Player is loading...

Embed

Copy embed code

Transcriptions

Note: this content has been automatically generated.
00:00:00
Uh_huh So quick announcement if we need
00:00:11
a translation device it's available
00:00:13
outside there's japanese korean and
00:00:15
chinese available. So please go ahead
00:00:18
that's just right outside the door
00:00:19
thanks okay yeah I guess we're gonna
00:01:32
get started thanks a lot for coming to
00:01:35
our first a panel of chi two thousand
00:01:37
fifty our topic today is technology
00:01:39
transfer media research innovations and
00:01:42
we're gonna talk both about some of the
00:01:43
challenges in the space and also some
00:01:45
of the opportunities I'm from each line
00:01:48
I'm an assistant professor at the
00:01:49
university of waterloo. And it's really
00:01:51
an honour for me to be the moderator
00:01:53
today as you can see we have a but I
00:01:56
know but I superstar it's the I
00:01:57
researchers and we're gonna have a
00:01:59
really looking forward to having a
00:02:01
great discussion and in particular
00:02:02
we're really excited about making this
00:02:04
interactive. And having an audience
00:02:06
questions and answers so I was ask this
00:02:11
question you know what do a TI
00:02:12
researchers do you know there could be
00:02:14
several answers right some of us try to
00:02:16
understand users and their context
00:02:18
others of us I try to solve problems
00:02:21
through design and theory. But many of
00:02:23
us are actually innovators and we can
00:02:25
mend a novel experiences for users are
00:02:28
also the engineering side of it see I
00:02:31
and you know I it's really satisfying
00:02:33
for eyes when we have our papers
00:02:34
published at venues like higher waste
00:02:36
or other cup your it's at conferences.
00:02:39
But at some point in our career you
00:02:41
know we wonder you know is my research
00:02:42
actually having impact in the real
00:02:44
world. Right it's great to have your
00:02:46
kite paper but you know beyond the
00:02:48
academic community was actually
00:02:50
benefiting from this research. Now
00:02:53
arguably there are many ways of having
00:02:55
impact and we're gonna actually tackle
00:02:57
this question later on in our panel.
00:02:59
But one approach that's kind of an
00:03:01
established another sub feels of
00:03:02
computer science and buy a Madison is
00:03:04
this whole idea of technology transfer.
00:03:07
So just to sort of set the stage for a
00:03:10
panel today you know what we mean by
00:03:11
technology transfer here is I just this
00:03:14
idea taking your research artifacts out
00:03:17
of the lab. And going through the
00:03:18
process of either privatisation or
00:03:21
commercialisation. So with
00:03:22
characterisation it doesn't necessarily
00:03:24
have to be about you know
00:03:26
commercialising it or making money off
00:03:28
of it doesn't necessarily have to be a
00:03:30
for profit manager. But there's this
00:03:32
whole idea batted in it that you know
00:03:34
your artifact is now being used by
00:03:36
people other than researchers. Now
00:03:40
there is this question raised I don't I
00:03:41
don't like this I high back in nineteen
00:03:43
ninety six you know where they talked
00:03:46
about this whole idea that there's a
00:03:47
lot of research that's I would in the
00:03:50
field and this is back in nineteen
00:03:51
ninety six two decades ago. But there
00:03:53
wasn't enough that we could see in
00:03:55
terms of products coming out based on
00:03:57
that research right. So they had the
00:03:59
right in researchers and practitioners
00:04:01
and they had a great discussion around
00:04:03
you know how to be data research into
00:04:05
the hands of actual users. And this was
00:04:07
in you know two decades ago and since
00:04:09
then I would put you know has more than
00:04:12
doubled. And it's been increasing every
00:04:14
year you know not just a high this this
00:04:17
data is from high alone but you can
00:04:19
imagine there's with the SCW you become
00:04:21
mobile eight CI and several other even
00:04:23
at conferences are emerging. So I think
00:04:26
it would be fair to say that many
00:04:28
people still have sort of the same
00:04:30
concerns around tech transfer in eight
00:04:32
CI right there certainly are more
00:04:35
avenues and there's more awareness
00:04:36
around tech transfer today but are open
00:04:38
if you think about it is still fairly
00:04:40
low compared to how much research are
00:04:42
putting out. And there's also this idea
00:04:46
that we face sometimes increase
00:04:47
pressure from universities and also
00:04:49
from industry research labs and they're
00:04:51
increasingly asking us to demonstrate
00:04:54
transfer ability and you know
00:04:55
demonstrating our impact of the work
00:04:57
that we're doing you know while we're
00:04:59
working as professors are students for
00:05:01
research scientists in industry. Now
00:05:05
sometimes we have informal debates and
00:05:06
discussions with our colleagues at or
00:05:09
organisations or even at conferences.
00:05:12
But surprisingly you know we've had
00:05:13
little formal discussion since that
00:05:15
panel twenty years ago at high. So we
00:05:18
can you know this year it was the time
00:05:19
to sort of revive this discussion. And
00:05:22
you know let's see what we can I what
00:05:24
we can talk about in terms of the
00:05:26
future of technology transfer and it's
00:05:28
a research innovations. So the goal of
00:05:32
this panel today is basically know to
00:05:35
bring together researchers and
00:05:36
practitioners researchers from academia
00:05:39
and tremendous research labs to talk
00:05:40
about their views on technology
00:05:42
transfer I'm really we're gonna focus
00:05:44
on two aspects what makes high tech
00:05:46
transfer challenging and also where are
00:05:49
the potential opportunities and at the
00:05:52
same time are also interested in
00:05:54
hearing about your experiences
00:05:55
questions and concerns. And just start
00:05:58
on you know what what are some of the
00:06:00
things that matter for each I tech
00:06:01
transfer and you can you know how much
00:06:03
should be actually care is this even
00:06:05
the right thing to discuss labour not
00:06:07
really here advocating that everybody
00:06:08
should go out and commercialise their
00:06:10
stuff or you know create products out
00:06:12
of the research but because of the
00:06:13
pressures that we're facing nowadays
00:06:15
you know how much should be actually
00:06:16
care and what are some of the avenues
00:06:17
if we were to go through this route you
00:06:19
know what that can enable tech
00:06:20
transfer. So far I save the only have
00:06:24
about eighty minutes I'm gonna finish
00:06:27
my introduction in about five minutes
00:06:28
and then all handed over to each of our
00:06:30
panelists we're gonna talk hopefully
00:06:32
for three minutes that's the time. I've
00:06:34
given them and then I together we're
00:06:38
gonna debate and discuss forty five P
00:06:40
questions. And then open it up to the
00:06:42
audience for the last part of the panel
00:06:46
Sir panelists are you know the probably
00:06:49
don't need a formal introduction many
00:06:51
of them have of them actually Reading
00:06:52
the kind academy in others are
00:06:54
superstars where emerging in the field
00:06:56
but just introduce them formally we
00:06:58
have first of all very Stravinsky for
00:07:00
Microsoft research say hi. And you have
00:07:04
humans I from google. I still because
00:07:07
meant from other desk research at
00:07:09
Patrick validation from has to apply
00:07:11
the institute Chris Harrison from
00:07:14
Carnegie mellon university and energy
00:07:16
that to mark from the university of
00:07:18
Illinois at Urbana champaign. So before
00:07:21
they do their are brief presentations
00:07:23
I'm gonna tell you a little bit about
00:07:24
myself and I'm by moderating this panel
00:07:26
today. So I'm like a mention of a
00:07:29
professor at the university of waterloo
00:07:31
I started there around two years ago
00:07:33
after finishing my PHD work at the
00:07:35
university of Washington in two
00:07:37
thousand thirty as a while I was doing
00:07:40
my PHD I co founded a startup I would
00:07:42
my advisers trick or bright and ninety
00:07:44
call and currently the startup I've has
00:07:47
become you know bigger rented space out
00:07:49
in seattle. And the way the started was
00:07:52
it was based directly I have was taken
00:07:55
directly based on my dissertation work
00:07:57
on on lemonade this is where I use this
00:08:00
whole concept of a novel selection
00:08:01
based crowd source help tool for about
00:08:04
a web applications that allows users to
00:08:06
find the help that they need in the
00:08:07
context of the application just by
00:08:09
making a few selections in the
00:08:10
interface. And then we got a lot of
00:08:13
industry attention at that time when I
00:08:14
was a student. And and there's a lot of
00:08:17
momentum around commercialisation and
00:08:19
startups under centre for
00:08:20
commercialisation at you down in about
00:08:22
really excited and the really push this
00:08:24
in this direction to explore
00:08:25
commercialisation options. So we got
00:08:28
some initial seed funding from them you
00:08:30
know they give us access to experts in
00:08:31
terms of our local entrepreneurs of
00:08:34
even about some incubator space to set
00:08:36
up the company and then eventually now
00:08:38
you know we have major finding and we
00:08:40
have several employees and it's been
00:08:43
going on for a couple of years. Um but
00:08:46
you know I wasn't directly involved in
00:08:48
the process because I got my faculty
00:08:49
job and I decided I had to go away but
00:08:52
my adviser yeah Jacob rocky became the
00:08:55
CEO and he's been only for three years
00:08:57
now trying to run this company another
00:08:59
quite visor and eco he's currently the
00:09:01
CTO the chief technology officer. And
00:09:03
so he's and handling the engineering
00:09:05
side of things. So it's been quite a
00:09:07
rough road you know we've had a lot of
00:09:08
lessons to learn in this process and
00:09:11
you know now we're managing over
00:09:12
thirteen employees and we have fifty
00:09:15
plus trial customers and paying
00:09:17
customers are still trying to learn a
00:09:19
lot you know where is the actual value
00:09:20
for this innovation because when I I
00:09:23
you know got this organ and at the this
00:09:25
time innovation it was more about
00:09:27
helping and users find help that they
00:09:29
need and the context of the application
00:09:31
that they're using right it was really
00:09:32
driven by end user needs the research
00:09:34
contribution was all about
00:09:35
demonstrating that users actually
00:09:37
benefit from this but in the real world
00:09:39
you know and users don't actually get
00:09:41
to decide what kind of help an
00:09:42
application actually offers right it's
00:09:45
a business district decision that's
00:09:46
made by the creators of the
00:09:48
application. So our biggest challenge
00:09:50
is really you know how do we provide
00:09:52
some other clear value to the business
00:09:54
that's going to integrate that will
00:09:56
either application so that end users
00:09:58
can eventually benefit from it. So this
00:10:01
some of the things that we're still
00:10:02
iterating on in trying to understand
00:10:03
who are are doctors you know where the
00:10:05
stakeholders in this process. So much
00:10:08
representing this in much more detail
00:10:11
tomorrow at two thirty and in this room
00:10:13
my paper on looking at you know how we
00:10:15
might from sort of user centred design
00:10:17
to adoption centre design and what were
00:10:19
some of the decisions that you know
00:10:21
were critical on this time. And this is
00:10:23
one of the reasons why this panel got
00:10:26
put together we're invited to do this
00:10:28
and you know based on this paper to
00:10:29
because some of the questions that we
00:10:31
raised you know require discussion at a
00:10:33
larger scale with the community and
00:10:35
that's why invited people from industry
00:10:37
and from other academic units to see
00:10:39
you know okay what are some of the
00:10:40
things that we should be discussing in
00:10:43
there we all more tech transfer as we
00:10:45
certainly don't have all the answers
00:10:46
and my paper is not really about you
00:10:48
know providing an answer were saying
00:10:49
that you know this is the way we should
00:10:50
do a tech transfer every to shared an
00:10:52
experience but it certainly raised a
00:10:54
number of other questions that are
00:10:56
worthy of you know tackling as a group.
