activity and then
do some things that i think in the way of actually no choice also to embrace this we are in a we're in the race
uh where everyone has a right older to stay outside you know on the surveys to be seen and so on so
and people asking more content everyone is asking more content even other machines are asking more count and so if
if you wanna stay there visible uh so that you can speak to sponsors so that
you can speak to your audience you need to be able to kind of
much what you can get off of a artificial intelligence and and what you
can get all the little things people ah watching sports on different devices
and that's one of the girls to adapt the content specific device
and the aspect ratio for example is something really important
i think this is the direction of frame that shows the you will have that in the best
possible content on the device of their choice and and this is where we should go
and machines can bring a lot of added value to that by the way if
there are questions you're just raise your hand i'll throw you the uh
or someone will bring you the the uh the box here so that you
can ask it um but please don't hesitate just jump in um
the question of style comes up and i don't know how important it
is really a colour uh is is where you mentioned before i
is is are we going toward say uh i uniform style of
production and of of content so that it's all gonna become the same cookie cutter that we're gonna see for everything
or is this only going to be at certain levels of of of for
that i think we're going total opposite actually on the machine is smart
cool artificial intelligence and is able to actually customise the content the way the receiver wants
to wants to see it so in our case for instance we get publishers
uh in in various different countries and they tell us okay i don't
care about okay i'm friends i don't care about uh spain
i just wanna get the clips of friends sent to my mailbox and a package
just this way with this intro video with the role within i'll probably do
no mention is able to do that in a complete different way then we'll be doing it for the for the spanish probably sure soul
i'm i'm really uh i'm not so sure that we're gonna go into something like a flat
boring uh 'cause the machines i was actually going to learn from from what people
and we'll have that so that i i fully agree on everything we had been because even that even when using machine
uh learning or that's artificial intelligence you can teach learn design
the system so that it implements this different uh flavour
i mean the example that i gave okay the food they just
selects the best the camera and produce has a program feet
you can tune that's this them to make it even if it the style of the french producer or
german producer that have different ways of uh of working so even in the a. on itself
you can difference hate and ultimately it's because the people
want to see that a different station but
the only is more really in in support it so you still have the director
really focusing on the creative apart but the director oscar the ice is than
a beer yeah system um give me a shot of that player
in current systems uh it will take time and you need these people are not waiters to really
assuming on the specific player and i can do that and they do a great job
but unfortunately they will only zoom in on one or maybe two by using this system
you can have the other system zooming in all all players on the field and having the director really big
the one that uh he or she wants depending on what's what's the personal style
depending on what the style of the country of the sport um depending on
the motion that the rector once but in the gain 'cause that's the little where
i see uh at least that this this moment the big difference haitian between
the humans and analysis analysis them it's it's in supporting that it will you will
off load things to the only system even things that look a complex light
following a player or selecting the best the camera angle but ultimately it's
the fall protection through the plane and the most um a little story to bring it home telling
viewing for the audience that's humans to the audience and that
is not in the standalone mode that actually is
bi directional that takes into account the motions that live in the stadium during the match
and that use that into account to really can create compelling a story and and the i. system
maybe we'll do it once but to they you really need to human humans to do that
i just add a comment about this that i and their bit the current and machine
learning techniques for example you can you would send a supervisor reinforcement learning and
given machine that trans to do some exploration you generate you know
what they use r. f. fan once the see like
it but i find cancer bought a set seventeens our second
play yes but you get a chance to machine too
and make some exploration and create some other contents for that you there
and then it eh images that interaction and see and it may
just be the user enjoys the new content or not and if it
is the right direction to go it probably does that more
to art and that kind of a theme that then issues and so
we can get mission the chance to do some exploration but
yeah all the mission learning models they on not
quite good beget there in future but at the moment uh you have some tools that we can leverage
but there there's been helpful questions that call them the passion i'm interested in a specific scene
and it's good like this so they they can give us i mean some some indications on okay view these other takes the okay
so i mean they can bring us a lot of things but that's where the human beings are so good that is
yeah it's something into generic this passion so for me it's it's really a complete i mean
something whether it be there's no there's no and no uh with
a i completely agree with your weak and three morphine
amended yeah i'm not going to remove humour from below piedmont to make things better for humans and they want to
they know is that make things is not meant not lives easier
as a producer as content producer i want to have hundred different that's the reason i chose the best
mark rather than looking at everything myself i want to get like that's a about this than that
asked them every day how slight interesting is the what is
that you've fallen uh in the past few objects
rather than myself looking at those things so it's basically the the like the
final decisions because uh it's a learning techniques are not there yet
