Dutch language

How to pronounce services in Dutch?

Full Transcript

so I'm Zoey borovsky I'm the librarian
for digital research at UCLA library and
I'm one of the sig convenors happily
here with my colleagues so this session
is designing fruitful we have six papers
and ninety minutes so I think that
calculates to about 12 minutes of paper
and Glenn said he can time us our first
presenter or our first presentation is
called the case for a citation index by
by the humanities and for the humanities
and we have three people but one
presenter Giovanni calivita Silvio
Peroni Matteo Romanello is going to
present is that correct okay it's your
money okay so I guess I will turn it
over do you want the microphone so you
can be streamed yeah okay and I guess
it's best on your picture yes yep no
thank you
good ok good morning
wooden is okay turn around yeah works
okay good morning everybody
so what I'll try to do now is basically
make a case and we will see if you are
convinced by saying essentially the
point is to make a case for a Citation
Index by the humanities and for the
humanities I'll try to explain what I
mean by that so why that might be
relevant and of interest and I also try
to briefly explain how we might go about
building that in terms of research
infrastructure and then what is the role
quite crucial role that libraries should
play into that if if they like starting
from a little bit of motivation it is
well known at least in the in the in the
community of Bible metrics community
that the humanities the arts and
humanities actually are not very well
captured not at all very well captured
in commercial citation indexes and by
the way commercial citation indexes
cover all citation indexes here for
example you have the proportion of known
academic journals that are indexed in
either Scopus or web of science on the
right instead you have the number of
contributions for every individual the
average number of contribution for every
individual researchers in in different
fields and you have the humanities here
quite at the bottom just giving you
essentially an intuition that the
humanities not are not well covered in
these indexes and also we struggle to
understand what is actually there
what is the coverage of what is there
hence basically these services are not
easy to use for Humanities color this
kind of fragmentation is a little bit
also familiar increasingly less so but
familiar to humanities scholars I would
say that are used to inhabit different
informational ecosystems so they need to
go and use different services to get the
information that they need for their
research to give an example that is more
personal I'm partially trained as a
historian and a dream that I had when
studying and working was to
see library catalogs connecting with
archival information systems so the
basic question is who has actually
worked on the documents that I'm looking
at so finding that information is very
difficult
this error-prone and it requires
basically a lot of domain knowledge and
a lot of time but this is somewhat
generalized so we have maybe digital
libraries that are replicating impart
metadata and data that is also contained
in library catalogs and digital
libraries we have Glam information
system that are struggling to connect
with what is on this side we have meta
engines such as a rogue Rana that gladly
are providing and trying to provide this
integration but I also struggling on
doing so and then we have commercial
solutions such as Google Books and
Google Scholar that you know we use but
we don't really know what is the quality
what criteria what is the culture and
basically what I'm arguing is that
citation data can provide a list for one
possible way to start pulling this
together to interlink all these
different services a little bit more
than what they are now so basically what
are the main motivations to build a
citation index for the arts and
humanities so first of all information
retrieval it's not as much as changing
the quality of information material by
the speed like finding all the of the
scholars that I've worked and then have
cited a certain book or a certain
article or a certain primary source or
certain work of art takes a long time if
we can reduce that to basically a click
that would be great secondly as I
mentioned before interlinking
information ecosystem that we need to
use together but they are at the moment
apart such as archive archives and
libraries but also museums and what else
thirdly by virtue of having a service
that basically you want to buy in
improving the metadata and also
reference behavior practices of all the
stakeholders of the system basically I'm
talking about potentially libraries but
also publishers and scholars themselves
our scholars when we refer to something
when we cast out the reference maybe we
don't all use the same standard maybe we
we take a somewhat eclectic view and
that is a pain in terms of doing
citation mining but if you want to have
your article indexed in the Citation
Index then you might pay a little bit
more attention and the same applies to
publishers for example but you know in
the Bible Metis community we're really
trying to to involve in to putting all
the metadata as open metadata and also
as machine readable and standardized
metadata and lastly a little bit so
certainly it would provide for quite
wonderful data to study basically what
is the social and intellectual
organization of the arts and humanities
through citation networks or
collaboration networks hence again you
know the point would be that it pulls
together it could help pulling together
all these information systems but I
would also say that we need to move
beyond what is available at the moment
take Google Scholar for example Google
Scholar I'm taking Google Scholar
because alledgely is the most
comprehensive even for the arts and
humanities and I'm assuming everybody's
is aware what it is but I say alledgely
because you know we cannot really study
what it contains
we cannot query it programmatically
there is no API in Google Scholar there
are possibly a lot of errors exactly
because they maximize recall so they try
to find everything what we think is that
we should try to have quantity and
quality at the same time so it's crucial
to have a comprehensive source coverage
especially in the arts and humanities
people really want to have everything if
not they may feel the reason issue and
they're quite right and source includes
basically all the materials that size
something hence - I have conferences of
sauce coverage you really need to have
substantial chronological depth we
cannot just cover the last 10 or 20
years which is what you can find in web
of science as coppice we need to go back
to basically when we started to use
references an interesting story but not
for now thirdly we want to be reaching
context what I mean is that
referencing the citation semantics in
the answer humanity is quite a little
bit more complex than what we have in
the sciences so what it means to cite
something can take a variety of meanings
it can be of course saying that there is
an intellectual bed but it can also be
criticism
it can also be contact relation
perfunctory citations and what else so
it is it's important to provide some
context say a snippet or a citation
context that allows the user to
understand why a citation was cast and
lastly we should try to be collection
driven what I mean here is that we
should leverage specific specialized
collections that we have in research
libraries in terms of secondary
literature in order to be able to divide
and conquer this problem and I'll tell a
little bit more in a second so the idea
that we have going in terms of potential
solutions is as follows
so we would like to take inspiration
from libraries and basically library
catalogs so how they are built they are
collaborative they are distributed and
they are open so there is no central
institution you know does the
cataloguing maybe there is a central
institution that does that amazing but a
cataloging is done distributed across
many libraries and then of course it is
it is open because it is done by the
public as a public service hence let's
try to do the same thing so let's try to
have a set of institutions say libraries
archives and where else that take
ownership of a specific part of the
scholarly literature say here we might
do the history of the Netherlands and
then this institution basically takes
responsibility to digitize and to index
in terms of citations that specific
literature so if there is a new issue of
a journal that is specialized on that
they take care to put it into the system
and index it how using a piece of
infrastructure which we call the scholar
library which is basically digital
library that exposes all the machine
learning components that are necessary
in order to make this process much
quicker than by hand because not it