00:10:59
So without wasting more time I'm gonna
00:11:02
invite on our panelist I had to do the
00:11:04
representation so you can learn about
00:11:06
their tech transfer experiences and
00:11:08
then we're gonna move into the
00:11:09
discussion part. So we're gonna begin
00:11:13
way Mary hi everyone thank you for part
00:11:23
B and it's a pleasure to see such a big
00:11:25
room of people interested in this topic
00:11:27
and suppressed like that that Andreas
00:11:29
once again for Microsoft research I can
00:11:32
actually move in between product and
00:11:34
research for twenty five years. So
00:11:36
summarising twenty five years of tech
00:11:38
transfer experience is obviously gonna
00:11:40
force me to the abstract level I hope
00:11:43
that when you ask questions we can get
00:11:45
more into more specific stories about
00:11:47
what actually happened. Um can I
00:11:51
transfer is kind of a an abstract term
00:11:54
in an of itself it can happen in so
00:11:56
many ways being in the user interface
00:11:59
field for so many years actually tech
00:12:01
transfer happens just put the ideas
00:12:04
around user interface design a lot of
00:12:06
what I do that's tech transfer actually
00:12:08
is just talking to prodigy members
00:12:10
showing them things we've done in the
00:12:12
past giving them ideas for things they
00:12:13
could do in the future and that happens
00:12:15
on many many levels from you know
00:12:18
literally algorithms for how you do
00:12:19
visualisations to you know hard or
00:12:22
designs software designs it can happen
00:12:25
in many many different forms sometimes
00:12:28
we just share prototype so we'll email
00:12:30
prototype to a whole team and they'll
00:12:31
all use it by early and then they take
00:12:34
what they want from it sometimes we
00:12:36
don't even get credit for that kind of
00:12:37
stuff and it's just really getting
00:12:38
ideas to teens and they just move
00:12:41
forward with that oftentimes we do have
00:12:42
a champion now and they will give us
00:12:44
credit for that and and that's great
00:12:46
but there are times when we find out oh
00:12:49
alright yeah actually shipped and no
00:12:51
one even told us that so that's kind of
00:12:53
a sad truth about what happens in
00:12:54
research sometimes because we're a
00:12:58
little bit ahead of the curve hopefully
00:13:00
a lot ahead of the curve we've worked
00:13:01
with hardware that product teams have
00:13:03
never worked with before I'll give an
00:13:04
example we worked a lot with affective
00:13:07
us to censor looking at a electro
00:13:10
dermal activity or goblin X can
00:13:11
response to somebody and we knew a lot
00:13:15
about how to clean that signal how to
00:13:18
analyse that signal to get real value
00:13:21
out of it in terms of a human arousal
00:13:23
and product teams now you know we have
00:13:25
the Microsoft and shipping functions
00:13:27
they're very very interested in using
00:13:28
EDS a signal and those three and a half
00:13:30
years of expertise that we've gleaned
00:13:32
from just working with that hardware is
00:13:34
actually have something we can ship in
00:13:36
terms of algorithms ways of analysing
00:13:38
data et cetera. So it it's it's always
00:13:40
unclear how these tech transfers are
00:13:42
gonna happen. And as I said just
00:13:44
brainstorm strumming with product is is
00:13:46
a big way of how I spend my time and
00:13:50
what does it depend upon in terms of
00:13:51
getting the transfer to happen the
00:13:53
number one thing I've seen throughout
00:13:55
twenty five years it's trust sometimes
00:13:57
you have to work with the product team
00:13:59
for years before they trust you enough
00:14:01
that they'll listen to you and a lot of
00:14:03
times researchers make the mistake of
00:14:05
coming in as if you know their holier
00:14:07
than thou or they come from an ivory
00:14:08
tower and they know the best thing
00:14:10
actually product teams know what's best
00:14:12
for their customers that actually study
00:14:14
that really really hard for a long time
00:14:16
and they iterated their designs a lot
00:14:18
with their customers. So it feels us to
00:14:21
be more humble kind say hey we've got a
00:14:23
lot of experience in this domain but
00:14:25
problems can we help you solve maybe
00:14:27
something we learned can actually help
00:14:30
you would the problems you're trying to
00:14:31
solve. So often times it's actually
00:14:32
listening more and then designing for
00:14:35
as opposed to you know transfer your
00:14:38
technology right per se. Um the timing
00:14:42
is interesting I found in my career and
00:14:45
maybe I'm not doing research far out
00:14:47
enough I found seven your horizon is
00:14:50
actually the most frequent oftentimes
00:14:53
product teams what call me up in a guy
00:14:54
I didn't even know you were Microsoft I
00:14:56
just read this paper and we'd like to
00:14:58
ship this to me that will seem really
00:15:01
holding funky seven years later and
00:15:03
I'll be kind of embarrassed but I work
00:15:05
with them it's like fine let's let's do
00:15:07
something that's really current modern
00:15:09
and they like the ideas that that's
00:15:10
great but seven years seems to be a
00:15:12
magic number for me I don't know why.
00:15:14
Um another thing I've learned is that
00:15:16
fit and finish when you send things to
00:15:18
a product team is crucial if you just
00:15:20
send them some crappy poorly designed
00:15:23
user interface even though the
00:15:24
underlying technology is great they're
00:15:26
not gonna take it we have a great
00:15:28
designer on our team she helps make
00:15:30
everything just look forges and that
00:15:33
helps us so much in terms of getting
00:15:34
people to accept the technology. So
00:15:37
finish even if you're just directly
00:15:39
shipping we do a lot of visualisations
00:15:41
where we just directly ship it at the
00:15:43
at the office I the unfinished turns
00:15:46
out to be crucial will get feedback
00:15:47
from customers right away that didn't
00:15:49
finish isn't there's sweep learn learn
00:15:51
the hard way to do that. Um like I said
00:15:54
sometimes you have to work with the
00:15:55
team for years to make the tech
00:15:56
transfer happen a great example for us
00:15:59
is Augustine we actually started
00:16:01
working with them three and half years
00:16:03
ago by wiring users up you can see
00:16:05
pictures of users there with EDA
00:16:07
devices on their feet and we have them
00:16:09
on their palms and this was during
00:16:11
windows eight out of the box experience
00:16:13
so we were having users for the first
00:16:15
time never having seen windows a
00:16:17
install it out of the box and we had
00:16:19
them all wired up I have never seen so
00:16:22
many team members stuff into usability
00:16:24
research lab behind the one way mirror
00:16:26
and out into the hall before in my life
00:16:29
because you could literally for the
00:16:30
first time see a users real
00:16:32
physiological reaction to the unveiling
00:16:35
of windows eight and what do we think
00:16:36
we saw some great delight moments but
00:16:38
we also saw some horrific frustration
00:16:40
moments that they immediately fix and
00:16:42
think it was already in beta two so
00:16:44
this physiological data was so powerful
00:16:46
to them that it made them immediately
00:16:48
realise not only did they wanted in
00:16:50
their use research labs but they could
00:16:52
fix things in the product that they
00:16:53
wouldn't have known that they needed to
00:16:55
fix without that physiological signal
00:16:57
data. So that was very cool and now we
00:16:59
work with them very very closely in
00:17:01
actual product delivery based on a lot
00:17:03
of the sensor data. And then lastly
00:17:05
I'll close with simply giving
00:17:07
technology away for free is often
00:17:10
extremely valuable experience I have a
00:17:13
picture you're a parent guardian this
00:17:15
is a virtual we designed for parents of
00:17:17
ADHD children. And it was a simple
00:17:20
little out that had little techniques
00:17:23
to teach parents in the heat of the
00:17:25
moment with their child with ADHD how
00:17:27
to behave to model good behaviour for
00:17:29
their child. And it was very simple
00:17:31
there were twelve little interventions
00:17:32
that included things like deep
00:17:34
breathing but other more sophisticated
00:17:36
techniques for handling stress. And we
00:17:39
we did give them some sort bracelets
00:17:41
and we were able to send interventions
00:17:43
in real time during during a meltdown
00:17:45
with their child but which the sensor
00:17:47
presents away with that just keep yeah
00:17:50
well to my surprise just six months ago
00:17:52
I looked into the data set and there
00:17:54
were eight of the sixty families we're
00:17:56
still using the that was a really
00:17:58
telling moment for me we've released
00:18:00
other apps like that before where we
00:18:01
get lots of feedback from users and
00:18:03
sometimes are star all we just released
00:18:05
them out there and getting all kinds of
00:18:07
feedback an experience from that so to
00:18:09
me that's a tech transfer to that we
00:18:11
can also share with product teams about
00:18:13
what users really want. So I'll close
00:18:15
on that nothing can ask me for more
00:18:17
specific stories later thank you thanks
00:18:25
Mary a next the humour So my more
00:18:40
personal experience and you that
00:18:44
represents the remote past or present
00:18:47
employer. Um I will start with this a
00:18:51
problem or controversy opposition that
00:18:55
we should focus on for a conference for
00:18:59
this is sometimes that majority as the
00:19:03
sense of parsing it provides for ideas
00:19:06
on wall whole sound engineers product
00:19:09
manager for for cash on that set off
00:19:15
and even if we do it. Um the outcome is
00:19:20
very unpredictable and on control no
00:19:23
sensible of an original researchers
00:19:25
don't get credit. Um instead as a feel
00:19:30
we should focus more on I and impact
00:19:33
okay maybe maybe forms including many
00:19:36
of your professors teach grade students
00:19:39
who eventually on to shape and design
00:19:43
you sure product that's impact
00:19:46
fundamental understanding of the human
00:19:48
interaction process models theories
00:19:52
that's also impact and arguably more
00:19:58
for then then prototypes or
00:20:01
applications but regardless both trends
00:20:06
for an impact requires quite hard work
00:20:09
other require. Um inside innovation
00:20:14
inside is both technical and human
00:20:16
factors to to problem space. Um not
00:20:21
requires a lot of of innovation but you
00:20:25
know nation that matters of our field
00:20:27
as well as many other extending feels
00:20:30
tend to value novelty of lot more to so
00:20:36
so in innovation is another important
00:20:39
aspect of of having also requires
00:20:43
persistence that work and off and it
00:20:45
may take a year or two or seven or ten
00:20:48
or even longer to actually see the and
00:20:52
and using also like many other things.
00:20:57
It's friendly very hard for this as you
00:21:02
pursuit. Um long since this teenager
00:21:06
that's chinese word for like a phrase
00:21:09
for it so you should TVU meeting. It's
00:21:13
right time advantages place advice good
00:21:18
people around you and that's what
00:21:20
they're requires so to achieve success
00:21:24
so with that I'll just point really
00:21:26
quickly a few examples well involving
00:21:30
from my past twenty years he's or more
00:21:33
or working mister HCI space I often
00:21:38
wonder about was my PHD work with the
00:21:41
former one the one where five to
00:21:44
conceptualise the input control device
00:21:49
design space to some fundamental
00:21:52
variables nationalities that has to
00:21:54
with human performance. Um and you can
00:21:58
see there's nothing or five for all the
00:22:02
devices were made of fox's rubber band
00:22:06
user a like a lot tennis balls and so
00:22:11
to refine their two or three or four
00:22:16
design space. I I use this example
00:22:18
because I often wonder if this is too
00:22:22
academic to try to boring. Um but now
00:22:27
not an idea just you have to go quite a
00:22:29
few years ago but to run two sounds
00:22:31
that I remember I got a call from the
00:22:35
idea of exactly was only the the power
00:22:38
chip basis. And he said I need you to
00:22:41
do me a favour I have two very
00:22:42
important customers. Um and they want
00:22:45
to meet with you. And I was so
00:22:47
surprised until I met with them bizarre
00:22:49
DCT owns the of Nintendo the actually
00:22:52
right before we actually know a lot
00:22:54
about the space very surprised. So we
00:22:56
can see that same enquired or transfer
00:22:59
or not but I often wonder about what
00:23:03
transfer really means so actually the
00:23:06
coincides with the idea of sign you
00:23:07
another from ah no actually control
00:23:13
space but more banal and I have another
00:23:19
example wow I IBM with the call that
00:23:22
like problems mean that soccer on
00:23:24
trying to reinvent the mouse. And I
00:23:29
still think it's a wonderful idea is
00:23:30
probably still for about hundred
00:23:32
million and maybe still available today
00:23:35
but then ultimately it's not terribly
00:23:38
successful in terms of changing the
00:23:40
very multi you pretty simple device in
00:23:43
in in in that straight and I wonder why
00:23:47
and I'll I'll say customer time is to
00:23:50
place has to do with the fact that
00:23:53
ultimately idea moved on to more recent
00:23:56
story business. And so the can is it.