it is the current techniques that are out there that this that
we can do is to follow it by hillman and
design something that behave exactly like backed producer but this is
something i don't want to be the ones that way
above a question production i think that uh a machine learning or activity intelligence can bring
any right older closer to lose something rather right now like
the currency that will everyone's talking about engaging religion um
what you were mentioning about you know being able to create stories based on the data that the remote to collect
now we have actually even robots we we tried that actually doing one or
more events would all the phase book robot that was able to
communicate with every fans already at a suburb run level every one was every fine was able to request
a particular type of content so it's pretty nice for every player uh
from that for every player i want bands or one yes it's
and and the next level is really having a conversation so watching the
game and say hey hey buddy whatever you wanna call it
uh what do you think your who's gonna be next core or how high is this guy jumping because
finds also getting should rejected you know they are used to to get this from other people so
i see a if you're a see huge player for engagement
if i can ask you that you censor face book is the faces of box part but
basically it was a perfect face book enable you to do that correct yeah so you build an interface
in the background and you use these book messenger i guess it's possible with any other messengers
and then you interact with the rubber soul ours what again very surface
level but i'm sure in the very near future we will have
these guys you're the new building machine are actually able to have a
conversation with you we saw it from good okay so it so
again questions if there are any please just stick your hand up and we'll here's one right here in the front
rather than throw this all the way across the room
thank you a question for leisure and you talked about a machine learning for contents creation
now if you can be right so many things on our past games what about using it
for training or coaching performance are you doing better thank god we have two different product
one is take back to art and a digital media and broadcast there's the
other one is for teens and then i did look at player's performance
they look at the historical data and and we provide
access actionable insights for co chairs for acting analysis
big weekend then then what might it around the are nothing to replace a coach
we want to build an assistant coach that provides information and digest information and
bring something to the core which that coach can make a better decision informed decision
also all they can and they have escorting product we can look
at a play yes in mind our legs and all
they are playing and we can't really comment players based on the playing s. i. based on
at that time and then we can say that these two might be a good faith
for your team if he is uh if these other players are eighteen days a these
new players have kind of similar ascii is and they are playing the same aside
uh that's a good team products that we have and uh
it's been useful um i'll be back more than a half of that trade
in their past three seasons in n. h. l. north american hockey league
so it shows that half of the trace have been back it they target real information not basically got thing
the the one of the things that was commented on uh you know this is something you said that the trend that
uh it's not the digital has certainly created a a completely new i
mean it has created new ways of consuming but that the personalisation
friend the the desire the market was all ready there or what was going to grow anyway
will i strongly believe that indeed this personalisation um something
that people want and they just isolation isn't
that isn't able actually just in the same way that i think that artificial intelligence as such
it's not the goal for a like paul tells if the means any turns out to be
i mean the most interesting means that we have at our disposal at this a
moment to get there but the intrinsic needs of other say of humans to
to get information to be entertained in to be that
in a personal way i mean that's that's there
and artificial intelligence uh i'm delighted to say that we're gonna handle better manoeuvre
superior intelligence add j. stuart good friend a colleague of mine to lead us through this session chase to it
and welcome all you're you're probably wondering who is this strange man with this odd wearable
around his back because i notice some one of the only people here with this wearable which is extremely interesting one dry clean only
and uh the thing that's interesting about it is that the italians do not
have a monopoly on this technology this was actually made in ireland
so there we go i'd like to move into our conversation here about new
technology it's extremely interesting one uh one machine learning and hey i
uh if you couldn't backstage is now as we were preparing a you would have heard
and i hope you will hear in our conversation that we're gonna have now ah
some strong opinions on the subject we have a very good discussion very good
debate going on where all this is going with machine learning and
a i and some of it follows on from andy washes comments at the beginning of the day which i hope you heard about
where things are going on how i machines are going to help us in in doing what we
do and what the limits are if they're already what's so i would like to introduce now
the people who are gonna help us in this discussion and uh we got a really
great wind up with people from sport can from technology uh first of all uh
also joining us your and one who is he a senior they hate innovation technology from the yes
boy away from the broadcast also distribution a java okay who it is founder of from canada
welcome all
channel as i was just saying that we were we were having quite a good discussion uh in the uh
it behind a curtain here about uh if a guy and a
machine learning first why should probably preface this by saying that
most of this discussion as i think many of you understand
is about the production of content and about the delivery
of content uh to the market so uh
oh the two gentleman from federations here uh and uh or other
colleagues or uh looking at how to produce content how to
do it effectively and and using the tools uh that are out there i should say that they they come from different
