Humanist Discussion Group

Humanist Archives: Dec. 10, 2021, 7:12 a.m. Humanist 35.401 - problem fixed & old messages rescued

              Humanist Discussion Group, Vol. 35, No. 401.
        Department of Digital Humanities, University of Cologne
                      Hosted by DH-Cologne
                Submit to:

    [1]    From: Willard McCarty <>
           Subject: long-lingering problem fixed! (24)

    [2]    From: Fishwick, Paul <>
           Subject: Re: [Humanist] 35.302: A Biography of the Pixel (with review & commentary) (108)

    [3]    From: Flanders, Julia <>
           Subject: Call for participation: Word Vectors for the Thoughtful Humanist (55)

    [4]    From: Robin Douglas Burke <Robin.Burke@Colorado.EDU>
           Subject: Re: 35.116: an oppositional artificial intelligence (46)

    [5]    From: AEOLIAN Project <>
           Subject: AEOLIAN Network Workshop Programme and Speaker Abstracts (276)

    [6]    From: David Zeitlyn <>
           Subject: Re: [Humanist] 35.2: low-level nittygritty (36)

        Date: 2021-12-10 06:53:14+00:00
        From: Willard McCarty <>
        Subject: long-lingering problem fixed!

My thanks to Kai Niebes (Köln) for finding and fixing the lurking
problem in Humanist software responsible for emptying messages of their
contents. Some of you will have received requests from me to resend
affected messages to me directly; some may quite rightly have begun to
suspect that some malevolent person had formed a dislike of them... :-)
No such thing. Just faulty software.

What if...? Yes, I know. There's software involved in processing
applications from prisoners for parole and worse. Apart from that, the
lesson here is that if something unexpected happens to your posting,
please let me know!

Following this message are those that Kai managed to rescue from the 
bin into which said software had tossed them. Other, time-sensitive ones 
I am not passing on.Apologies for the following miscellany, through which 
I very much hope you comb for still relevant postings.

All best,
Willard McCarty,
Professor emeritus, King's College London;
Editor, Interdisciplinary Science Reviews;  Humanist

        Date: 2021-10-17 17:42:08+00:00
        From: Fishwick, Paul <>
        Subject: Re: [Humanist] 35.302: A Biography of the Pixel (with review & commentary)

Alasdair makes an excellent point. I am centered in computer and information
science (CS) but very
interested in the digital humanities--especially, creative practice in the arts.
I agree that getting
the opportunity to study a humanities-related topic is something we are all
interested in.

Here are 2 approaches (#1 and #2):

1. The tool approach suggests that the DH scholar is the  consumer and the
product is created
by CS. This has been shown very useful and can work effectively.

2. The knowledge approach where discussions centered on the humanities can be a
way for
CS (Education) to be brought to the forefront. When I was in Exeter a few years
ago, my goal was
to get computer scientists interested in learning from the arts and humanities.
For example,
see that medieval bridge or Roman wall? Model these with a JSON structure or any
data structure. Make a computer program to synthesize the wall.

#1 is not really up my alley. #2 is where I spend all of my time.

There are significant hurdles. The main one being the culture of computer
science on
utility and vocation. Modeling the medieval bridge in JSON is not practical and
not solve "a problem." Who cares about these things? Most CSers do implicitly
like this, but not all.

I am glass-half-full and wanting to bolster #2.


Paul Fishwick, PhD
Distinguished University Chair of Arts, Technology, and Emerging Communication
Professor of Computer Science
Director, Creative Automata Laboratory
The University of Texas at Dallas
Arts & Technology
800 West Campbell Road, AT10
Richardson, TX 75080-3021
Twitter: @PaulFishwick
ONLINE: Webex,Collaborate, TEAMS, Zoom, Skype, Hangout

Date: 2021-10-16 07:25:33+00:00
From: Alasdair Ekpenyong <>
Subject: Re: [Humanist] 35.300: pubs: A Biography of the Pixel (with review &

Dr. McCarty,

In response to your question about how to get digital humanities scholarship to
focus more on the mechanics of how computing works, in addition to the current
focus on social impact, I think the solution would be to get more data science,
computer science, etc. scholars to be interested in the digital humanities. We
as a community do a lot of work to introduce English students, history students,
etc. to the idea of using coding or programming to visualize their work, but I
don’t know that we do as strong of a job helping technically-trained students
feel comfortable and welcome joining into humanities conversations.

I’m in a Big Data Analytics course right now in a masters program, and when
given the chance to choose a topic, my team of classmates quickly chose an
analysis of healthcare industry data. I had the option to explain to them that
we have the option of studying a humanities-related subject and that this, too,
could be considered Big Data, but I didn’t feel comfortable investing the energy
to try and start that conversation and probably get the idea shot down.