00:24:01
Um and here's another fibreglass for
00:24:05
right now fifteen years a initially
00:24:08
called shark and then call shape writer
00:24:10
Michael Jackson typing today Google in
00:24:13
at different places. Um well it started
00:24:17
eating enough time the first four five
00:24:19
four paper published was very high in
00:24:23
on this idea of using a stroke gestures
00:24:27
representing a whole world. And uses
00:24:29
this method to process them. Um and it
00:24:34
was very exciting but also we felt the
00:24:36
time was not quite right at the time
00:24:39
the whole mobile industry with moving
00:24:40
away from us right okay so I had to TV
00:24:46
not like like very based plastic stuff
00:24:51
on so the whole industry was moving
00:24:53
from that are that paradigm of of
00:24:55
gestures and and and and computing and
00:24:59
I debated with many important enough
00:25:01
paper no future that's going until two
00:25:05
thousand seven one Steve jobs make this
00:25:08
keynote. I don't know I've fallen and I
00:25:11
was still most people don't believe
00:25:13
even mine calls colleagues would would
00:25:16
imagine believe in the future oh we
00:25:19
worked very hard on it my former
00:25:21
student problem presents and I work
00:25:24
with other engineers started a startup
00:25:25
company the first are on the right we
00:25:28
literally created a a whole new
00:25:30
paradigm today's adopted in all major
00:25:33
more of window on the mobile like
00:25:35
forms. And I don't know has well over
00:25:39
appealing users right that's quite
00:25:42
exciting we change a lot of people's
00:25:45
you like to experience on a daily basis
00:25:48
or even hourly basis just as exciting
00:25:51
terms of impact is there with see then
00:25:55
a five six years ago six ago a mother
00:25:57
whose daughter has severe disability.
00:26:01
And fancy are we okay you say that I
00:26:07
after all this time. She can finally
00:26:09
right send text messages and journal.
00:26:12
And the amazingly increasing the
00:26:15
collective for like an hour and that is
00:26:18
also quite exciting part. So the point
00:26:23
is that I felt very important to be
00:26:25
that you can find you can have you have
00:26:31
two many things to yeah I think human
00:26:47
up next year to be out of this research
00:26:50
it it's a pleasure to be here and I'm
00:27:01
as well very pleased to see the tendons
00:27:04
here. So my name still because when I'm
00:27:06
user interface researcher at other desk
00:27:09
research just to give you a bit of
00:27:11
sense of the desk is a design software
00:27:14
company we make software for all sorts
00:27:17
of three D modelling design task simple
00:27:20
two D and three D this is a you work
00:27:22
three about a the before visualisation
00:27:24
done by just I can or our group. So we
00:27:27
have about seventy five hundred
00:27:28
employees the research group is about
00:27:30
fifty people within UI research are
00:27:33
actually quite small there's five full
00:27:34
time employees. And then something
00:27:37
interesting we also have as a research
00:27:38
transfer team. And that's also the five
00:27:40
employees is a really interesting model
00:27:43
so how works is when we want to
00:27:44
transfer technology we have a team
00:27:46
that's with in research that's
00:27:48
dedicated to doing this. So they're
00:27:50
close enough to the research team they
00:27:51
actually get the whole from us that
00:27:53
they understand understand the research
00:27:55
the other interesting thing you'll
00:27:56
notice about this visualisation is a
00:27:58
close forced were to the centre and the
00:28:00
centre is basically the how far we away
00:28:03
we are from the the top of the company
00:28:05
just the C up. And we report to the
00:28:08
chief technical officer of of the
00:28:10
company and both the CO and CT or are
00:28:12
very forward thinking which really
00:28:14
helps for doing tech transfer so let me
00:28:16
give you some examples of of what we've
00:28:18
done been able to do in the past five
00:28:20
or six years so type twenty two
00:28:23
thousand and nine we publish a paper
00:28:25
call to a club switch embedded videos
00:28:26
inside of tool tips. And actually
00:28:28
within the same year it was integrated
00:28:31
into our all of our flagship products
00:28:32
including autocad which is used by over
00:28:35
ten million users. So is a bit
00:28:38
different from the a magical seven your
00:28:40
number that Marius was talking about
00:28:42
chronicle is others that list two
00:28:45
thousand and ten and this this project
00:28:50
in looks at capturing video but also
00:28:52
work flow information related to the
00:28:54
video to help users learn how to use
00:28:57
use the system and and provide you are
00:28:59
thought you content in the software.
00:29:02
And in two thousand twelve we listed as
00:29:04
a a free tech previously it was
00:29:06
referencing. And in twenty fourteen is
00:29:10
actually launched as the official
00:29:12
protest prod that was twenty fourteen
00:29:16
we developed a system called extreme
00:29:18
for embedding electronics within three
00:29:20
D fabrications. And it is now
00:29:23
integrated into much mixers and you
00:29:25
tool. And finally tracker which was
00:29:29
published that chi twenty fourteen last
00:29:32
year system for doing sketch based
00:29:34
animation where currently in the works
00:29:36
again with the research transfer team
00:29:38
of developing this as I thought
00:29:40
application and hopefully what we're
00:29:42
gonna we're gonna see that this year.
00:29:43
So also very exciting. So how do we do
00:29:46
all this in in such a short time frame.
00:29:50
So here's overview of the transfer a
00:29:52
process we start with a research paper
00:29:55
publication usually comes with the
00:29:57
interactive prototype then we developed
00:30:00
in collaboration with the transfer
00:30:02
group we work very closely with them we
00:30:04
wanna make sure the research is
00:30:05
maintained the concepts behind the the
00:30:07
research innovations once that's done
00:30:10
we actually start working with the the
00:30:12
product teams and now we're more hands
00:30:14
off this is done by the research
00:30:15
transfer team and they work with the
00:30:17
the product teams to integrate and then
00:30:19
the ultimate goal is to have the
00:30:20
research transfer team completely
00:30:23
handed off to to the actual product
00:30:25
teams. So they can begin working on on
00:30:27
the next new innovation. So just some
00:30:31
lessons learned from from doing this it
00:30:35
really helps to have a culture
00:30:36
innovation within the within the
00:30:38
company. So obviously within the
00:30:39
research group there's a culture of
00:30:40
innovation that all the desk as I said
00:30:42
is quite quite for thinking out it's
00:30:44
not just coming from the top you need
00:30:46
at at the lower levels sometimes you
00:30:48
have a base present saying I love this
00:30:49
stack but it's the developer that says
00:30:52
oh not so interested and it doesn't go
00:30:54
anywhere so you need that culture of
00:30:55
innovation in seconds is very
00:30:59
comfortable some of the other talked
00:31:01
referred already you really need to be
00:31:03
persistent you actually need to be
00:31:04
willing to adapt as well so this is a
00:31:06
screen shot of protocol that we publish
00:31:09
that that west. And we're very excited
00:31:12
to show to the CEO of the company. And
00:31:14
you responded what the hell is that I
00:31:16
but with no further work adaptation and
00:31:21
and design to really focus on on the
00:31:23
customers now have to say that a bass
00:31:26
is the CE O.'s not only advocate of the
00:31:28
system is a user of the system
00:31:29
uploading his own work flows for some
00:31:32
of our products finally also just like
00:31:35
to just the mindset of the academic
00:31:37
community for us that be part of a
00:31:39
research group we need to be publishing
00:31:41
we also want to have some impact on on
00:31:43
products and sometimes we get responses
00:31:45
like this from your so in this case the
00:31:47
one is he was completely supportive of
00:31:48
the research and was novel that was a
00:31:50
contribution. But said well you can
00:31:52
develop this today so is that along the
00:31:54
chi conference and that something we
00:31:57
have it here discussion about yeah am I
00:31:59
don't do it. Right make can be a a next
00:32:10
year and the first of Illinois I am
00:32:24
Reggie that and then as this then
00:32:27
professor of computer science at UIUC
00:32:30
I'm also the co founder and chief
00:32:33
scientist at all proposals which is
00:32:35
that a bear yes for start up. And my
00:32:40
tech transfer experience actually began
00:32:43
dad type twenty thirteen when we
00:32:46
publish the paper website guys design
00:32:49
mining the web and it was the the
00:32:52
culmination of my dissertation work at
00:32:55
stanford. And website guys look that
00:32:58
how to use the same data mining and
00:33:00
machine learning techniques never you
00:33:03
know successfully used to analyse
00:33:06
content to understand design. And an
00:33:10
interesting thing happened when we
00:33:12
problems that paper overwhelmingly
00:33:16
people would come up to us and ask us
00:33:19
when can we use the the system and the
00:33:21
the software you felt. And so the the
00:33:24
website guys team we decided to think
00:33:27
about this in figure out how we wanted
00:33:29
actually you know release this
00:33:30
technology we found that the academic
00:33:34
setting was an ideal because there I
00:33:37
you know the incentives incentive
00:33:40
structures were basically misaligned
00:33:42
there was a lot of more engineering
00:33:45
work we would have had to do to
00:33:46
actually release this to everyone and
00:33:49
so you know if we did it in academe
00:33:53
yeah it we basically I would need to
00:33:55
spend two more years and not getting
00:33:57
you research credit for it. And so you
00:34:00
know we decided that the best way to
00:34:02
actually get this tool and the and
00:34:05
system in the hands of users was to
00:34:07
start a company. Um and you know we had
00:34:11
never started a company before so we
00:34:15
weren't you know we we were thinking
00:34:17
about how to do this and I it really
00:34:20
how because we have the support of the
00:34:22
entire website guys team out of the
00:34:25
seven people who worked on the system
00:34:27
by robots you know wanted to be
00:34:29
involved in one way or another to make
00:34:33
this possible. And so in September
00:34:36
twenty thirteen yeah I wanted a
00:34:38
proposed and we bootstrap four year I
00:34:42
postpone my you why you see a
00:34:45
appointment for year Maxine and was
00:34:48
just about to I go off to MIT well she
00:34:53
didn't and and Jerry quit his job. So
00:34:57
we basically gave ourselves I your to
00:35:01
you know build the core infrastructure
00:35:04
and really figure out what what are
00:35:07
users wanted. And at the end of the
00:35:10
year we raised a small see round a
00:35:13
couple million dollars from any and
00:35:16
interesting weights. And now you hear
00:35:20
all these horror stories about how
00:35:22
difficult it is to work with you see
00:35:24
forums and we actually you know in fact
00:35:27
have found it to be the opposite from
00:35:29
day one indian increasing a really
00:35:32
supported or vision. And you know have
00:35:35
provided value beyond just the money.
00:35:37
So you may be wondering you know what
00:35:42
is our proposed you will make
00:35:44
impossible for users to search explore
00:35:48
and understand design data across the
00:35:51
web to make better design decisions
00:35:54
more quickly with greater confidence.
