re
okay like to welcome everybody see this
a special session a natural language generation for dialogue systems
i'm just gonna give a five minute overview of the session then we'll have a
couple of long talks accomplish very short talks and then we'll have a panel
and
about a lot of this kind of came together to organize this so the other
organisers of their temper like a kind reader david how corrupt
showing or read and very in a research
so the
what the organizers task is to do is to say like why we would have
a special
session on a nlg for dialogue systems
and you know because you might think well this
areas been around for a long time indeed some of the earliest work
on computational models for natural language generation was work done in the context of a
dialogue system or question answering system like kathy make humans
early work for phil collins early work
there's also a lot of earlier much earlier work you note going back almost twenty
years
people doing starting to do statistical
work on statistical natural language
generation
starting with the kind of seminal work of like alien like who showed that you
could kind of have a very loosely configured hybrid
linguistic statistical representation where you could overgenerate and then you could learn
rightly rules and filter out
this filter out the data's you could produce and so that works as old as
nineteen ninety eight so you might
still say like why would you have the special session now and also in the
context of darpa communicator conversational dialogue systems
i edited a special issue with computers
speech and language are not sure language generation for spoken dialogue
and the number of people that you know had papers in the special issue including
am and i lasso and ridge alex rudnicky mari ostendorf stephanie seneff so a lot
of long time people been working on conversational dialogue systems
for years
but the reason that we wanted to have this special session despite the fact that
these a lot of that
its of generation for dialogue have been around a long time is that there's been
a recent kind of resurgence of interest in a natural language generation because of kind
of the
ai renaissance i guess we should say all the interest in chat bots and all
the consumer products that are out there now like collects and google system
the other thing is that the availability of large online corpora like open subtitles or
twitter or
i you know corpora like that have led have may people wonder whether they could
actually use a purely
statistical kind of and machine translation approach to produce dialogue turns in an open-domain way
so there's been a lot of activity in that area for the last five years
and
so
so there's a lot you know seems to be a lot of different stuff going
on of this field and one of the reasons why i wanted to organize the
special session was the kind of
be able to look at what we can do now with different generation techniques into
bring you know
it especially in the panel to bring in a panel of experts people who worked
on one language generation for dialogue systems and try to get some different perspectives
on from
from their from their points of view what kinds of techniques which ones don't work
which things the ready for primetime it could go into consumer products in which things
are still just
kind of basic
research ideas
and when am i am particular interest and i think that of many of the
other people to organize the panel who have put out these other challenges but e
two e challenge is also the web nlg challenge now is in interest and stylistic
variation in some of the classic things the natural language generation sabine able to do
in their sentence planner and like and a kind of interest in whether
and to and framework is actually without a lot of extra architectural details that kind
of model the traditional natural language
architecture whether they're actually gonna be able to produce different kinds of
stylistic variation like the
previous generation of statistical language generations could do
so that's kind of why were here
and
we have
too long papers
redundancy localisation for the conversation of unstructured responses and that a neural language generation in
dialogue paper
we have to short papers that will be presented in five minutes the for the
panel
and they'll be in the poster session later i wanted to have them in
in the discussion in our in our minds before we start doing the panel
because i think
there's a really interesting thing here of a new generation challenge that aims to be
a little bit more complicated than what people then you seen in the neural generation
framework
and then somebody was really
hot of the starting box
and that
the corpus was released in two weeks later they had a they this
character the character out of the box model that uses the corpus of this is
all kind of very much
breaking news
and
so we can go ahead and
get started for the for the main papers and then we'll have the panel at
the
at the end
okay so sebastian