In their 2006 account of one of the first DH projects, "Sorting things in:
Feminist knowledge representation and changing modes of scholarly production,"
the scholars talk about how a collaborative team approaches to DH involves
bringing together humanities scholars and STEM scholars. (Link:;data=04%7C01%7Cp

I definitely see more room for opportunity for bringing STEM scholars feel
invited to the DH table and if necessary empowering the junior STEM scholars to
feel confident and capable of joining humanities conversations. I wonder if the
serious humanities seem as intimidating to some STEM scholars as the idea of
learning Python sometimes seems to some humanities scholars.


Envoyé de mon iPhone

Le 16 oct. 2021 à 00:48, Humanist <> a écrit :

Not only are most of us
undereducated in mathematics, hardware and software engineering
and so on, but the sources of instruction one turns to tend to be written
for people within the technical disciplines, so it is an uphill battle.

A student recently complained to me that her lack of training on the
digital side of digital humanities made the path I was laying out close
to impossible. Is this not a problem we need to fix?


        Date: 2021-09-20 17:46:21+00:00
        From: Flanders, Julia <>
        Subject: Call for participation: Word Vectors for the Thoughtful Humanist

Dear all,

Applications are invited for participation in the final event in a series of
advanced institutes on text analysis, sponsored by the Northeastern University
Women Writers Project with generous funding from the National Endowment for the
Humanities. These institutes introduce teachers and researchers at varied levels
of expertise to the text analysis methods and interpretive questions arising
from word embedding models, which represent connections between words as
computable spatial relationships.

The full program includes four institutes, three of which have already taken
place in 2019 and 2021. On May 16–20, 2022 we will hold the final event in the
series, an intensive seminar focused on pedagogical uses of word vectors. This
five-day virtual event will run from 12:30–5pm Eastern and will offer a
thorough, well-scaffolded introduction to working with word vectors in R and
RStudio through commented code samples that can be adapted for use in
participants’ own teaching. Participants will learn how to build corpora and
train models of their own, and will also explore the challenges of teaching
command-line tools in a humanities context.

Each event is followed by a period of virtual discussion, consultation, and
support. Participants will be encouraged to share research and teaching outcomes
(syllabi, assignments, blog posts, research papers) and will be given the
opportunity to post preliminary results and work in progress on the WWP blog.

For information on how to apply please visit: and

Application deadline: January 31, 2022
Participants notified by: February 25, 2022
Preliminary schedule:

The previous institutes in the series are:
• An introductory institute focused on research applications of word vectors,
using the WWP’s web-based Women Writers Vector Toolkit (July 17–19, 2019;
• An introductory institute focused on pedagogical applications of word vectors,
using the WWP’s web-based Women Writers Vector Toolkit (May 24–28, 2021;
• An intensive institute focused on research applications of word vectors, using
R and RStudio (July 12–16, 2021;

Please contact with any questions.

Thanks and all our best,

Julia and Sarah

Julia Flanders, Director
Sarah Connell, Assistant Director
Women Writers Project
Northeastern University

        Date: 2021-07-05 22:25:01+00:00
        From: Robin Douglas Burke <Robin.Burke@Colorado.EDU>
        Subject: Re: 35.116: an oppositional artificial intelligence

Re: the idea of an opposition artificial intelligence, I offer the following
thread of recent research in music information retrieval and recommendation. The
authors in these works have been seeking to build systems that can spur musical
creativity through "oppositional" suggestions / recommendations:

Collins, N. (2010). Contrary Motion: An Oppositional Interactive Music System.
In NIME (pp. 125-129).

Knees, P., Andersen, K., & Tkalcic, M. (2015). " I'd like it to do the
opposite": Music-Making Between Recommendation and Obstruction. In DMRS (pp.

Knees, Peter, and Kristina Andersen. "A Prototype for Exploration of
Computational Strangeness in the Context of Rhythm Variation." UMAP (Extended
Proceedings). 2016.

Bauer, C., & Schedl, M. (2017, July). Introducing surprise and opposition by
design in recommender systems. In Adjunct Publication of the 25th Conference on
User Modeling, Adaptation and Personalization (pp. 350-353).

Some of these papers appeared in the two SOAP workshops (Surprise, Opposition,
and Obstruction in Adaptive and Personalized Systems) from 2016 and 2017:

These workshops were a bit broader as the title implies. I think the music-
related work is closest to the spirit of the term "oppositional".



Robin Burke (he/his), Professor, Chair
Department of Information Science
Department of Computer Science (by courtesy)
University of Colorado, Boulder
I may send email outside of working hours; I do not expect you to.

        Date: 2021-06-02 07:25:10+00:00
        From: AEOLIAN Project <>
        Subject: AEOLIAN Network Workshop Programme and Speaker Abstracts

Dear all,

The AEOLIAN Network (Artificial Intelligence for Cultural Organisations), a
project funded by the New Directions for Digital Scholarship grant from the US
National Endowment for the Humanities (NEH) and the Arts and Humanities Research
Council (AHRC), is hosting their first online workshop on Wednesday 7th July,
12:00 to 17:30 GMT. Please see below for our updated Programme and Speaker
Abstracts and Bios.