00:35:59
And so if you'd like a sneak P you can
00:36:03
centre for our mailing list on the side
00:36:05
here. Um you can also start by or boost
00:36:08
at the exhibition hall tomorrow thanks
00:36:12
thanks for you to next we have press
00:36:21
yeah I like yours okay so how the story
00:36:34
for abruptly became thinkers and start
00:36:36
at Carnegie mellon university and in
00:36:38
two thousand nine I was by myself
00:36:40
Julius Schwartz and our comment
00:36:42
advisers got hot and we were working on
00:36:45
a tabletop system or you can see at
00:36:47
sovereign over right there. And here's
00:36:49
a clip from our original submission
00:36:51
video which basically sure that we were
00:36:52
doing a finger a part differentiation a
00:36:56
touchscreen knows this big clunky sort
00:36:58
of large and and slow apparatus that
00:37:01
sort of mid nineties accuracy and at
00:37:03
the very beginning to really have a lot
00:37:05
of kind of commercial potential and so
00:37:07
when we submitted it to risk twenty ten
00:37:12
it was a fairly rejected and not
00:37:15
particularly convincing and then we
00:37:17
actually retool didn't submit it again
00:37:19
yeah the next time I twenty one where
00:37:21
it was also rejected. And then we
00:37:23
submitted it all to me to withstand
00:37:25
twenty eleven where was accepted and
00:37:27
during this sort of year iteration here
00:37:29
it is we extended it quite a bit and
00:37:31
most notably added on mobile snare you
00:37:34
know it's not just the tabletop but
00:37:35
expanding it to mobile devices a mobile
00:37:37
use cases and then really sort of a a
00:37:39
record. And by this point kinda going
00:37:42
through this forced evolution we sorta
00:37:44
get it we started getting more more
00:37:46
excited about the technology and so we
00:37:47
continue to work on the technology
00:37:49
after it except in so we build a
00:37:51
portable demo that we could go around
00:37:53
and start showing people in here's an
00:37:54
example that prototype to still fairly
00:37:56
crude but we get a show people in of a
00:37:59
during the time that we met us and one
00:38:02
lady who to make come Marcia when we
00:38:04
found it back excel in twenty twelve
00:38:07
and then we did a series a we wrapped
00:38:08
up in twenty fourteen and women gonna
00:38:10
startup company sense. And we're now
00:38:13
can employees into offices of
00:38:15
engineering office in Pittsburgh in a
00:38:16
business office in San jose. And things
00:38:19
are are going really well we actually
00:38:21
just really get traction in the market
00:38:23
you're sort of a quick guide "'em" of
00:38:24
real of what it looks like today just
00:38:26
software only solution one of the
00:38:28
biggest things we had to contend with
00:38:29
was making it very high accuracy low
00:38:31
power and also make use of the
00:38:33
commodity built an accelerometer
00:38:34
actually this was not possible to
00:38:36
recently the only use the accelerometer
00:38:38
the built in accelerometer. So users
00:38:41
that and a pile of machine learning to
00:38:43
be able to do real time classification
00:38:45
of touch type. And of course we've also
00:38:47
expanded use cases so here you can
00:38:49
scroll the finger but do for example
00:38:50
tech selection with a knuckle. Um also
00:38:53
be able to taste for example with a
00:38:55
kind of a context menu and similarly
00:38:58
you know other forms of media such as
00:39:00
you know web pages and images and so on
00:39:02
PDS we can also reserve the strip
00:39:04
knuckle invite for different actions
00:39:06
also notably being able to lasso for
00:39:09
example. And tap right kinda knuckle
00:39:11
tapper context menu to summon a
00:39:13
interactive options so it's all got
00:39:15
filled out by virtue of this sort of
00:39:17
forced evolution and then most recently
00:39:19
we've also been able to port our
00:39:20
software the the smart watches which is
00:39:22
an interesting form factor after
00:39:24
"'cause" it really small having some
00:39:25
extra dimensions in your input can be
00:39:27
really interesting and powerful. And so
00:39:30
we started show this around tool on a
00:39:32
client we sort of got traction this
00:39:34
recently especially these accelerometer
00:39:35
have improved a member finally kinda
00:39:37
getting traction in the market. And
00:39:39
please do them on devices really soon
00:39:42
one of the biggest challenges I've
00:39:44
noticed we can talk about this more
00:39:45
questions is actually we think of the
00:39:47
mobile device industry we think it is
00:39:48
this really fast moving innovative a
00:39:51
domain but actually found is that
00:39:53
people are really quite deeply
00:39:54
conservative in changing user
00:39:55
experience that even adding a single
00:39:58
gestural put into the experience
00:40:00
actually can be a tremendous shock even
00:40:03
if they find it very sort of obvious
00:40:05
and sort of this should be a slam dunk
00:40:07
do include still a lotta conservatism
00:40:09
in terms of what to include in the user
00:40:11
interface and how to change one that's
00:40:13
been incredibly successful it really
00:40:15
since the advent of multi. So we can
00:40:17
talk more about that in questions okay
00:40:25
last but not least okay bye which I
00:40:34
think so much so I hear any good and
00:40:36
it's a little controversy so it's
00:40:38
thought since like relaxed me at some
00:40:40
here. Um plus my have tech transfer
00:40:44
wreck it is really ridiculous all
00:40:46
rights oh my slides not supposed to
00:40:49
green right. Maybe then standard one
00:40:53
green as well. Right. So we're supposed
00:40:56
to think about a bunch of questions and
00:40:57
I think the last one to what extent
00:40:58
should we care about tech transfer. So
00:41:01
zooming all the way out snow we talked
00:41:03
about the perspective of individual
00:41:04
researchers let's think about the
00:41:06
perspective of labs for a second so
00:41:07
imagine you found it a research lab.
00:41:09
And there's like two things you could
00:41:11
do you can phones on the right side a
00:41:12
research lab. And the mission of that
00:41:14
would be to have people think ten to
00:41:16
twenty years out. Um and I think bell
00:41:18
labs is kind of the traditional model
00:41:20
that we've kind of still living by and
00:41:22
some places and then you can have the
00:41:23
idea of nor you lap and so used to work
00:41:26
for their sparking Microsoft research
00:41:28
and I left send the starred in academia
00:41:29
seven years ago and since a better
00:41:31
chance actually consult a larger number
00:41:33
of companies and got a bit of an
00:41:34
overview. And so for example bosses
00:41:36
nick's insanely reject model where like
00:41:38
if your research or you can work as a
00:41:40
research until you get the first
00:41:41
product in the you go with the product
00:41:44
into the protein you can Dakota you
00:41:45
wings essentially just thing in in the
00:41:47
product that so the interesting
00:41:51
question is you know what this is about
00:41:52
so the goal of research lab was you
00:41:54
know big companies have these things
00:41:55
because they want to have someone with
00:41:57
scouts at the future twenty is out so
00:41:59
if the main in a of the company's birds
00:42:02
managing they wanna have something they
00:42:03
can go into right some that's
00:42:04
imaginative much the comp the copy
00:42:06
machine company wouldn't be greatest
00:42:08
when years later when the patents kind
00:42:09
of feel you know if you had another
00:42:12
business such as the PC business for
00:42:13
example and in this particular case
00:42:16
with this didn't work out I guess it.
00:42:18
So if you were in academia you know I
00:42:19
mean why do we have research labs what
00:42:21
you think of like an a witch and so on
00:42:23
you know this is idea that someone has
00:42:24
to check would the questions of
00:42:25
society. So and if you think of the
00:42:28
traditional successes like large
00:42:30
successes of traditionally our last
00:42:32
research labs like things like the
00:42:33
transistor you mean augmentation by
00:42:35
talking about certain is well known are
00:42:38
you laser printing in all things that
00:42:40
serving and so on. And the and I think
00:42:44
the satin usable being in that research
00:42:45
lab is that ninety nine percent of
00:42:47
stuff will never come out. It's a high
00:42:48
risk high gain kind of business. So the
00:42:51
like the outcome is no tech transfer.
00:42:53
Um no this is weird case where there's
00:42:55
some sort of an accidental thing you
00:42:57
were planning to use of screwed up and
00:43:00
it happened to be you know work out in
00:43:01
your that's great. Um depending on up
00:43:05
the perspective the company. And then
00:43:07
no the weird thing is I think it really
00:43:09
works out this typically no tech
00:43:11
transfer just more or less replace the
00:43:12
company work for right. So for example
00:43:14
had zero exactly how this work out with
00:43:16
the PC you wouldn't think of zeros as a
00:43:18
copy machine company anymore just to
00:43:19
continue to be you think of you know I
00:43:22
would not be staying from of my apple
00:43:23
mac poke your would be standing in
00:43:24
front of my xerox machine. And not to
00:43:26
make photocopies alright so the idea of
00:43:29
of O capital or lapses to make huge
00:43:31
kind of step forward side of very
00:43:33
unlikely ratio obviously right. So if
00:43:35
you think of surfing you know you can
00:43:37
think of. I think of them than even
00:43:38
more love surfing kind of looks like
00:43:40
this you win the warrants sewer wave
00:43:41
comes in then you hop on it if you've
00:43:43
ever been surfing you realise is
00:43:44
totally not whole surfing works. Um
00:43:47
this what circuitry looks like standing
00:43:49
in the war is sitting them. It's very
00:43:50
realistic actually that they're looking
00:43:52
at the rising at some point someone
00:43:54
start Spelling something directly like
00:43:55
what is going on. And then you know
00:43:58
because people go to some extrapolated
00:43:59
point with I think the waves gonna hit
00:44:01
and then hopefully when the way gets
00:44:02
there that's when the getting on which
00:44:04
means that you know most of surfing
00:44:06
looks like this and I think that's a
00:44:07
bit of the experience of the of the of
00:44:08
the research lab and the and and the
00:44:10
way. So alright so couldn't we have
00:44:14
both couldn't you have the capital R
00:44:15
research lab and say like you know well
00:44:17
but we're gonna give it little button
00:44:18
incentive to researchers and you know I
00:44:21
think the research labs of work for
00:44:22
certain there was some incentive to I
00:44:23
mean it was a good incentive to do tech
00:44:25
transfer anything can we get a little
00:44:26
bit of tech transfer in the side. And I
00:44:28
think it's interesting idea so on the
00:44:30
one hand I think there's not shot and
00:44:32
of once this debate a lot of research
00:44:34
that right know this kind of thing
00:44:35
question roll over and over. So you
00:44:37
could say like well we're gonna be I'm
00:44:39
sure there's I think we such as what
00:44:40
think a little bit about the stuff I
00:44:41
have couldn't that make an interesting
00:44:42
product and I think there's something
00:44:44
to be had on the flip side I think I
00:44:47
think and that's I think that should
00:44:48
danger of the whole thing is if now the
00:44:50
incentive model kicks in to the extent
00:44:52
that people say like wow I really would
00:44:53
like that. I think there's a certain
00:44:55
risk that people start thinking oh you
00:44:57
know what I really work on things they
00:44:59
can actually sit and six to twenty four
00:45:01
months in in our importantly you've
00:45:02
converted your lap tool already at and
00:45:05
the you know maybe instead of having
00:45:07
people work on ten euro things you know
00:45:09
this is not personal random appear
00:45:10
right smart matter rain the faces you
00:45:13
know up to the singularity which is
00:45:14
kind of the timeframe are really you
00:45:16
know you make your own map of what you
00:45:17
think the tenant when your questions
00:45:19
are you might find yourself working on
00:45:21
things we're actually a lot of
00:45:22
companies on it you might be work
00:45:23
working on mobile devices for example
00:45:25
I've a lot of that right so I'm not you
00:45:28
know I'd much shorter time frame. And
00:45:30
including things where a lot of
00:45:31
industry has already happened at this
00:45:32
point. So at the same time a piece that
00:45:35
puts up the rest of like who's gonna be
00:45:36
the next going apart and honestly I
00:45:39
mean are we all know that's a tall
00:45:40
order but it would be great if a few of
00:45:42
us actually you thought so in
00:45:44
conclusion I think should we detector
00:45:45
and I think absolutely and I think this
00:45:47
to break case we can do it number one
00:45:49
nets Schuman is the great example here
00:45:51
if you can research long enough if you
00:45:53
have attended when your agenda well you
00:45:55
know just wait ten years and you have
00:45:57
something to transfer I remember a an
00:46:01
advert for things to contact one asked
00:46:04
with the great way of making this creek
00:46:05
on TV says well just by the best rating
00:46:07
then we two generations all right. So
00:46:11
the other is you accidentally get
00:46:12
something I think we all occasionally
00:46:13
do that you know I've got a bunch of
00:46:15
projects in the space and you need to
00:46:16
recognise that and that's a good time
00:46:18
for doing a transfer on the other hand.