perspectives uh on the the subject and big well i i think i'll start with you
uh which wind a question about that because i think about
has already dropped with both fee into this space of of a of a a i and
machine learning in terms of what it is that you do so if you could
maybe take us through we're it's sets within what you do and why it's important
so jumping with two features are quite a strong word i think we're still scratching the surface but
four years ago or management folders logos we need to accelerate the
growth of feeble we need to make sure the basketball continued
uh being a being out there and and we need to be more
aggressive on the market uh and and be more more visible so
how it quickly related to to me to to my my uh my department is the fact that
as have all digital wars digital upon when we need to be
more reasonable uh create more content and distribute more content
and this with a very common constrain let's budget or sooner kind of budget
and that's how after couple discussions with some company we come up uh
uh we met some company could actually provide a scenario solution
that would allow us to to do that to do that exactly so create really a lot more content in the shoulder
uh i'm off time and to distribute it to a a much much bigger audience that we never did
so talking about funds obviously but not forgetting all federations those are guys or consuming content a lot
uh stakeholders like players as well and media uh and broadcasters so yeah it's really are
enabling this uh this type of content production and and enabling us to scale
i just think at what we're saying before hand is that it's responding to
a a trend uh it's not technology in search of
a of an application it's it's responding to something
that's that the audience wants and that the market what's yeah i mean we
discussed this earlier in digital is creating new ways of consuming content
a final becoming more and more more fines and and others actually are becoming more and
more particular in how they wanna receive the content only wanna consume the content
and i'll bring a traditional way of producing on then we'll just know enable us to to answer those
uh see someone would not necessary lawyers are set or label because you have to remember that fee but i still like uh
a midsize international federation we don't have the means of others like maybe you are five
so we need to for me to find ways to to accommodate these new trends
i'm keeping in mind also that and be a huge player
in where the digital is driving or extending fans ought
to consume content is giving them the condom they want so uh such feeble being also basketball we would be
in the you know we won't be uh looking good uh if we were not able to actually repeated something similar
but you're saying in terms of the volume of of what we're talking about you were saying tens of thousands of
oh yeah of clips annually literally yeah uh with the same who
more or less the same production team uh about five
six seven people depending on on on events we've managed to
multiply um no i can even know exactly but
thirty forty e. x. the number of you do that we're creating a on a yearly basis
nick let dale you you at the way for continuing to use traditional means let's call it human
means of doing this are you planning on making a tradition a a transition again
i mean not as well it would be the the different solutions uh
it's important and that's what it is just uh uh i think that the the human beings as you said
i mean the production done by human beings for one that's something we really need to uh to remember to make sure that
the the contents with the awful human beings as the sound okay or for everyone yeah
that we need a louder for particular place yeah so that that's yeah something which is important if
we want to go for the best possible then the worse and uh for the moment
he's the base the really is
sure for that
but yeah it's of course see uh you know we can use different ways and the f. is it but
based on the well typical uh uh we i to sort of
is this still i i think it's still quite here i don't know if we can um we can turn it up
um one of the things that you would mention before at the very start you said that this is
a kind of you can get into this subject publicly because in the beginning was the network
yes explain what that what that means in georgia just gonna picture of your headset other the thing is that
being that was the the network the the development of of the the the telecommunications a network
capacities is really the enabler of uh yeah it without the network nothing works including
really need to make sure that uh we're doing the right thing yeah
the problem it working place and then we can use it
uh was also mentioned you know your that one thing that has really changed
a lot of thing is the fact that that everything is bi directional
so you can send the content uh was with the viewers can also
send information and is a complete change compared to the previous situation
any as a it gives the the proceed from a i used in the beauty that is as as a
fantastic to to create this to create additional compares to multiply the the the the content so it
it's a new tool that is enabled by the the network that that's what i
am doing basically but yes it's it's it opens a lot of gets
yeah and if you could be before we jump into it if you could tell us a bit about what e. d. s. is doing and
where you are in the spaces of a a i already so because
we've had to federations your perk a provider of solutions like
well maybe before really don't need to some examples um it's very interesting to hear i mean to i
mean nothing opposing but complimentary point of view because that's also what uh with with what we think
if they it there with the need to create more content uh more cost effective uh content
um and for the same or lower a clearer productions the question is how can we also
made as much as possible to produce we've limited for in higher rents productions is
not getting with the crew because they need to focus on storytelling on creative things
and what we want to do is retrieve aunt production techniques that can help
and for the automation and make suggestions help the creative of storytelling part
and we do that by automating some part and by off loading let's say more cumbersome activities
so one of the things that we're other rather recently
product tallest is based on automatic camera calibration today
the operators need to calibrate uh the camera and the creative value in that

Share this talk: 

Conference program