would be unfeasible then second point
we take all the cetacean data and we
federated into a single linked data
repository which already exists it's
called open citations and thirdly then
once that data is there we can use it
for whatever we like
including a Citation Index which is
another application that a user can can
basically use so if you think about it
this is really a triple if' for those
who are familiar with that
so for either the standard to share
image is just done for citation rate and
this is the same idea just different
perspective with many instances of this
digital library that allow librarians
say to create citation data then we
aggregated into a unique database open
citations and then from there or
visitation exposes the citation data to
machines especially now from there we
create services including a citation
index so we have built prototypes for
all of this
they're already online for the digital
library in the Citation Index in the
context of the vanish time machine which
is a project that is within the umbrella
of the time-machine European consortium
I can tell you more afterwards which is
really just a digital library but embeds
all the tooling that I was mentioning so
once you put an image there you get a
text but you also get parsing results on
all the foods finding all the references
possibly with a few errors because it's
complicated and then you can go about
deciding if it's good or not and there
is also a link to a permanent identifier
if we have access to that taken from
libraries or archives or where else now
this is online but it's
password-protected free source of
copyright as you can imagine open
citations instead is is a service that
is already there since a few years you
have links here and a recent publication
describing it it's basically a triple
store for citation data and a lot of the
work has been done in terms of devising
a model that can represent this data and
then expose it
and lastly Citation Index again in the
context of the Vanus time machine this
is online so you can go and play with it
if you like
it's a Citation Index with so
think like Google Scholar just a
slightly improved interface and also a
few things that are the liking of
historians for example we have
connections with the archives in Venice
so you can see even a scholar or given a
publication which documents they have
being cited and vice versa
taking a document you can see who has
sight on it which is quite amazing Alex
for me it was essentially my dream and
now to sort of wrap it up basically what
is the role of research libraries I
think it's quite massive as you as it
might have guessed and especially it
comes into play with this first layer so
it's really the layer of curating
citation data to then pull it together
into into a federated database so but
before that there is something that
really we struggle with still and we I
really would encourage you to to
consider it and I'm really sure that you
are already working on it but to build
this tool we really need to have
metadata that is uniform and then we can
query to an API because if not we will
end up creating internal identifiers and
we won't be able to use your metadata
and to point to your identifiers when we
have material that we have indexed this
is really not what we want we would like
to basically to use your metadata as the
authoritative metadata and as we just
provide the links through citation so
this is crucial piece of infrastructure
that we need then digitization and
creation of citation data and lastly
copyright issues which are really
painful and the idea of allowing
libraries to keep their collections in
place and just federated citation data
comes also from that experience thank
you very much
[Applause]
in my library yeah no I mean in in our
view this is something that we point to
so basically if you also go to that to
what we point to is the specific webpage
of the archive say the archive of Venice
or another archive where there is
metadata and there is a link allegedly
or hopefully to the finding aids related
to that specific fund or or record group
so that the user can use it so we
replicate nothing we simply link to it
and it would be ideal of course to find
out online but the struggle is really
the permanent identifier at least for us
right we need a permanent identify to
that page so that we can link to that
that's and then it's the business for
the ARCA I would say please
I'm wondering what the approaches in
terms of it seems like the central model
is the idea that you're citing a
particular object or particular
documents so sort of an object document
based model and what to do with when
you're encountering citations that are
broadly gesturing towards someone's work
yeah implicit citations for example say
yeah so yeah that that's a good point so
we have that a little bit for archives
I'd say because the archive the
granularity with respect to the specific
citation can definitely go down to the
document in terms of extracting that
information and making that uniform from
a footnote that's really painful so what
we try to do is to go down to the extent
possible keeping at certain quality so
oftentimes we have citation to the
record group or down below but we don't
go to the document we just provide for
the text of the reference that might be
perhaps a solution that we can use for
authors for example if there is a
generic citation to an altar we could
add it as a without pointing to a
specific publication and then counting
it for the other but just counting it
for the author then I would say we would
need to make it very clear in the
interface that that is the case so we
haven't yet done that once yeah it's
quite a good point thank your ass I
think we're going to switch to a
broadcast this is going to be digitizing
for impact from pixels to scholarship it
was a work by three people Matt Knutson
Sarah Rubin now and Eric shows from the
New York Public Library and lotos going
to
and I'm here with Matt Knudsen
conservative to discuss digitizing for
impact from faithful to scholarship the
New York Public Library is the third
largest public library in the world with
what is that number over fifty three
billion items these collections
represent a broad spectrum of material
types ranging from maps to prints photos
periodicals books audio recordings
filming off video objects and more
digitizing everything in the library
would be next to impossible and
impractical for a myriad of reasons so
selection criteria of prioritization
strategy and corpus development have
become increasingly important as in my
pls digital collections have grown over
the last decade or so there are several
different streams of digitization at the
library driven by different types of
demand but the bulk of still image
digitization is curatorial Griffith
through a committee selection process
over the past few years the projects at
the top of that prioritization queue
have been identified by staff is the
most heavily used printed or manuscript
collections to relieve service staff
time wear and tear on commonly requested
items and to provide broader access the
challenge that remains for these heavily
used and digitized collections is that
once digitized a collection item becomes
a very different object in the digital
realm with a vast increase in the
complexity of expectations and
possibilities for conversion of data and
interconnectivity with other collections
one such heavily used collection the
directories of New York City has
traditionally been served as hardcopy
books here in the Milstein division
which supports research in local history
and genealogy a precursor to telephone
books the city directories attempted to
record every working person living in
New York City by first and last name
street address and profession the
earliest was published in 1786 and lists
about 900 people or about 3% of New
York's population while in 1886 there
are more than 300,000 individuals listed
which was a quarter of the city
directories have long been used by
geologists
searching for ancestors that are also
used by researchers looking to study
social mobility industrial activities
and other
directories were structured
alphabetically by surname so if you're
searching for a person you can locate
where they lived and what they did for
work interrogating the data by any other
means such as street address or
occupation however it would be if they
didn't leave time-consuming so for an
economist studying bakeries for example
in the 1870s and 80s that work would
mean leaving through about thirty-nine
thousand pages or about six million
entries one publisher though John dajin
reorganized the data to create this New
York City Street director published only
in 1851 in parallel to the city
directory it's structured alphabetically
first by street name then number which
raises the question wouldn't be great if