00:46:21
I think if we all get sucked into like
00:46:22
thinking oh my god we only to going to
00:46:24
kick starter I think there will be you
00:46:25
know we feel impact on the research
00:46:27
communities in the sense that the
00:46:28
products projects a research project
00:46:29
would be less exciting. So I think
00:46:31
that's a bit of a danger. So if we
00:46:33
wanna wanna go to kickstart or maybe
00:46:35
you don't have to go through is
00:46:35
tremendously complicated exercise
00:46:37
you're really easy ways of doing that
00:46:38
right now it wasn't this one here. Um
00:46:41
buzz Aldrin when Walker would like a
00:46:44
word with you you promised me more
00:46:45
school news and so they got face book
00:46:46
alright okay I right thanks Patrick
00:46:58
extraction dollar panelist so now
00:46:59
you've heard their perspectives are
00:47:01
actually a little bit over time it's
00:47:04
not surprising but so we have about
00:47:08
half an hour in thirty five minutes
00:47:10
I'll pose a couple of questions that
00:47:11
may be cut down my one questions an
00:47:13
opening up to the audience pretty soon.
00:47:14
There's we have a great turnout and
00:47:16
never really wanna know what you think
00:47:17
and what you wanna ask our panelists.
00:47:21
So one question that I wanted to ask
00:47:24
was okay what re it's resources and
00:47:26
skills are needed to pursue tech
00:47:27
transfer you know this is one of the
00:47:29
first questions that every student or
00:47:31
professor or a research scientist asks
00:47:34
you know how do I navigate my internal
00:47:36
organisation and how do I get by and it
00:47:38
will be mentioned some examples for
00:47:39
example your transfer team is within
00:47:42
the research unit so that makes things
00:47:44
easier rise in other cases you have to
00:47:46
actually go out and figure out okay
00:47:47
with the right person to talk to just
00:47:49
wondering if you guys can comment a
00:47:50
little bit on that and just based on
00:47:52
your own experiences well I think you
00:47:56
already for several of the same
00:47:58
personally perseverance product and
00:48:01
adapting to the to the world that
00:48:03
you're trying to transfer to listening
00:48:06
you know being a good listener and I
00:48:09
think something people leave out along
00:48:11
as being fun you know just being fun to
00:48:13
work with those are really long way I I
00:48:18
think the just adding to this when I
00:48:20
was a Microsoft they remember I had a I
00:48:22
had an outing with Chris Bradley was
00:48:24
they had of all the slapped. And you
00:48:26
give me a piece of his mind white a
00:48:28
terrible to work with people like us.
00:48:29
And the and you said well interesting
00:48:32
just dumping idea A into the product
00:48:34
things and now. It's a kind of a one
00:48:36
night stand. And and we're not really
00:48:40
committed to the things that are and I
00:48:41
think here's a good point. And so I
00:48:43
think the for me I think eating sushi
00:48:45
playing guitar work used to work with
00:48:46
Perkins the commitment is a big thing
00:48:49
that came out in the talks to rely
00:48:50
currently that was say hi Jerry quit
00:48:52
his job and you know you postpone your
00:48:54
faculty positions that you wanna talk a
00:48:55
little bit about yeah I think you know
00:49:00
there's two ingredients especially I've
00:49:03
you when you're thinking about startups
00:49:04
that matter the first one is have you
00:49:07
you know come across the problem that
00:49:09
you're trying to solve that will have
00:49:11
an impact and then the other one is do
00:49:13
you have the right team to you know be
00:49:16
solving that problem and so you know we
00:49:19
just felt like we were the writing this
00:49:21
all that particular problems so that
00:49:23
may be easier. It's all sort of about
00:49:27
personal choice as well so I decided
00:49:29
not to participate in my started
00:49:30
directly you know I always wanted to be
00:49:31
an academic I was raised by academics
00:49:33
right side it's great that you know
00:49:35
chicken and you wanted to take we done
00:49:37
this and they took time off and they've
00:49:39
been taking time off for the last three
00:49:41
years but eventually that you have to
00:49:42
go back. So there's that bottle like
00:49:44
you can just you know once you commit
00:49:45
to something you can't just leave it
00:49:47
right you can expect other people to
00:49:48
take over and well things forward I
00:49:51
think there's a long term commitment
00:49:53
that's required the W were saying this
00:49:55
persistent thing as well. Um but you
00:49:59
had a lot of success in fairly short
00:50:00
time how does that work you seen
00:50:05
innovations call from high papers to
00:50:07
products and like really necessary or
00:50:08
in some cases it's it's really
00:50:10
important to demonstrate you when
00:50:13
you're when you're going to pursue
00:50:15
transfer no organisation and if and
00:50:17
it's it's really important to
00:50:19
disseminate your results to all levels
00:50:21
of the company both right right to the
00:50:23
top. But also what the the lower levels
00:50:26
such as that product managers that'll
00:50:28
actually be implementing things. But
00:50:30
the the real key to what we've been
00:50:32
able to do is that the transfer group
00:50:34
that we have within our a research team
00:50:36
and and for our part there is a lot of
00:50:39
dead occasion as well like for
00:50:41
chronicle for example I did a lot of
00:50:43
the the design work so I was removing
00:50:45
myself from what my my primary role at
00:50:48
the research group is is to do more
00:50:50
more research and the reason I did that
00:50:52
was I wanted to ensure the the research
00:50:54
ideas were were maintained very talked
00:50:57
about trust and I think it's a bit too
00:51:01
much trust it to just hand off your
00:51:03
research "'cause" you never know it's
00:51:05
gonna happen when when that happens I
00:51:06
said it's only at the final stage for
00:51:08
us that we we do that. You know
00:51:10
complete and off we we switch some soft
00:51:12
and at that point the the product
00:51:15
hopefully is already taken care of
00:51:16
itself anybody have a comment about
00:51:21
this from the audience question related
00:51:23
to this and then you come up to the
00:51:32
microphones there's one at the back as
00:51:33
well yes yes RS D.'s are really
00:51:58
critical they have to be a special
00:52:01
breed is they have to be not just a
00:52:02
software engineer but a software
00:52:04
engineer that can work with researchers
00:52:06
I hope you all know what I mean by that
00:52:08
because soccer engineers tend to be
00:52:10
extremely detail oriented they want
00:52:13
things very nailed down and specify
00:52:15
whereas a research ST has to be able to
00:52:17
handle people like me or my designer
00:52:19
becomes then with very high level
00:52:22
nothing's really set in stone yet but
00:52:24
we want to build it anyway. And so
00:52:26
those those RST D.'s have a magical
00:52:29
blend of being able to handle people
00:52:31
like us and still build systems that
00:52:33
work and sometimes we have to make the
00:52:35
specifications as we go we do
00:52:38
eventually get there are but they have
00:52:40
to be very patient and tolerant
00:52:41
actually enjoy the process yes Oh I
00:52:52
list that racial Google so I that this
00:52:56
idea back in cracked and what I would
00:52:57
let you won't comment a little bit is
00:53:00
we don't want to just throw things of
00:53:02
the wall off and I end are getting more
00:53:04
more more fascinating by the teachers
00:53:06
that collected or my things in use. And
00:53:10
so you know I and that's that of
00:53:11
working with people we get product out
00:53:13
completely fascinated about what the
00:53:16
engineers working with the
00:53:17
instrumentation ingenious. And
00:53:20
redesigning the measures and metrics
00:53:22
that actually stray what we eat it so
00:53:27
that it becomes clearance that you
00:53:29
could have your company understandable
00:53:31
measures and metrics is actually ran
00:53:32
something like HCI we're very concerned
00:53:35
about people how to use things it's
00:53:37
just not audience numbers but it's
00:53:39
actually the utility to people about
00:53:42
the metric style that we need to
00:53:45
demonstrate the value of whatever it
00:53:47
was that we design so I would love to
00:53:50
have you like maybe one time whatever
00:53:52
how you re so what the measures and
00:53:54
metrics are a success rather than as I
00:53:56
think human just that you just handed
00:53:59
over make goes right. And I love
00:54:01
marries points about persuasion that go
00:54:03
forward and I think that the
00:54:04
persuasions starts to become about the
00:54:06
evaluation piece as well which HCI is
00:54:09
very focused on so I I'd like to get
00:54:11
your thoughts on how you have to be
00:54:12
created about that as well. remember I
00:54:17
set up sometimes just letting tech
00:54:19
trance or just go out to the field for
00:54:21
free one of the main reasons we do that
00:54:23
is then we log everything and we see
00:54:25
exactly what areas of the user
00:54:27
interface people are excited about or
00:54:29
use a lot more than comment back to us
00:54:31
about I don't actually believe in just
00:54:33
lost without some contextual data about
00:54:35
what the task is the users trying to
00:54:37
sell. So we always per provide a way
00:54:39
for them to email us back and ask us
00:54:41
questions about things then we get a
00:54:42
better sense of what the context around
00:54:44
these lot numbers are but yeah the way
00:54:46
the way we can collect data today oh my
00:54:48
gosh it just makes the field so much
00:54:50
more exciting so I totally agree with
00:54:52
you no sorry though yeah I just I just
00:54:56
wanted to follow but Protestants very
00:54:58
very similar we have these sort of two
00:55:00
main channels which is what we LCIP
00:55:02
which is the data that get sent back
00:55:03
was based on the usage behaviours but
00:55:05
we also have you know he usually email
00:55:08
addresses for feedback for specific
00:55:10
prodded things things like that and
00:55:11
it's really nice to see people sending
00:55:13
us messages saying how you know they
00:55:15
really enjoy using a system us or
00:55:17
something like that the particle
00:55:19
research we yeah we were a little
00:55:21
nervous about you know when we put the
00:55:23
switch genders on it's a community a
00:55:25
server so we would see people creating
00:55:26
content if they were using it. And so
00:55:28
we saw gallery just start filling up
00:55:30
with video content so we knew we we're
00:55:32
on to something for sure yeah just add
00:55:36
to that definitely had a proposed we
00:55:39
spent a year building the port
00:55:41
infrastructure because we knew that the
00:55:43
first time we build a product that
00:55:44
would be the one one but we knew we
00:55:46
wanted you know it'd be idea driven
00:55:49
tool so we actually spent the time to
00:55:52
build that infrastructure so we would
00:55:53
be just a matter of weeks to prototype
00:55:56
different products. So that we could
00:55:59
get user feedback and see which one
00:56:01
stuff what the comment on the broader
00:56:07
notion fun data merry touched on this
00:56:10
briefly earlier that this when you work
00:56:13
with the existing product team. Um
00:56:17
there's also a lot a lot to learn from
00:56:19
them. And often researchers will come
00:56:22
with ideas that aboriginal or driven by
00:56:26
innovation I'm driven by some Diego
00:56:28
inside. But unattended product
00:56:32
engineers and project managers the
00:56:35
especially when it comes to user
00:56:36
experience everybody's a human being
00:56:38
everybody think they are expert at
00:56:40
understanding innings to so but
00:56:42
sometimes we do have very pretty deep
00:56:44
inside. It's the process of deepening
00:56:48
the understanding of the users needs as
00:56:50
well as the ecosystem to get to your
00:56:53
original vision. I think it's a quite
00:56:55
challenging process but but the
00:56:58
activities you want to continue to
00:57:01
lower the problem space and that's all
00:57:04
for the challenge of technology
00:57:05
transfer he's because you are you don't
00:57:09
own the problem space then you propose
00:57:12
a solution often the very recent
00:57:14
because you don't think that you
00:57:16
actually understand the problem space
00:57:17
the deep now maybe it's you know
00:57:20
selection ideas is required but when
00:57:22
you think about these other factors
00:57:24
associate with it. Um that mean on me
00:57:27
me me not. BS audience of them anymore.