Welcome Words by David Eades
David Eades
15 May 2018 · 9:06 a.m.
Welcome Words by Philippe Leuba
Phillipe Leuba, Chef du Département de l'Economie et du Sport
15 May 2018 · 9:08 a.m.
Welcome Words by Grégoire Junod
Grégoire Junod
15 May 2018 · 9:13 a.m.
Welcome Words by Benoit Mariani
Benoit Mariani
15 May 2018 · 9:15 a.m.
Hacking the Future
Andy Walshe
15 May 2018 · 9:25 a.m.
Talkback Session: Smart Venues, Moderated by David Eades
Daniel Marion, John Rhodes, Claire Lewis
15 May 2018 · 10:18 a.m.
Talkback Session: Wearables and Data, Moderated by David Eades
Terho Lahtinen, Jean-Christophe Longchampt, Christophe Ramstein, Patrick Schoettker
15 May 2018 · 11:04 a.m.
Talkback Session: Fans and Data, Moderated by Rebecca Hopkins
Horesh Ben Shitrit, Pete Burns, David Lampitt, Clemens Schnellert
15 May 2018 · 11:46 a.m.
Interviews, Moderated by David Eades
Stéphane Guerry, Joey Tan, Olivier Glauser, Albert Mundet
15 May 2018 · 1:45 p.m.
Talkback Session: Machine Learning and AI, Moderated by Jay Stuart
Nicolas Chapart, Nicolas Déal, Johan Vounckx, Mehrsan Javan
15 May 2018 · 2:37 p.m.
Jochen Färber
15 May 2018 · 3:14 p.m.
Talkback Session: NextGen Video & Engagement, Moderated by Nicolas Henchoz
Christoph Heimes, Spencer Nolan, Pedro Presa, Alexandra Willis
15 May 2018 · 3:30 p.m.
Interview on Security
Earl Crane
15 May 2018 · 4:08 p.m.
Talkback Session: Cyber security and Digital Security, Moderated by Sébastien Kulling
Dang Duy, Thomas Shorrock, Jean-Pierre Hubaux, Simon Trudelle
15 May 2018 · 4:19 p.m.
Introduction of Day 2
David Eades
16 May 2018 · 9:35 a.m.
Social Business and Sport
Muhammad Yunus
16 May 2018 · 9:38 a.m.
Talkback Session: Designer Bodies - Yes or No? Moderated by David Eades
Roland Sigrist, Vincent Gremeaux, Carlos Canto Alvarez, Véronique Lugrin
16 May 2018 · 10:27 a.m.
Talkback Session: Designing for the Future, Moderated by Jay Stuart
Ali Russell, Emilio Risques, Véronique Michaud, Thilo Alex Brunner
16 May 2018 · 11:08 a.m.
Talkback Session: Protection through Innovation, Moderated by Mike Miller
Liam Mc Tiernan, Laurent Mekies, Eric Nauman, Martial Saugy, Mathieu Saubade
16 May 2018 · 12:04 p.m.
From Racing to the Road
Laurent Mekies
16 May 2018 · 1:32 p.m.
Talkback Session: Understanding the eSports Ecosystem, Moderated by Jay Stuart
Brent Barry, Anna Baumann, Julien Delalande, Michael Journot, Carlos "ocelote" Rodriguez Santiago
16 May 2018 · 2:05 p.m.
Talkback Session: What's next in eSports? Moderated by Lars Stegelmann
Brett Abarbanel, Stefan Kuerten, Jan Pommer, Federico Winer
16 May 2018 · 3:05 p.m.
Closing Words
David Eades
16 May 2018 · 4:06 p.m.

Recommended talks

Closing Session 1
Clément Epié
11 Feb. 2016 · 12:24 p.m.