we could organize all of the city
directories in this way or in any way
that we want digitization of these
images mining the data and divining
their structure would give researchers
the opportunity to interrogate the data
set from multiple perspectives how are
they doing since they first started
appearing on digital collections the new
york city directories have consistently
been in the collection most frequently
accessed by users 136 volumes are now
unlocked which is about one hundred and
fifty-seven thousand pages this virtual
library however has some limitations
did you know the station produces high
resolution images made up of pixels and
image of the city directory however
isn't about the image but about the
semantic information it contains
baseline expectations for
discoverability in digitized collections
have shifted in recent years most of our
users now expect full-text search an
increasing number of users report the
lack of full-text searches above so the
full potential research on packaged
digitized directories remains unrealized
for the current display while providing
digital surrogates online creates an
averaging
the researchers to world optical
character recognition OCR given vastly
improved search and collections initial
work has been done to extract text from
the directory walls use of OCR and that
data is currently stored here you can
see a page of listings on the left and
it's OCR output rate inconsistent letter
spacing and formatting results an
imperfect output the OCR process which
requires mobilizing and perhaps more
importantly the content professional
address needs to become structured to
give these two researchers the
inconsistent data structure in the
original volumes compounded by
inconsistently messy OCR eliminates the
effectiveness of regular expressions to
parse data so as part of a year-long
collaboration and when PR worked to
address the problem difficulty the
person with the goal of extracting
structured OCR after generating the OCR
the supervised machine learning
algorithm signed and returned labels for
each element of each entry
my entry into this work came during my
14 years in public libraries map
division where I advocated for a
digitization of our New York City Fire
Insurance map collection this collection
was heavily used because it details the
environment building levels starting
from about 1850 to they enable us to see
how the city transformed across time
which is valuable to urban
archaeologists and historians to
genealogists sociologists economists and
many others this demand led to about a
dozen years of digitization of around
13,000 Atlas sheets like this one
digitization of directories and maps
offered an even more exciting prospect
that if street addresses are the primary
key in these two data sets textual on
the left and graphical on the right then
we can join the - but there's no OCR for
maps so about ten years ago we begin
developing tools to first-year reference
images then used computer vision
algorithms to mind building footprints
then developed crowdsourcing tools to
check the accuracy of our data and
finally develop consensus algorithms to
establish data competence measures
connecting multiple sources like this
takes advantage of the network effect
increasing the value of the whole beyond
that its parts and provides
opportunities for researchers that have
never existed before this led us to
another question what if the data from
the city directors could be organized
geographically well this slide shows you
what happens here you can see the
personal names addresses and occupations
are assigned to specific buildings
aspect effectively collating the
directory entries which were previously
scattered across pages into a single
building this is the kind of navigable
and searchable information that our
researchers are really looking for
ongoing research projects using
directives include two one uses a
localized extract of East Village
addresses to ensure to encourage student
data gathering projects focused on 19th
century Irish communities the second is
a capstone project by obtaining students
at the Center for urban science and
progress they build upon the OCR machine
learning algorithms and the capstone
project is providing foundational work
for future scholarship as heavily used
in access collections begin to share
space online and tools are built to
enhance the experience of drawing
connections between them the gravity
begins to develop around a particular
fear of geographies and eras the
strength of NYP ELLs collections by
nature concentrated on topics relevant
to the history of New York City that
over overlap geographically and
temporally as a base layer of map and
directory data's that develop through
earlier projects it's become easier to
prioritize new digitization by looking
at space and time adjacent collections
that will lead to a multiplying effect
and richer corpus of material for
researchers to draw taking cues for
digitization from you
partners the local academic community
the researchers across the globe
interested in accessing and working with
our material they've begun to develop a
digitization strategy around entire
corpuses of material rather than one
offs to increase the value of these
collections and unlock the potential for
a high degree of interconnectivity one
project currently being digitized to
support Connexions already drawn between
the maps and directories is the
committee of 15 records the committee of
15 was established in 1980 the prominent
New Yorkers concerned about the dramatic
increase in place particularly gambling
and the social media evil of
prostitution in New York City and the
unwillingness of the police to suppress
it the committee scoured the city to
document or report their findings the
finding aided metadata for this
collection are arranged by street
address making it clear and simple to
connect with existing data one
interesting chapter of the committee of
15 stories related to Chinatown in the
nap as they commenced that they were
unwittingly drawn into a turf war
between two gangs or control over local
gambling prostitution and opium houses
the committee received highly detailed
unsolicited information about many
illegal activities from two members of
the Chinese YMCA who were in fact
working for an upstart gang leader vying
for the business of his rival queen -
expect was targeted right here the
committee work catalyzed the exact
opposite reaction to their progressive
intent this story is enriched by other
documents related to the same place and
era these two blocks are represented by
an incredible density of documentary
evidence so much more than can be shown
here they highlight threads woven into
numerous and wildly complex and
overlapping narratives the maps connect
to the affadavit a lottery ticket
business cards mortgage applications of
the far-left which was the home base of
popular slum tours of fake brothels
opium dens and authentic local
restaurants as you can imagine
developing tools to transform digitized
collections into machine readable and
interconnected data is highly resource
intensive which is what is key for
us to partner with research institutions
that have a strong digital humanities
and computer science infrastructure with
our robust digitization services and
world-class collections we're uniquely
positioned to advance scholarship
through building relationships with
strategic partners essentially the
library can serve as a bridge between
our collections of the research
communities who rely on them and help
facilitate their progress among digital
surrogates to corporate the actionable
additional data and the fall of 2019
we'll have a new space to convene and
facilitate partners in conversations
practica and symposia to explore
possible deeper connections with
academic partners can provide a forum
for dissemination sharing and discovery
of the process and products of
research-based historical collections
our decision efforts focus on developing
a strong corpus of collections from a
variety of formats to enable greater use
by scholars and researchers leveraging
our relationships and partnering with
academic institutions and the broader
research community amplifies the
potential impact of library collections
thank you I guess we'll have to propose
some questions online and maybe we'll
get some some responses later I'm going
to call Zack Coble up yay Zack is here
and it the title of his paper is next
generation of scalable and sustainable
d-h services
I know in pocket that's good okay and
can I set it does this pick up the audio
if I put it here yeah right you weird
there master I don't think so
sure like that okay all right thank you
all very much and so we'll jump on the
express train from New York Public
Library at 42nd Street and head down to
two boats library at 4th Street stain
and in New York City so I'm Zack Coble