00:57:30
Um so it's it's a balance oh Uh okay
00:57:37
here your whole sub UC Berkeley I only
00:57:40
ten to ask this question about getting
00:57:42
at that kind of fundamental idea of
00:57:44
research and I think partially all you
00:57:46
have tremendous success in the car
00:57:48
outcomes that you had I think first or
00:57:50
will you be celebrating this across the
00:57:52
community out words of people
00:57:53
understand this. But also wanna revisit
00:57:56
this idea you know the kind of bell
00:57:58
labs these things that come up about it
00:58:00
almost I sent some little bit of a
00:58:02
moral obligation to really do these
00:58:04
high risk sort of research projects
00:58:07
that have it's not clear they're gonna
00:58:09
pay off immediately and I think one is
00:58:12
modelling that for the next generation
00:58:14
of students and sort of practitioners I
00:58:16
I see more students coming in and it
00:58:18
really stated how am I gonna have a
00:58:19
startup coming in to do these things.
00:58:21
And that's a great outcome and very
00:58:23
positive the push them to that but I
00:58:25
also worry about maybe being too near
00:58:27
Owens them having a hard time Reading
00:58:29
the landscape I spent a lot time you
00:58:31
know my courses are treated the same
00:58:33
trying to show and demonstrate how lots
00:58:36
of really visionary work way before
00:58:38
it's time now you're using everyday I
00:58:41
kind of think about in you know you
00:58:44
might see people on the streets that
00:58:45
are wearing cool hipster kind of shirts
00:58:48
about dan's that broke up and they're
00:58:50
kind of like these are like the cool
00:58:51
hands like you don't know this band it
00:58:52
was really cool back in the day so what
00:58:54
the call would be could we have like
00:58:56
hipster researchers that are wearing
00:58:58
all T shirts interval bell labs that
00:59:02
you know it's Intel research and so
00:59:03
they would become like this was a call
00:59:05
at like there was really cool stuff
00:59:06
that happened here like you know like
00:59:07
"'cause" I wanted to concerts that that
00:59:09
was really cool but time side I mean
00:59:12
really just pushing on this idea that I
00:59:15
think it'd be great to showcase in
00:59:17
really a discussion of the parts are
00:59:19
that you're such a great handle have
00:59:20
together about how to get just not lose
00:59:24
this almost like I do feel talk of
00:59:26
moral obligation to really push people
00:59:28
to take those challenges if a a lot and
00:59:30
some things here. And some don't things
00:59:35
are so the when when I was living in
00:59:39
the bay area in two thousand you could
00:59:41
get any research on the planet of the
00:59:43
time because no one wanted to be in
00:59:45
research at the time everyone wanted to
00:59:47
be in the dot dot com I think it was
00:59:49
called at the time. Um I remember kids
00:59:52
were like we could of the bachelor's
00:59:53
degree like your prior on because they
00:59:55
could get a job as a web admin that's
00:59:59
on like this and they would make eighty
01:00:00
thousand dollars a year and then you
01:00:03
know eighteen months later they were
01:00:04
all really driving captioning going
01:00:06
back to school and I think a little bit
01:00:08
of this is kind of I I think this and
01:00:11
external pressure we haven't
01:00:12
acknowledge the for example the the
01:00:13
presence of the I phone for example. So
01:00:16
I'd have to London to do that I mean so
01:00:17
the lab time at has a tremendous
01:00:19
entrepreneur culture because the
01:00:21
founder of the institute is hostile
01:00:23
planet the founder of SAP right is if
01:00:25
you use germany's most valuable company
01:00:28
very successful startup you could argue
01:00:30
and so the students wanna be like him.
01:00:32
Um and at the same time right now if
01:00:33
you look around there is so they still
01:00:35
do by phone hype right and then there
01:00:37
is always kick starter too. And so I
01:00:40
think the the this is a little of the
01:00:42
forces that push right now and so I
01:00:44
think the certain extent these other
01:00:45
force to be reckoned with right now and
01:00:47
I think you know if the economy goes up
01:00:48
and down so it's just really hard to
01:00:50
focus right now on research when
01:00:52
there's so much opportunity I think
01:00:53
that's a businessman assume on my
01:00:54
student yeah want to response that I
01:01:00
thought that was great question
01:01:01
interesting also there's a very deep my
01:01:07
dealer deep cultural difference between
01:01:09
silicon valley and the rest of the
01:01:10
world. So frankly knows bell labs
01:01:13
viewed as well IBM research or even
01:01:15
talk to some degree not review that
01:01:17
typical signal experience. And if you
01:01:20
look at it beginning things and out
01:01:21
like it started such a clean others for
01:01:24
when you know what price for for their
01:01:25
work. But the the the Sperry was very
01:01:29
much more to for Norman research is
01:01:32
more about owning the information and
01:01:34
on it all the way. And it so I overall
01:01:39
today I feel like trend is more towards
01:01:42
that kind of model that if you really
01:01:44
make about yourself and you really only
01:01:47
on it and you you put yourself rather
01:01:50
than having a a large industrial
01:01:53
research lab where that you could base
01:01:57
for a long time or present really
01:01:59
really can't you it's more as a sort of
01:02:02
simple model that is something we
01:02:04
breakthrough happen then it that often
01:02:05
happens naturally. Um it's really
01:02:08
interesting that how how that works and
01:02:10
I would say well drawings the canal you
01:02:12
you know being Google or half or
01:02:14
faithful the the model bell labs this
01:02:18
just not regarded as as as the models
01:02:22
that's that's probably very true I just
01:02:28
wanted to mention that experiment
01:02:29
that's going and Microsoft research
01:02:30
right now we've actually split
01:02:33
Microsoft research into two groups and
01:02:36
we'll see how this goes but researchers
01:02:38
that we're very interested in the one
01:02:39
to two year horizon I see john nodding
01:02:41
maybe he was gonna ask that's but
01:02:44
researchers that were interested not
01:02:46
shorter term horizon were invited to
01:02:48
move over into MSR next which is a new
01:02:51
organisation under purely that is
01:02:54
intended for all those researchers are
01:02:56
expected to ship in the next one to two
01:02:58
years there just expected maybe being
01:03:00
given a ton of resources to enable that
01:03:03
to happen and business assistance et
01:03:05
cetera et cetera the rest of us who
01:03:08
wanted and desire to stay in long term
01:03:10
basic research have have stayed under
01:03:12
Jeanette waiting. And it operates is
01:03:15
business as usual with the same
01:03:17
resources we always had in the same you
01:03:19
know drive to change the world. So
01:03:23
it'll be an interesting experiment to
01:03:24
watch and just throwing it out there as
01:03:25
a possibility of one model who knows
01:03:28
how it works I mean I'm a sars tried
01:03:30
many many things to address this issue
01:03:32
about getting some immediate tech
01:03:34
transfer and some really long term you
01:03:36
know groundbreaking foundational
01:03:38
research. So we'll see how it goes me
01:03:40
check back in in two years oh okay hi
01:03:46
right on top marks are research
01:03:48
actually but darker question that I
01:03:50
then that I mean I've had worse
01:03:52
frustrations where the research that we
01:03:54
do basically doesn't have a Khomeini
01:03:57
overeating sponsors. And I are back on
01:04:00
some of your stories I mean academics
01:04:01
tend to start and then you start up
01:04:04
that kind yeah surety but your story is
01:04:07
that get that going to a place where it
01:04:10
did have a home but it meant kind of
01:04:12
moving to someplace and you we tell I'm
01:04:15
really kind of intrigued by this notion
01:04:17
that there's this group finds a place
01:04:20
within but it just gonna and how that
01:04:23
yeah okay yeah research labs especially
01:04:26
the cap powerpoint and coming up with
01:04:28
ideas that could don't have a home in
01:04:31
excess budget and then when you do. So
01:04:36
yeah I can I can start with that within
01:04:37
our our group or actually really
01:04:40
fortunate to be working on you know
01:04:42
research lab potentially based on
01:04:44
Patrick specification are are D because
01:04:46
we are doing short term answer we also
01:04:50
have the flexibility to work on the ten
01:04:52
or twenty your products and like you
01:04:54
say a lot of these don't have a whole
01:04:57
minutes interesting some these more
01:04:59
forward thinking prod X of the ones
01:05:01
that get to the media and the press. So
01:05:04
when we we talk to the you know the
01:05:06
higher levels of our company that the
01:05:07
senior vice president three no P you
01:05:09
know was the projects they're very
01:05:11
supportive as well because they want
01:05:13
the company to be you know seeing as I
01:05:15
thought leader amongst amongst others
01:05:17
software companies amongst the research
01:05:19
community. So there's supportive of the
01:05:22
one to to your tech transfer for sure
01:05:24
but we're not constraining to only
01:05:27
working on that's we're actually
01:05:28
innocent but uni unique space we we do
01:05:30
get a nice blend of those two
01:05:32
activities so just about that that
01:05:35
sounds like tech transfer by going to
01:05:37
the press and so Reading demand on that
01:05:39
side is that after I think in the the
01:05:43
point of the press isn't necessarily to
01:05:45
do tech transfer it but it definitely
01:05:47
gets as recognition as as you know
01:05:49
about you bowl a valuable group so for
01:05:51
sure when we have opportunities to do
01:05:53
that we take advantage I think at mikes
01:05:56
of research and gadget was the best way
01:05:58
to reject reaching the proteins and
01:06:01
granted I think this is I I don't think
01:06:09
there's a magic for or I mean it really
01:06:12
depends on company size and it's a
01:06:16
product portfolio as well as the is
01:06:20
present model so you remember one
01:06:24
bright idea research there's no lots of
01:06:27
debate every now and then about for
01:06:30
example should you have a larger
01:06:33
projects that are driven by a
01:06:35
collective we gonna mostly more
01:06:37
commonly American reason about where
01:06:39
the next technology scum from verses a
01:06:43
more individual smaller team invention
01:06:47
or research driven initiatives that
01:06:50
doesn't really have critical mass. And
01:06:53
you can see success in both both sides
01:06:55
and often some of the biggest excess
01:06:58
like well let's unix or or all these
01:07:00
things come from likely very small in
01:07:02
the UI for small teams so it's it's
01:07:06
hard to say but generally. Um I feel
01:07:09
just most companies really prefer to
01:07:12
have a larger critical mass are fewer
01:07:14
problems "'cause" that's just easier to
01:07:17
manage and and and major success. Um
01:07:22
but I don't think there's a magic from
01:07:27
the actually I and one line to the so
01:07:30
so indifferent got I mean it live along
01:07:32
lines what you said the question is how
01:07:34
far should allow to be allowed to
01:07:37
deviate from the mission statement of
01:07:39
the company. So I was a park in two
01:07:41
thousand and two thousand one and so
01:07:44
those if you were not there at the time
01:07:46
this was kind of when the dot com
01:07:48
bubble kind of burst and everyone got
01:07:49
real nervous and what you could see a
01:07:52
park at the time and I'm sure you can
01:07:54
see the same phenomenon other companies
01:07:55
is oh my god things we going that we
01:07:57
need to focus on the core agenda item
01:08:00
everyone gets nervous and instantly
01:08:01
printed becomes important again for
01:08:03
xerox and you know and you know you
01:08:05
people might have seen this in other
01:08:06
places so in some ways I mean I think
01:08:10
at this moment what you see is I think
01:08:11
the companies we actually we allow a
01:08:14
research lab to have with the bell labs
01:08:15
model which you know if an invention
01:08:17
happen it would surely replace the
01:08:18
company itself rather than just being a
01:08:20
transfer somehow other companies were
01:08:22
doing really really well so if you
01:08:24
research your your research overlap in
01:08:26
some ways you check out the stock you
01:08:28
see like well the stocks go not company
01:08:30
ridge they're gonna be super relaxed if
01:08:32
the if there's if the going gets a
01:08:33
little tougher than the company will
01:08:35
try to pull you into more the core
01:08:36
general somehow in but I but with the
01:08:40
reason one saying this is because I
01:08:41
think ultimately I think the real value
01:08:43
proposition of research lab is the
01:08:44
thing where you do something that is
01:08:46
you know not completely off topic but
01:08:47
something that the company can so
01:08:49
related and the example of the peace
01:08:51
you mentioned earlier with just a
01:08:52
little bit outside the conference on a
01:08:53
xerox right but I mean that you know
01:08:55
have been a little bit closer somehow
01:08:57
right I mean could completely
01:08:58
transformed xerox right. So so I mean
01:09:02
that's well that's a difficulty right I
01:09:05
mean some ways the question is you know
01:09:06
how far can they be so you can still
01:09:08
like we market when it comes. That's
01:09:11
allowed enough that actually make some
01:09:13
real change the company rather than
01:09:14
being the feature in the product have
01:09:20
them unusable that wood is you and I I
01:09:24
just want to if I listen to you it's is
01:09:27
really interesting and I think for
01:09:29
community very important discussion
01:09:31
because it if we can prove for if we
01:09:34
can show ways of of how innovation move
01:09:38
from research into into companies it
01:09:42
would would strings and also a few oval
01:09:46
and that's why I think it beyond to the
01:09:48
the the stories and the in the
01:09:51
individual insights which you present
01:09:54
it you very very interesting and very
01:09:56
different also if we if you look at all
01:09:59
we ooh ooh ooh almost is able to to
01:10:03
bring products research results very
01:10:06
quickly into products to this much more
01:10:08
difficult things which I think maybe
01:10:11
just just to it gives testimony or so
01:10:14
what I think that we need to research a
01:10:16
German more a profound research agenda
01:10:20
in our community where we really look
01:10:22
in detail at all high specific inside a
01:10:26
moving from research into into practise
01:10:29
into into product development and with
01:10:32
all this different mechanisms where
01:10:33
startups always in huge companies like
01:10:36
Microsoft I think I mean there's it
01:10:38
would be very helpful for the community
01:10:40
to know more about it. And I also think
01:10:43
that second instantly high unique
01:10:45
things that are specific to better
01:10:47
let's insides because because it will
01:10:50
all on the level of of natural signs
01:10:53
inside cot results in human computer
01:10:56
interaction of a very align to context
01:10:59
to to to certain cultures of usage and
01:11:04
things like that so so it's not not at
01:11:07
all that easy to even to express and
01:11:10
even to express them and share them
01:11:12
among within the research yeah that's
01:11:16
an excellent point I mean in some sense
01:11:19
having been involved in you know the
01:11:21
executive committee the committee of
01:11:23
second for so many years that is what
01:11:25
we try to do it this conference right
01:11:27
that is the reason we're presenting our
01:11:28
work here is to show you you know what
01:11:30
the future holds and what might become
01:11:33
an actual product but speaking from
01:11:35
practical for practical direction we
01:11:39
try to track are eight CR contributions
01:11:41
to for instance microsoft. And it's
01:11:44
really really hard for all the reasons
01:11:46
that I presented it my first slide
01:11:48
because just having a conversation with
01:11:50
the product and ends up in a tech
01:11:52
transfer where the you know we're not
01:11:54
and so how do you map that stuff it's
01:11:56
very difficult but I agree with you
01:11:58
that as a field it will us to take the
01:12:01
really big and obvious ones and map
01:12:04
those map those in time and history and
01:12:07
show you know maybe even how much money
01:12:09
was made there that might be very
01:12:10
convincing the some people are just how
01:12:12
many millions of people influence which
01:12:14
might be very interesting but I agree
01:12:16
with you be a great maybe a great side
01:12:19
project puts occur community take on to
01:12:21
kind of map history be awesome I know
01:12:23
bill Buxton has done that for input
01:12:25
devices it took a lot of work but it
01:12:27
was a great example of that I was
01:12:30
thinking even more I mean the six
01:12:32
sisters would be nice to but also I
01:12:34
mean speaking about the problems you
01:12:36
know because you learn best if you see
01:12:38
how complicated it is the to bridge the
01:12:41
gap between the research functions and
01:12:43
development culture and and it's it's
01:12:46
it's not only about the celery and and
01:12:49
and and temporal perspectives so just a
01:12:53
one just to add to that is one thing
01:12:55
that we're not really talking about is
01:12:56
just how difficult it is from a legal
01:12:59
perspective detect transfer obviously
01:13:01
within sight of something like an S are
01:13:03
we can talk more freely to your
01:13:04
employees and and you can still it's a
01:13:06
challenge a lot of the community that
01:13:08
we see today we dissipating in in high
01:13:10
are academics and and take tech
01:13:12
transfer from universities is
01:13:14
extraordinarily difficult. And then I
01:13:16
think within the last five years we've
01:13:17
seen research lab models move to be
01:13:19
increasingly restrictive in terms of
01:13:21
sharing so pretty much everything
01:13:22
lisa's I know that happening inside of
01:13:24
and the star next. I I do not know why
01:13:26
not really prevent anything remote for
01:13:28
now I yeah I mean so even inside of
01:13:30
Microsoft office is this bottle forming
01:13:33
and certainly you know Google is
01:13:34
relatively secretive about a lot of the
01:13:36
things certainly inside the to relax
01:13:37
and so on so they atmosphere has
01:13:39
changed to mow to be from sort of the
01:13:40
bell labs arab where people used to go
01:13:42
to conferences speak much more freely
01:13:44
now I know why isn't contracts to look
01:13:47
at people the few apple people that are
01:13:48
here are all you know mum's the word
01:13:50
even people that are we know from
01:13:52
Amazon and so on it so I think there's
01:13:54
a whole cultural share that further and
01:13:57
maybe not that it was never easy but
01:13:59
it's further inhibiting this kind of
01:14:00
cross a nation of ideas a lot of people
01:14:03
are kind of speculating it kind even if
01:14:04
they see something interesting they say
01:14:06
you know how how difficult is that
01:14:08
gonna be Dana tech trend to read of
01:14:10
senior otherwise and and all these
01:14:11
things are adding up not doing research
01:14:14
single innovation process sis inside
01:14:17
companies would be also very and as a a
01:14:20
and a very strong suspicions if big
01:14:22
companies that all would along too
01:14:24
because I mean that's basically that
01:14:26
well a value proposition. yeah I would
01:14:30
say going back to my somewhat bases
01:14:35
that if I feel I really should think or
01:14:38
feel should focus on impact and should
01:14:40
fit for some fundamental research
01:14:42
should focus on the the foundations of
01:14:46
the field rather than focusing tech
01:14:49
transfer because tech transfer okay
01:14:51
having many forms you can be randomly
01:14:54
can be a startup or could be just know
01:14:56
someone else eventually copy idea. And
01:14:59
run with that there's many forms but
01:15:02
it's very hard to say have one sort of
01:15:06
still tickle tickle path that that
01:15:08
these two not transfer but rather a few
01:15:11
of the more fundamental research and
01:15:13
more fun than inside models and the
01:15:15
theories. And a new paradigms and I'm
01:15:20
to conventions and really straight and
01:15:22
by you reaching user experience and
01:15:25
addressing real use any these I think
01:15:29
oh well you should do I should do more
01:15:31
it's a great discussion we only have
01:15:36
five more minutes that I think a couple
01:15:38
more questions and I am Caroline from
01:15:41
KTHCN and well I actually agree that we
01:15:46
should prioritise impart over thinking
01:15:49
on how do we transfer the research
01:15:52
innovations on the other hand. But I I
01:15:56
like to label myself as the baby
01:15:57
researcher I'm just starting in this. I
01:16:00
had to write my master thesis and that
01:16:02
took so much work and writing papers
01:16:04
takes so much time and then they get
01:16:07
rejected and you don't want to show
01:16:09
them to everybody else because maybe
01:16:10
you can save it to another conference
01:16:12
so the thing that Chris was mentioning
01:16:14
about secrecy in between research labs
01:16:19
in the industry. I think the way you
01:16:22
don't so happens a round of any combine
01:16:26
environment or mean it's not secretly
01:16:28
but I don't hear that we get extra
01:16:32
points for thinking how to transfer our
01:16:37
innovations in a way that helps getting
01:16:42
that impact that they could actually do
01:16:45
we have we we write papers that are for
01:16:48
us we organise this conferences there
01:16:52
are for us and maybe not for other
01:16:54
interpreters or people that are willing
01:16:56
to take our ideas and bring them to the
01:16:59
hands of the you so you that's wrong
01:17:04
right. But we should change it and how
01:17:09
how that reflecting the indication of
01:17:12
baby researchers that we have today and
01:17:18
if I think it's sort of partially a
01:17:21
roll of the academic community to
01:17:23
decide what what papers are accepted to
01:17:26
the to the conference supposedly those
01:17:28
are the ones that are gonna that are
01:17:30
gonna be making and and yeah so I think
01:17:33
for students and I know like the
01:17:34
culture when I was a a grad student
01:17:37
tech transfer not only do not get
01:17:39
points you probably lose point if
01:17:41
you're thinking about that and not
01:17:42
thinking about what are you gonna
01:17:43
publish next year. Um so as long as
01:17:46
those papers are making it in fact
01:17:48
don't see anything anything wrong with
01:17:50
that and it's up to the sort of
01:17:52
academic community today side you know
01:17:55
which direction there we going and what
01:17:56
type the of papers are we are we
01:17:59
valuing and deciding are you can gonna
01:18:02
have enough impact that they they
01:18:04
belong at the conference my my biggest
01:18:06
concern is that when we write papers
01:18:09
maybe they are not targeting the
01:18:11
brother audience so of course. they
01:18:15
decided okay this is something that
01:18:17
maybe we have a huge impact and
01:18:21
something important that we need to
01:18:22
communicate but doesn't I don't think
01:18:24
that it it it goals you on these
01:18:26
community it stays in the paper. And on
01:18:31
those about boxes to the completed but
01:18:34
I'm very things like that so and and we
01:18:37
don't get extra points for making who
01:18:39
are you to be the O.'s or going to
01:18:40
attack dog or things like that. And if
01:18:43
we want to stay at in the research
01:18:46
career you need to focus on getting
01:18:49
those the the points that counts can
01:18:51
can you use more specific and gives you
01:18:53
point. Um well that maybe something
01:18:59
that I need to learn. But I don't get a
01:19:01
station for example when I do you to be
01:19:03
you yeah and somebody TV and so yeah do
01:19:11
you think that that's something to
01:19:13
change in the academic environment to
01:19:15
support I think it's a technology
01:19:17
transfer to a broader audience. I and
01:19:23
always put those you two minutes on
01:19:25
your CV well I definitely empathise
01:19:29
with that situation I think you know as
01:19:32
a missing the incentive structures are
01:19:35
misaligned like you did take the
01:19:37
research and make it a real product.