I'm head of digital scholarship services
at New York University
so digital scholarship services is a
small department that's based within the
library at NYU
we support D H digital publishing things
in in that area and we're relatively
small departments so there's myself a
digital scholarship librarian and very
soon next month a digital humanities
technologist the department was
co-founded in 2013 by Jennifer Venna pal
and Monica McCormack and the two of them
had written an article around that same
time about scalable and sustainable
digital scholarship services which kind
of staked out kind of our our approach
towards supporting D H situated within
the library so kind of thinking about in
a field that's still evolving and
and rapidly expanding how to be able to
to do things right to to give
appreciation to to the craft of D H so
that's something that I've inherited and
and spent a lot of time thinking about
and in terms of how we organize and
structure things at the library to be
able to to be sustainable scalable but
also have room for innovation to not do
these things at the sake of trying to do
something new so rather than scalable
and sustainable and innovation at
opposite ends of the spectrum of perhaps
something of a three-legged chair and
how those those can balance with each
other part of our service philosophy is
to provide the technical infrastructure
and training for four scholars which
usually ends up being faculty and
graduate students to do their work so we
don't embed ourselves in people's
projects we don't do digital humanities
projects for other people but we want to
to give them the tools and the training
that they need to be able to take
ownership and be comfortable with doing
their project so we do a lot we have a
robust workshops and training especially
in collaboration with other departments
I'll talk a little bit more about that
we do the carpenter ease consultations a
whole suite of services with some other
departments in the library especially
our data services department which
offers the the GIS services quantitative
qualitative support we also have digital
publishing scholarly communication
metadata everything that you would
expect so one other thing that the
article that I mentioned staked out was
this kind of so it wasn't the this is
the faculty website problem so the
problem of faculty who want to create a
website of some sort it could be a no
mecca exhibit it could be something a
bit more
customized but but how do you provide
hosting for that so which has been a
perennial challenge especially in our
department and so one of the things that
we found that I've been working on is
it's what's called the web hosting
service we've kept it very vague and
we've been offering this for for about
two and a half years now and it's the
same as a commercial web hosting service
like dream host Bluehost GoDaddy that
sort of thing but it's offered by the
library so so we underwrite the the cost
of hosting we use a vendor called
reclaim hosting some of you might have
used them before they're amazing the the
price is very fair the technical support
is amazing and it allows us to
effectively solve the the faculty
website problem so it meets our our
service philosophy again we don't build
web sites for people we will give them
accounts where they have the cPanel they
can do one-click installations of the
the content management system of their
choice they can build something more
custom specialized there's no js' you
can build Python apps with flask and
that sort of thing so it kind of fits
into this intermediate web development
need we also don't provide technical
support I kind of started as an
experimental thing of again being a
small department we just don't have the
time to do this so we we vet users ahead
of time and we told them upfront that we
only provide very minimal technical
support and it's been very successful I
thought we would get a lot of pushback
but but it's been good we're able to do
this because our IT department offers a
wordpress service which has a lot more
technical support for comparison they
have about 10,000 sites in there we have
about 400 sites and in our web hosting
service so you know like one of the
things that that we've been thinking
about is you know what is the right
scale for projects
so something that you know with this is
you know technically available to all
60,000 students and however many faculty
at the university but we only have about
400 and we can support it in a
department of two people and that meets
the need that we have on our campus so
I've given a kind of very very focused
view of some of the things in digital
scholarship services but I want to zoom
out a little bit to talk about some of
our partnerships across the the library
and also across the university as well
on how we're able to provide some some
services around that because as we know
most most good d-h work is collaborative
so I want to talk a little bit about our
partnership with our central IT
department which is very very strong
it's it's a very good relationship I
know at least with a lot of universities
in the US there's there's been kind of
varying degrees of merges and unmerge
--is of library and IT departments and
nyu we've been fortunate enough to have
a continue to have a strong working
relationship with with our IT department
in particular we work a lot with the
research technology group within the IT
and so they do support for a lot of
different things HPC collaborate on on
some of our departments so we we have
several joint departments that are
managed and run by both libraries and IT
the the data services that I mentioned
we have our digital library group which
does digitization and kind of research
and development for for digital
publishing and scholarly communication
projects and soon to be my department
digital scholarship will be a joint
library and an IT department
so the organizational charts can be very
complicated it's not even worth going
into that but the partnerships are there
and so what we try to do is you know for
our patrons for our users to make that
as seamless as possible don't expect
them to you know understand you know the
organizational chart how things are
structured but just know that you can
ask pretty much anybody and and will be
able to you know get you connected with
the right people you know very good at
doing handoffs and and bringing a lot of
people into the picture we have we've
we've started in the the past year a
digital humanities steering committee
which is libraries and IT so again kind
of like a clearinghouse for conversation
talking about upcoming projects and and
how to make sure to get everybody on the
same page which is always a challenge at
a large institution so we did a couple
summers ago a peer benchmarking series
of reports again as a joint libraries IT
thing which is basically just gathering
data to ask for new positions in several
areas digital humanities one of those
and also in data repositories so that's
what I want to talk a little bit about
here and and they were successful and
and getting some staff and thankfully a
lot of these you know somebody that's
hired in data repositories you know
turns out to you know have a lot of
impact and usefulness within our
collections as data work so within data
repositories we've been undergoing this
this large project under the umbrella of
digital repository services for research
DRS our or dr. senior which is kind of
you know as more and more scholars are
you know getting these increasingly
sophisticated needs for for using their
day
you know what if we just kind of like
take everything apart and start to think
about what kind of infrastructure we
need we have buckets we have our
institutional repository Google Drive
box that sort of stuff but scholars need
more than that so they they divided it
into these four groups basically and I'm
going to focus mainly on the the second
one which ended up creating a service
called research workspace so just for a
bit of context the the other of the
other four areas that they were talking
about one is for basically temporary
high speed storage so that's basically
the the data storage for
high-performance computing
there's also digital publication and
preservation which is the stage we're in
now I'll leave it at that and the the
fourth one is basically secure data
environment so for people who are
working with sensitive data issues of
patient privacy that sort of thing how
to get a very lockdown storage
environment but the the research
workspace the one I'll focus on today so
some of the things where we were trying
to accomplish with it is is mountable
storage so this is on on my laptop I've
mounted one of our our collections here
so it's just if you're on a Mac command
K connect a server and then you can it
pulls up the the data set as if it were
on your computer so we wanted mountable