01:19:39
It's you know you were you were going
01:19:42
to get research credit for that so it's
01:19:45
a completely kind of different mind
01:19:47
shift that you have to kind of you and
01:19:49
your team have to commit to doing yeah
01:19:52
yeah I I I think if if you ever change
01:19:57
jobs latest that's the latest norms
01:19:59
that which you realise that any
01:20:01
currency is made up citations papers
01:20:06
all made up if you have or switch your
01:20:08
product in your TV collapses into one
01:20:11
line that says sixty three peer
01:20:13
reviewed publications if you want for
01:20:16
yourself to say that you two
01:20:17
compressions or startups I mean in your
01:20:21
whole life I don't just mean in this
01:20:22
current situation you can choose what
01:20:24
you want to cheapen your life you
01:20:25
consider metric for yourself you know
01:20:27
same way people say wanna walk ten
01:20:28
thousand steps a day whatever it is you
01:20:31
can say like my goal is so in in a
01:20:33
little bit earlier Johnny lee you is
01:20:36
that at Google right now he was in a
01:20:38
couple of years ago he was similar
01:20:39
popularise the idea of having you two
01:20:41
videos you just did a couple of do we
01:20:42
just felt you'd is new to the reach
01:20:44
several of them which six minutes. And
01:20:46
when I started also supplying institute
01:20:48
have well I have my class not knew
01:20:50
Johnny lee it into anyway does anyone
01:20:52
else in the room believe me right and
01:20:54
the one thing is you know when you
01:20:56
think about some gives you points the
01:20:59
multipoint systems as you use to choose
01:21:02
yeah concurred that with that that this
01:21:05
is a note of any better system in me
01:21:07
change quite dramatically and I had to
01:21:10
agree that you know it it there a
01:21:16
sometimes there's a competition in
01:21:18
terms of having ideas but generally
01:21:20
over time you'll fine. Um ideas
01:21:23
actually plenty and it's really the
01:21:27
deep knowledge the inside game over
01:21:31
long time period time that actually and
01:21:33
matters. So often when you started to
01:21:36
feel like I have this idea first and
01:21:38
nobody else thought of it. Um but
01:21:41
you'll find a lot of ideas people can
01:21:43
think of. But but it's getting to the
01:21:45
next step getting eventually to then
01:21:48
think like is really what matters more
01:21:51
so this is also to like global for
01:21:53
example there's no some parts sound
01:21:57
like Google keeping secrets actually oh
01:21:59
I think about doesn't really care that
01:22:02
much about secrecy in terms of
01:22:04
technical technical know how it's
01:22:07
mostly just more the focus is on really
01:22:11
and using impact rather than academic
01:22:12
publications that for the cultural of
01:22:15
of course mobile access different
01:22:17
because we wanted to be projects that
01:22:19
one of the generate certain ecosystem
01:22:22
inside so it has to manage that process
01:22:26
different. But anyhow well think it or
01:22:31
almost out of time actually are out of
01:22:32
time if people don't mind spending five
01:22:34
minutes and we can take one more
01:22:35
question and then the analyst will be
01:22:36
here for couple of minutes if you wanna
01:22:38
come talk offline that'll be great too
01:22:40
but before we get to that I do wanna
01:22:42
just say some perspective has been
01:22:45
highly north American now obviously we
01:22:47
have battery from Germany but there are
01:22:49
other ways in the czech transfer
01:22:51
happens in other companies and there
01:22:53
certainly are there examples that we
01:22:54
couldn't include today but we hope you
01:22:56
know future panels good to include
01:22:58
these perspectives. And you know we
01:23:00
want that this discussion does become a
01:23:02
regular part of DCI community as
01:23:04
something you've talked about you know
01:23:05
there's still are lots of unanswered
01:23:06
questions anonymous a real experts in
01:23:09
this area. So would be great to keep
01:23:11
this discussion cory and I really wanna
01:23:13
see a big thanks to all of our panelist
01:23:15
you know what we to be on this panel I
01:23:17
think the provided a really valuable
01:23:19
perspective and individually each of
01:23:20
them right and really just doing
01:23:22
stories and unique experiences so gonna
01:23:24
be kind of I wanna stay will take a
01:23:33
couple more questions than if you need
01:23:34
really for lunch at least go ahead So
01:23:37
we can to the end of the session
01:23:38
without talking about incentive That's
01:23:40
a very strange to me I'm because that's
01:23:43
from from case the mobile like grew I'm
01:23:46
I'm wondering because it's it's
01:23:49
obviously important to me it wouldn't
01:23:50
be asking how you see a correlation
01:23:53
between what incentive structures there
01:23:56
are for financial gain and ownership of
01:24:00
what comes out. And success in the and
01:24:03
that was all or does it not matter to
01:24:08
people just do it. Because they love
01:24:10
research does it not matter that they
01:24:13
get the patent a hundred percent ten
01:24:15
percent they get millions or is a good
01:24:18
start so since I switched from research
01:24:23
lab to academia the incentive has
01:24:26
increased tremendously financial
01:24:28
incentive incentives also I can tell
01:24:31
you what it was also and so at the
01:24:33
research labs what was before if you
01:24:36
look at the incentive structure I think
01:24:38
there was an incentive in the sense
01:24:39
that it would look very good on your
01:24:40
performance review would have tech
01:24:42
transfer. But I think IBM actually
01:24:44
Schumann should come on this I think in
01:24:46
in IB and there's actually clicked back
01:24:47
at the Microsoft network that was non.
01:24:49
So once you transferred there was no
01:24:53
personal incentive to see this go
01:24:54
through except in the I think this
01:24:55
actually led to this thing that Chris
01:24:56
probably told me when you said like you
01:24:58
just wanna P Gideon into the product
01:24:59
you move wanna move back right next
01:25:01
paper there was no incentive model it
01:25:03
but maybe maybe in that so maybe
01:25:04
American but come on this as well maybe
01:25:06
it's changed from the company
01:25:08
perspective there was a very strong you
01:25:10
were perspective for like you know to
01:25:11
see you thing come to fruition but the
01:25:13
company did not give you an incentive
01:25:15
to just opposed to know since the left
01:25:17
research labs and I mean academia at
01:25:20
this point and so my lab for example
01:25:22
just allows me to find long patents of
01:25:24
a like two and then these owned by me
01:25:26
at the expense of paying for them
01:25:28
complete different game. So Chris of
01:25:32
research were given instant to throw
01:25:34
patenting we get several thousand
01:25:38
dollars per patent depending on how
01:25:39
many people are on the pattern with
01:25:40
you. So actually more of an incentive
01:25:42
in a way to just pat pat pat instead of
01:25:44
two track transfer most the people that
01:25:47
do tech transfer do because there's
01:25:48
extremely passionate about it yeah I
01:25:52
think the university models is really
01:25:54
interesting and I only know the kinda
01:25:55
north American model but you'll find
01:25:57
that academics generally minutes or a
01:26:00
personality driven thing but generally
01:26:02
are not particularly motivated by money
01:26:04
which is arguably good reason to do a
01:26:05
startup company. And all the CD tracks
01:26:08
from your students in your publications
01:26:10
and so your department may not actually
01:26:11
look upon a particularly favourably
01:26:13
other than sort of assess a success
01:26:14
story for the school or something. And
01:26:16
so there's not a very good incentive
01:26:18
structure in a across any metric for
01:26:20
professors to do an if you look at a
01:26:22
lot of that if you think about the the
01:26:24
most well positioned exactly like the
01:26:25
HCI faculty at Stanford or Berkeley a
01:26:28
lot of those guys don't have any
01:26:29
startup commissioning of Michael burns
01:26:30
generic Paulson setup it might advise
01:26:32
just are to be single those guys are in
01:26:33
the supreme spot that spent also media
01:26:35
startup and they don't and so it in
01:26:38
this is really you know true almost
01:26:40
across the entire field and and that's
01:26:41
why non this panel you know there's a
01:26:43
handful of of H instead of that it
01:26:45
happened out of no twenty years of
01:26:47
research at chi is pretty startling
01:26:49
given that you know we're a three
01:26:50
thousand person conference now. So I
01:26:52
think there's a very as your correctly
01:26:54
saying there was this big difference an
01:26:56
incentive model I think you see the
01:26:57
people that actually do push it out as
01:26:59
a startup or push it through to a
01:27:00
product group and kind of step back
01:27:02
from research tend to be up the people
01:27:04
that want to see just come to fruition
01:27:06
get out there so they're very impact
01:27:07
driven as opposed to monetary or
01:27:09
whatever yeah see very reward is also
01:27:13
you know somewhat broader sense that is
01:27:16
that is not just transfer but but when
01:27:20
you have success that really make a
01:27:22
difference to and user experience then
01:27:25
you are more likely to get anymore
01:27:27
anywhere any company are more likely to
01:27:29
get more investment or resources more
01:27:32
engineering more even product
01:27:34
management not marking to to go with
01:27:37
the product and already mention or
01:27:39
holding back that is not a formal both
01:27:44
recognition reward. So a lot of things.
01:27:47
You you've actually scale and that that
01:27:49
record store. And that itself is some
01:27:53
formal formal or actually on the
01:27:57
academic right I I do feel like I kind
01:27:59
of incentive structures vary from
01:28:01
university to reverse the actually UIC
01:28:04
recently started kind of
01:28:05
entrepreneurial program for their
01:28:07
professors to encourage more tech
01:28:10
transversely you would have the
01:28:12
opportunity if you it's something
01:28:14
through the university to you know your
01:28:16
own startup and actually higher your
01:28:18
own grad students to make it possible.
01:28:21
So yeah I think we right hi hello
01:28:30
morning this is one from by the centre
01:28:34
for by implementing to columbia's with
01:28:36
some of them are you can I get a a very
01:28:39
practical insured question if I am the
01:28:42
leader of the innovation team in my
01:28:45
company and I got the new product I
01:28:48
built a new problem is it my
01:28:50
responsibility to go out and sell it to
01:28:54
make money over it oh well that's
01:28:56
something I have to chair without
01:28:58
comment sensation teamwork some like
01:29:00
that what's the relation between making
01:29:02
money and and innovation as a leader of
01:29:05
innovation I'm I'm not sure I can
01:29:12
completely understand just that the
01:29:14
question if the casing you you were
01:29:17
researcher within the I am my
01:29:18
researcher on my on that either of my
01:29:21
innovation team so I came up with the
01:29:23
new product. And that's my word to make
01:29:26
new problem. But to sell it out to go
01:29:29
to the customer and get money already
01:29:30
we is that my responsibility as
01:29:32
innovation leader it would probably
01:29:36
depend on what comes you work for what
01:29:38
they're what they're "'em" model is I
01:29:40
mean I would I would hope that the
01:29:43
researchers not involved in that
01:29:45
because there's probably you know that
01:29:46
their skills that is probably better
01:29:48
better used in a in other areas but I
01:29:50
can imagine if you're a smaller company
01:29:52
then you're probably doing almost
01:29:53
everything every aspect of it to to get
01:29:56
your technology out there okay but also
01:29:59
back to my point about you impact of
01:30:01
this. So how how do you know your your
01:30:03
sister innovation that has that are
01:30:06
that would matter so what what is so if
01:30:11
you say this is non down with something
01:30:13
no new and novel and something else job
01:30:16
to like to transfer that yeah then out
01:30:18
and like we do you don't know whether
01:30:20
that's actually a good you mention a
01:30:21
not so that is the impact part it is
01:30:26
hard sometimes takes a lot more time
01:30:28
than the coming up with the initial
01:30:30
ideas or I so I can see companies to
01:30:33
all the require you to go for the last
01:30:36
mile. And that evening okay. Thank you.
01:30:40
right now we've got ten minutes over if
01:30:43
that's really great again thanks for
01:30:45
coming and participating this for you
01:30:48
guys are really great audience so think

Share this talk: 


Recommended talks

SIGCHI Lifetime Research Award Talk
James Hollan
April 21, 2015 · 2:35 p.m.
234 views
Multiple ways of building a recommender system with ElasticSearch
Andrii Vozniuk, React-EPFL
May 11, 2017 · 1:04 p.m.
1537 views