storage shared access among different
groups so granular user permissions and
it's intended for for working data not
intended to be long-term storage for
archive data and we found that it ended
up being very useful infrastructure for
our collections as data work NYU we call
it text as data you can kind of see
where our
emphasis lies and so some of the work
that we've been done have been doing in
this area we have a lot of our subject
specialists are curating collections
that kind of started with one person who
was working with some South Asian NGO
documents that would kind of appear and
disappear in very short time windows the
previous presentation talked a lot about
some of the work partnerships that we've
done on the street directories a couple
of my colleagues give a shout out to Ted
Nick Wolfe and Steven Balog who were the
ones working with NYPL they were doing
the the OCR analysis for for the street
directories but within that getting more
and more groups asking for larger data
sets especially with the Center for
urban science and progress which was
mentioned in the previous presentation
we also have a Center for data science
data Sciences I think in many places one
of those just kind of like boom areas in
the university and NYU is no exception
so a lot of people who want just kind of
large corpora for a machine learning AI
analysis sometimes they're they're not
that picky they just want a huge
corporate adjust feed to the Machine and
and sometimes they have specific needs
so so one of those that I'll talk about
is a professor in the history department
and the the Center for data science
who's working in a group called the
history lab which is a joint initiative
with with Columbia University and so
what they wanted was the West
Congressional Record database and all
that have use of the historical
newspapers data set from ProQuest so
after extended negotiation with with
ProQuest we were eventually able to get
the data sets and the ability to share
them with the NYU community which gave
us the question of where do we put this
data so this was at the same time that
the
research workspace infrastructure was
being created initially we put it on our
high-performance computing environment
as a data set that you could load in
there the problem being that it's you
know you have one have to know that it
exists then two it only exists within
the the HPC environment so not everybody
wants to do that so we were able to
within our research work space
environment so something that anybody
can can call up and and be able to work
on locally so the nice thing about HPC
is that you know it has that kind of
like scale of infrastructure but if you
want to be able to work locally with
particular tools that you've developed
for analysis it provides that that
option as well so that's it right ok I
will I will wrap it up there then thank
you all very much
[Applause]
[Music]
yes so within the funding one of the
things that is important for us is to
have hard lines of funding so kind of
that's at the cost of having like a
smaller Department that we're not
chasing grant funding to have hired and
in temporary positions that is
unfortunate or can be for for the people
who fill those positions so kind of
taking the longer route - we do a lot of
work with gathering data so we can make
the case and kind of pull that up the
ladder at our institution to convince
them of the use for it but it's
something with within my area that the
funding is kind of built into the budget
at the library thank you so next up we
have the title of his paper
yeah first of all thank you so way for
this idea to translate the slides into
English so if you like to follow my talk
in English by reading the English slides
it's now possible but I will talk in
German and I will also show the German
and slides and now I hope we'll get
along with that as a man for African
focus of Fr Granada Kane is a league DD
central periodic foolish in between
organizations it adds forums people tick
akiza melt hat is measured by a specter
in meinem for track
neopor handle its antibiotics which is
that 5000 acting as forums partner an
active fewer and washrooms play actor
University series resistor Johann Kaspar
lava Tapley Foust carbon I want to make
the time Bishop III playing organ is a
tuition and financial ALPHAN that the
people you take had only measure of
Tygra a lesson learned via patent a
student project or by lesson learned
evening and few entertain in agony
valley goin on cat an individual who out
in the about witness project givers and
mussenden which teachest wooden quotes
also some a project this is for songs
corporations polar project ba weather is
on the plot project it is a belief edits
on DBQ further from SWAT salvation
eternal form it and comic millions
rights have found in the Aston project
father thus tedious project is an
historic literature online edits you
insane he briefed this teacher of clear
us you on Casper Lovato SNF is -
violation not so now for the National
Forum organizers own institution as a
fact life permit a torch in DF game on
Ralph cabinets and her bibliothèque is
this ultra Ebert van 2000 preferred DVF
on Yonkers pallava to harm to digitally
Xion does Vietnam ended an affair
910,000 image design event VA deny
insulin brief inspiration Demeter -
listen papyrus here guns guns Buzzell
Keith Emerson from junior as with an
absent and finger or on Tatum Ashley is
an old Angeles and theater Stan even
ended at Bank edits you'll and export
yet known via alpha in Manhattan via in
seek his personals as word ina
50% Turkish often petard together as
national force these 50 potential of
water in that sent her periodic anger
wasted that hides an outfit salt and
this is Dell Tuesday's leash split CD
bibliothèque about no and true shows
from F thousand fronting that sue while
pursue an ultimate arbiter in an
institution noon abated Suzette's leash
suit ISA employee Stella harm via novi
travnik finish 10% Josephine Ines odds
in the hunt - a balloon and masticators
only Ashley song done about Amina
bastante heart in the restorations up
tile on all mm digital easy important
digital is young sent home even a
descendant of digital to become incident
also Martin Dustin personal cost in
France vow at film fans 80,000 Franklin
and poem digits entropy pratik offended
damnit is a banished Gatan rehab Zack
Austin as a physic metal name diao stain
as a V mission Philip leave a missing
baby handle and Tridentine first open as
I'm innocent in the Documentum mm shoots
Anant West are earrin we're missing
about ow
ana technician avoid home in one
software her perfume uber hopped a stint
at and oust house so a million we have W
by he now saw by Tau Alpha Newman as a
theaters hosting for special fear the
infrastruc ture cons alcaman via bite
plates etc the highest we have most in
schism of here vida and cause now fund
and in the Elgin our fund even from nap
oxy thousand nap 90,000 tried Safran
erection does our Suzanne and fewer does
Lovato believe Eddie Jones project and
come be so made of Causton
Fondren let's ban 50,000 fans indeed
it's and how people take Jewish offended
it seems wise even few camp cause if it
is despoja Cuba Anna million my it's a
fountain on DC million writer fountain
pathology and people who take house
their Agnon chasm a mineral Moffat
liking us on our funky badass team be
tracked in via from the forums Institute
soon from for Oxford Arabic omen from s
and F M Danzig man does karma film
simple
sent on Zahra cost and which is a Forex
further along in the a protectant rehab
first 900,000 rights of huntin DV invest
en or Madhavi Aiden I can click on
terminus on indirect for Oxford rumba
tribe and then only Disney our fund
currently does project this business
after she project and furnish exists Ian
and that's English right and it was um
green stickers it was also a perspectiva
and then the same tibideaux item the
inhd bibliothèque hat in this
enforcement sparked a backlash Vulcan
and ISA cost in each permits an alpha by
and Harvard DC discussion Miss Manners
rights during infinite vastness vision
and on land and house it impossible
mission interface in does out to her on
no strategy Frogger 900,000 found in the
Hansen name and also gave me up vomit
pandas are CLE via Mohandas our son
conspiracy vision conned by Weber toy
comes in test is a four shows corporate
zone the anaphor few mundos people
decoction of drugs is vanish pits
efficient and context noon Susannah guns
concreted in Sly Stone few Anakin's
concrete Khulna frog you can totally see
and understand Suzette Lee s listen we
can in present even online in the Edit
shown above catalog via
Conan understand a Hutton Yamato hominem
divot was shown via console catalog and
Russian owned we currently are on the
metadata person by the in the business
afternoons reciter happen ultimate owns
this project tuition amend of Utah
standard soon as we want an agonist and
of language in context present Ian Cohen
when I even denial of the Edison's
platform shinin theater in vitamin D no
to answer is correct
after sue and in internal workflow and
India paradigm to opt in young and did
about an actor to technician aversion
and Sue and week on DVD on fire under
project not sunken or Olivia our unknown
and partner Danny back even curl and
you can't need some biography and
fighters forums project and he's and
Vikram in lava topic Newton whom even in
Darden transfer to a mutation as fats it
offseason here no quote Suzanne go fast
liebelei guman some Lutz Nana for
schmooze corporate zone at fat it keeps
atomic dust is each packet is agonist
Anders and a huge Viet am on data which
to the agonist and gruza inclusive a
neutral effect attractive are assigned
owned and flighty mediaeval Danzig avoid
for for gap reacted lack succeed order
in handle for inspect the even dilution
effect wooden this apportion scale a
demon becomes Val Venis Abbas deemed on
dock in Mahindra to and that man vitae
which tender abortion can ordered some
innocence of an entire boost and
fatigued Ashley's in can commune other
student lessoned learned the fortunes
project can be as people think niche in
daily business African valium did at man
flew to close is fortunes practice
interim up reacted by rice Ibaka a
friend up chemical Bernie Sanders
highest Manhattan chief napped allows
despotism noctilum Dan Parkman deeper
stanza Halton owned man hottest ticket
Alice Young sent home on Ferguson
Ibaka - Dan connects and Brockman
an episode of a platinum owned
coordinates you own this hypothesis and
it's and her people you take Jewish
that's wench Lawson
five thousand after now Anna specific
Estella answer rationed in product so
does product sales management 3d digital
easy on does also IANA that feeds the
coordinates you and attracted to change
is lights idea about Santosh partner and
fewer divorces proactive Dean's all
owned two-thirds laziness an avid
Hrothgar become mushy plan and
coordinate zone from eigen digitally
songs for happen commander of like no so
what
vámanos abortions corporate zone else
below takashi-kun of track first aid and
unseat consult Amanda's ugly Oh fine and
Cl
so hunt happened as highest as the d4
expect nice to be free state project
metal M fine and seein Sun and even meet
equally and with heaps Milton fine and
seein the sisterly Mohan guns class -
give Islam as an anti vision shaft and
and abortion and dustman ads for Leslie
her partner's ish outran will dismiss
for Leslie her partner Kelton can there
even that to buy tweaked thus the
project in the edge play human laughs at
after safely uma Gazette's Verdun you
know concrete has known or ahista Snoopy
did send her baby or take via neue
stylee which often give you our visit
Disney aha doesn't it mean it's finished
and the cooking now for this a foursome
project a given someone on Katy signs
own and you seen Diana sites on Estella
in big Italian sent home I Nevada first
L in abbastanza Hartung old then happy
it's vanished illness is 510 percent D
Indian exam pretty apparent eighty
science all as a gimmick no hands -
Ashley since all zone and Mr Al to occur
or the illustration graphic in cotton
etc etcetera
dan is this of carbonic new proportion
expected to up items on an email I can
get Elysium project um Suzette and I can
t digital easy expect haven't beat s up
in their plan on Val Venis vision can be
filled fortunes projectile from sukham
Zoltan be my conversions proactive so
can be on the lighter and trotzdem log
Bischoff Deacon 30 zone I thought plant
gave is Amazon is man in variable on
topics can efficient and resource and
iron sights on her image by hunting's
balaam force inspector so other states
so feel tank
[Applause]
you will take questions in English and
in German as well yeah at the moment
we're discussing this issue but we have
no solution for this because we are to
begin everywhere at the beginning of the
process and you saw we started 2018 so
we're quite late yes generals are a
little bit late but now we are coming
and now we now we're talking yeah now
we're talking about about these issues
and yeah yes and then we have another
virtual presentation this is
collaboration with researchers content
presentation systems and techniques we
have two lapekin and you see Pekka harka
Renan and library it's part of this
conference I was good to say is this a
subtitle no it's the conference
wonderful thank you ok ok
so you'll answer questions cooperation
with the region s responder presentation
systems and techniques I'm told of
acronym and this talk was
human ancestors also wear it briefly a
partisan apparently not so we are part
of University of Helsinki an independent
Institute there and our origins are from
the 1614 and the Academy of Thor who was
founded and from then on the library has
existed in some form or another so we
have been industry destroying and then
international rain and then - during the
independence also yeah but let's go
forwards to the shore topic so we have
had this kind of inter library regional
services virtual which has been
developing these kind of researcher
services so for a year - and our aim
there has been kind of how to how we can
build upon the existing strength so it
goes to the certain key things that we
feel that would be important for our end
users so first thing there is correct
increasing the awareness so having the
opportunity to to tell people tablet
chest and our material is they studied
everything where these materials can be
used then we want to increase the
finally it is so for example photo in
corporate materials how to find it
excessive then the other pressing need
is for the collaboration because
recently done in collaboration with for
example memory organization so that's
important and then we want to support
the research in publishing of the data
so how they could utilize it later or
then how other speak until a certain
excessive data and we want to be there
for the complete research take a life
cycle song from the beginning to the end
and the next rounds of the research and
the natural competence development is
important also for within the library
but also to gods researchers so are we
could learn from each other and
develop things on was with the joint
effort so what we have currently done
research resistant for one example there
is for example the hackathons where we
have artists that are already impact
let's say so we have been in the digital
humanities hackathons from year 2015 and
then in the digital humanities in Nordic
the hugging technique works off in 2018
part of England events and then in the
white 2018 that was the University of
Helsinki innovation hackathon of last
year and usually it is pin so that we
have in the material provider so we have
given an API or dataset for the
hackathon the use or then we have filled
up even in the positive tone there one
publish there's also being that that we
need to go there researchers are so we
have had this kind of research on it
event so to speak so for example we have
this inspection portion in the fall in
the spring side which you supercars are
gonna start organizing now then there
has been this hell dick summit which is
the symposium of the digital humanities
in Finland and then what kind of other
research event what some of this recent
data humanities that is coming in the
halls or nvivo and then their language
studies course works well for somewhat
of environment language Bank of England
in there so we have had our research
down there also this he has been in
effort has been this data current enix
so we have been visiting various
universities in finland so you should
have been a colleague of mine have been
going to the five universities so over
time very you just connect or but you
also and there any possibility showed
National Opera finest materials
presentations of steps and possible
tools that you could use with the money
but manipulating the material and there
has been about four hour session with
15,000 people and this has been
organized together with the library of
the University so
that as far also being one good conduct
point so thanks for them and and they
have been opportunity for questions and
odds for so long
so getting some gonna even side of what
the researchers are doing and and and
then discussing about the features and
the data that we have which has been
very valuable
well nationally international opera or
figure we have many the presentation
systems which have got a very different
aims and one thing that we keep thinking
that should we have specific features
for researchers or is it okay to have
kind of generic features that are fit
for anybody
let's say so them so it's not you know
so what's up on one example is the Orion
which is kind of this institution
institutional repository or service so
it's possible to store your documents
Schwein sets and so what they then we
have the in the field it is turns the
collection of the Hispanic white you
live resolution so it is kind of
national aggregator of the older
metadata that if they exist for all of
the materials across because across them
these these organizations and they are
nice developer API as metadata Ava
available and Phil know it's actually
part of the national research
infrastructure assets and then we have
it in our collection stuff time mostly
avail oh no so then we have newspapers
journals FM area
music speeds books manuscripts and maps
here it's possible to this kind of
text-based search to the home contents
of the material and then it's possible
to take graph visualizations or getting
the data to the Excel and then money for
it for us but naturally one thing that
people want is the day that data so we
have also
the catalog whether it's possible to get
the listing of all the datasets that we
have prepared and there's also these API
these things or giving instruction how
to use of us about api's in the
different percentage of systems it's ok
ok so I have talked about the
presentations and the data that we have
a bit so next thing is that hotels how
to have these techniques so we have
think I'm doing this term self help
instructions for the researcher so the
idea expanded how to how to get results
to help themselves so that they could
get further on on their own and then
when they get stuck so they would get
help from us directly and this is
something that is evolving and
innovating and seeing the power um how
much district requires and how this
could be enhanced the one last year a
couple of our people in this building
library lab sister and here is an
illustration of library lab that they
were discussing there and exactly what
what we have the things that we have not
oriented things within the national
library but the heaven they don't have
the overlap concept so the thieves
kismet various experiments the spaces
all of these are common aspects that
will be happy already ventured but we
don't have the overall laughs team
existing yet so that's one thing that we
need to think about one thing is that
should development be a project so this
is the University of Helsinki project
medical management what also it is it's
quite an extensive but in a way why not
this will give a good structure for the
lab development and it would kind of
give guidelines how to think is holding
forwards that so that they would have
successful outcome or at least that we
would have learned from the from the lab
development assets
so what next
so we are here so it is important that
we would learn from
other laughs Oh what to me make you
thick and what worked and what didn't
and from then on stop traveling doing
this roadmap photo doesn't every lap I
guess one key thing there is that how we
incorporate the existing library
services the black content because
that's kind of the whole like we're
saving starts are how we kind of modify
the existing services through this lab
concept and incorporate that there and
kind of the main thing is that how we
fit within our own University for ten
universities nationally and they're also
there's different reasons within Finland
so I am for example I need Kelly so how
will power a local or a channel laughs
and you can do with the whole app let's
say so can we have same things or how
they organize them and I guess they'll
kill anything is that we need to start
from with small steps and and then build
on ones so so okay thank you
can we have questions yeah there was a
comment from Stella wisdom and thanks we
did stole your ideas of this Roadshow
and it did work out it was tedious work
since Finland is bit larger than the
Netherlands and it's quite a lot of time
for travelling across Finland but it did
paid out because through these state
clinics we managed to integrate our
services into th teaching and I think
that is cumulative process then but now
if you have any questions or I can try
to answer the length
chuckling thank you for pointing out on
Twitter your your data catalog that's
I'm just playing with it right now it's
very useful are these are the data sets
here all open data sets and if so how
would you approach the problem of
delivering licensed yes that's a problem
in Finland we do have this collects a
collective licensing system for those
newspapers and items that are under
copyrights so we release only the open
and public domain data over there in
Colaba where we have had pilot projects
in this year and last year in finland in
which we have managed to open the in
corporate data for researchers and but
but they must lock in and and they must
be verified users that they can use the
data and there are some restrictions how
they can release the date set for
instance and whether that is in
copyright or not that is a question that
we're working with no no and we are pro
we are trying to create another fellow
approach for three years for with all
universities in Finland where they all
the use users that have this possibility
to access our database can have all in
corporate dates
[Applause]
I think this is our sixth one which is
great
partnering up with researchers in a
national library we have Martin Stephen
Sarah and lotta are you all going to
present no okay hi all my name is
Martine Klepper I was the guy who was
shifting around the presentation all the
time that's basically all my role for
today everything is done by a law signed
by Sarah and I'm presenting four types
of collaborations that we do with
researchers and that's basically a joint
effort of several colleagues so they're
all up here so it's lata but it's also
Stephen oozing over there and Sarah
who's at home and we all work at the
National Library in the digital
scholarship team on several departments
and this is basically where you are
right now if you're going on a tour I'm
not sure if the tours are going through
the stacks I hope so yes okay wonderful
you get a nice impression of all the
books and 125 kilometers of books that
we have downstairs but if you take the
elevator up to the fourth floor you get
into our hi-tech data center and this
elevator is really a nice time capsule
where you go from the from the books
right now to the library of tomorrow and
what we are going to show you are some
of the collaborations we do with
researchers who are using all our
digital collections so we have all sorts
of services for the general audiences
services that are free of charge for
example Dell for data Nell it's a
full-text search engine through our
digitized books periodicals newspapers
and radio buttons you can do any query
you get results you can narrow down your
results through using facets and you can
read to the newspapers get the OCR text
it's a nice service very popular in my
neighbor of 75 years old uses every day
to search this is family backgrounds and
it's a nice way to do a a close reading
approach of the historical sources that
we have so that's Dell firm another
service that we have and that's actually
run by staff and results McGann who's
hitting it back it's called data
services and that's aiming for
researchers who want to not only do
close reading but do decent reading get
all the data that's out there we saw
some amazing approach this morning by
the French National Library so this is a
service that we provide and we also have
agreements with the rights owners that
we can provide these data to researchers
once they sign the contracts and do with
it what they want as long as they don't
redistribute it and they do all sorts of
great stuff with it where we also
benefit from they for example make all
sorts of tooling we also make tooling
that we make available through our lab
website it's also another another
service that we have and a completely
different service that we have here at
the National Library is is our our
service that we provide for the Public
Library System
so since a couple of years we are
responsible for what is called the
digital infrastructure of the Public
Library System I can tell you all about
a lot about this but let's do that over
lunch but one of the services that we
provide this is the online library where
members of the public library can borrow
ebooks under a year or there's a
different app it's very popular very
nice actually just downloaded ten books
children's books for my daughter
yesterday I guess or the evening and you
can browse through them you can download
them you can find all sorts of books so
we have quite a large range of users as
a national library and for me this
pyramid always helps me quite well on
what type of users we are targeting with
our services and the parent doesn't say
that the top users are the most
important one where they're in size the
smallest one but maybe intensity the
biggest ones so at the bottom they're
the people who are using services that
we build for the for the public library
and it's a lot right so two million
adults two million children so that's
four million people that we serve the
middle layer are are what's called the
professional researchers or also my
neighbor for example who's using Dell
for doing distance or doing close
reading of books we have several
services of them and on the top of the
pyramid are the are the data scholars or
the data scientists or the humanities
scholars using for example our data
services or a lab it's a small group in
size but the amount of data that are
using is big and they do a lot about it
and in my presentation I will focus on
the Culebra
issues that we have with this top group
the the data scientists and what types
of collaborations we have with them so
what I want to show you is how do we
collaborate with them but we do not only
collaborate and serve them but we also
collaborate with them to improve our own
services and I show you a couple of
examples and I'm quite inspired by two
people in in this approach where we do
not only serve them but we also try to
learn from them to improve our own
services and one of them is Peter
Leonard currently in Utrecht attending
another workshop which he hosted
themselves otherwise he would have been
here but he gave a really nice keynote
three years ago already at a Libre
conference where he said what we want to
do what Yael is putting TDM or text and
data mining in a mainstream importance
for big audiences then research it only
and that's actually what we're doing as
well and I also liked this keynote by
Alex Humphries of JSTOR he gave a really
nice presentations about text analyzer
wonderful to buy JSTOR where you can
upload a text it's it's analyzed on the
fly and get all sorts of recommendations
of the data that is in JSTOR but is
using all sorts of d-h techniques and
what he says well what we are basically
doing at JSTOR is applying or is doing
applied to digital humanities supplying
techniques from the da's domain in
bigger bigger portals and that's also
what we are inspired by and try to do so
I'm going to present you five types of
collaborations that we have really brief
without any technical details but just
to give you an idea of how we
collaborate with researchers and one of
them is providing data actually already
showed it giving access to data that's
out there one example that worked out
quite well is this this is Martha marks
he's a computational political
communication science yes at the
University of Amsterdam he does a lot of
research about debates historical
debates that have been digitized have
been made available for text and data
mining purposes and he also wanted to do
analysis of newspapers how they were
linked to the debates in the parliament
next door and he said well can I have
the data of your newspapers because I
want to do all sorts of stuff with it so
we were allowed to so we provides him
with
and he says well I want to make an
Engram viewer where it can analyze the
word frequencies over time so I build
this by himself together with with
within a UX company and we liked it a
lot we thought this is useful for your
research but those who follow
researchers but maybe also for other
other users so we incorporated in our
lab to make it available for the users
and in the end and this is one of the
stories that we're really happy with we
were able to incorporate it into Dell
firm the service that I showed before so
actually you can now type in any query
in the offer and you can get these
engrams we completely rebuilt the code
ready-made but it was a great
inspiration for us to use and how my
neighbors using it which is which I love
another way of collaboration is the
research and residence program that
we've now been running for five years
which is great is actually the way how I
got to work at the KB because there was
one of those research and residents it's
a program that we set up five years ago
to invite early career researchers to
come and work with us for six years part
time and we did that first to understand
their changing needs we've seen all
sorts of presentations already on
changing scholarly practices and by
inviting them over to comment work at
our offices we really got to see what
their questions were so they had a
primary a research question that they
were allowed to well to answer within
six months working with us they team up
with a research software engineer and a
digital scholarship at Pfizer and
actually release it so this is our
offices upstairs
they really sit over there and it helps
us a lot by gaining understanding what
they do but I also built all sorts of
stooling there is of interest so that we
publish in the lab but it's also used
for for other services so an example
this is Frank the guy buff he was one of
researchers I think already three years
ago two years ago okay okay well so
anyway he's an old research researching
residence and he's a media historian and
he is interested in shifting types of
journalists in newspaper articles and so
he wanted to know if the interview how
it was published changes over time
or the column or the repertory and II
wrote a and he uses del fer to do that
before but in Delft where we have a
possibility to to select types of
articles so you can select
advertisements or illustrations an
article but articles basically
everything so there's another
distinguishment of other types of
articles but he wrote a dissertation
where he manually coded with a group of
researchers a lot of articles saying
this is an interview this is a column
this is a reportage so what he did
together with Juliet is a former
colleague for Marie's research software
engineer is building a classifier based
on machine learning techniques to
automatically classifies newspaper
articles based on the work that he did
before which was really nice so they've
also built a demonstrator you could try
it out where you could put up a a text
and then the algorithm says well I think
with a 87 point 8 confrontation rates
this is a new Jean and this is really a
nice nice example for us also to learn
more about machine learning but also
where we think hey is this something
that we could incorporate in the offer
to help
how big a big audiences we collaborate
in all sorts of Research Consortium in
different ways providing data but also
as an active research partner so what is
quite nice is about the project of
Franck that I showed we managed to get
additional funding relatively big with
to our research partners so they
continue working working on this and
largely we formulated a research agenda
to prioritize the topics that we want to
pay a lot of attention within research
projects out there Wow okay other
project that we also also have I'm
pretty there a fourth one which I liked
a lot we did this year was participating
in a workshop organised by the Dutch
national funding body of scientific
research it's called the ICT with
industry workshop and we're as a
organization you can submit a case on
all sorts of topics and researchers then
can register to participate in this for
a full week they get a hotel and big
food great food and a lot of knowledge
so we did this and the topic was about
creating automated metadata and was a
wonderful way for us also to kickstart
these kind of researchers and build a
new network
and this year we participated for the
first time in benchmark um tests can
tell a lot more about this actually we
participated in to where we provide data
to researches on different types of
tasks one was really academic tasks one
was more toast that was fitted our needs
and where researchers could use the data
in the work with it so as a wrap-up cool
so how do we collaborate and how do we
try to improve our services by
collaborating with them so these are the
users these are basically users they're
using services like like we have and in
the case of Martin we really managed to
build it into a services for bigger big
options and then if I go to these guys
saying well are we done yet I not yet I
think we are not there yet
because we also have these users the
public library users which is a great
group where we can do a lot of great
work for for example for the ebooks
finding ebooks is difficult so we're
looking at how can we make better
transparent recommendation systems so in
the end we would like to drill down the
the insights that we have for for these
audiences from the read more here is and
thank you for attention
[Applause]

so I'm Zoey borovsky I'm the librarian

Toggle Transcript

services::noun | Spocien Dictionary

services

noun

performance of duties or provision of space and equipment helpful to others

Type	Words
Type of	employment, work

the mayor tried to maintain city services.

the medical services are excellent.

How to pronounce services in Dutch?

services::noun | Spocien Dictionary

abandonment

Related words

Related

Subscribe to our Newslatter