There is still time to apply for a place to attend the workshop, but
applications will need to be received by 18th June 2021. Please see our website
for more details:

AEOLIAN is designed to investigate the role that Artificial Intelligence (AI)
can play to make born-digital and digitised cultural records more accessible to
users. The project will make a ground-breaking contribution to this field
through carefully-structured workshops, innovative research outputs, and the
creation of an international network of theorists and practitioners working with
born-digital and digitised archives. Please visit https://www.aeolian- for more information.

Thank you,

Katie Aske
Research Assistant for AEOLIAN
Twitter: @AeolianNetwork

AEOLIAN Network’s Online Workshop 1: Employing Machine Learning and Artificial
Intelligence in Cultural Institutions

Wednesday 7th July from 12:00 to 17:30 GMT

12:00 – 12:10: Welcome from Dr Lise Jaillant (Loughborough University) and Dr
Annalina Caputo (Dublin City University).

12:10 – 13.30: Panel 1. Chair: Dr Maria Castrillo (Imperial War Museums)

Dr Giles Bergel (University of Oxford / National Library of Scotland)
Title: Visual AI and printed chapbook illustrations at the National Library of

Einion Gruffudd (National Library of Wales)
Title: Describing the Welsh National Broadcast Archive

John Stack (Science Museum)
Title: Machine Learning and Cultural Heritage: What Is It Good Enough For?

Followed by Q&A

13:30 – 14:30: Lunch Break (1 hour)

14:30 – 15:10: Panel 2. Chair: TBC

María R. Estorino (University of North Carolina at Chapel Hill Libraries)
Title: Rabbit Heart: Archives + the Machine

John McQuaid (Frick Collection), Vardan Papyan (University of Toronto), and X.Y.
Han (Cornell University)
Title: AI and the Photoarchive

Followed by Q&A

15:10 – 15:30: Interactive Session

This session is designed to generate casual discussion, share research
interests, and get to know other members of the network. Attendees will have the
option to attend one of 4 breakout rooms:

Room 1: Digital Management in Cultural Organisations
Room 2: Machine Learning and AI Projects
Room 3: Working Across Disciplines
Room 4: Developing International Projects

15:30 – 16:00: Comfort Break (30 min)

16:00 – 17:00: Keynote Presentation. Chair: TBC

Thomas Padilla (Center for Research Libraries)
Title: Keep True: Three Strategies to Guide AI Engagement

Followed by Q&A.

17:00 – 17:30: Roundtable. Chair: Dr Katherine Aske (Loughborough University).

Roundtable discussion with the AEOLIAN Project Team: Dr Lise Jaillant, Dr
Annalina Caputo, Glen Worthy (University of Illinois), Prof. Claire Warwick
(Durham University), Prof. J. Stephen Downie (University of Illinois), Dr Paul
Gooding (Glasgow University), and Ryan Dubnicek (University of Illinois).

Followed by Q&A.
Time Conversions (US)
(GMT-5) East Coast: 7:00–12:30 (break 8:30–9:30)
(GMT-6) Central: 6:00–11:30 (7:30–8:30)
(GMT-8) West Coast: 4:00¬–9:30 (5:30–6:30)

Speaker Abstracts and Biographies

Dr Giles Bergel (University of Oxford / National Library of Scotland)
Dr Giles Bergel is based in the Visual Geometry Group in the Department of
Engineering Science at the University of Oxford, where he works on the
application of visual AI to cultural heritage datasets. He has personal research
interests in book history, particularly cheap printed forms such as broadside
ballads and chapbooks, and has worked on a number of digitisation and
accompanying digital scholarship research projects on these forms. He is also
interested in the development of reproducibility standards for AI in cultural

Title: Visual AI and printed chapbook illustrations at the National Library of
Abstract: This presentation describes a project undertaken within the National
Librarian of Scotland’s Fellowship in Digital Scholarship programme for 2020-1.
The National Library of Scotland’s Data Foundry repository was created to
encourage the application of digital research methods to the collections: it
includes a large dataset of images, metadata and transcripts of Chapbooks
Printed in Scotland. Chapbooks are small, cheap books sold by travelling
pedlars, or chapmen, which comprise one of the most innovative and widely-known
forms of popular printed literature of their heyday (c.1700-1900). They are
frequently illustrated with relief (woodblock or stereotype) prints, which can
aid in printer attribution as well as providing evidence of popular visual
This project employed both a variety of computer vision methods to aid in the
analysis of the chapbook illustrations. Object detection, using a pretrained
classifier retrained on a small sample of the chapbooks, was employed to
identify the illustrated pages and to extract the illustrations from the corpus.
Next, the illustrations were matched using a visual search algorithm, clustered,
and made browsable per visual match and by means of the Library’s structured
metadata. Candidate matches were registered to provide a means of verification
of the closeness of the match, providing also a means of sequencing the printed
impressions, and chronological order of publication. Last, an image classifier
was applied to the extracted illustrations in order to explore intra-class
relationships and similarity to other relevant data.
The presentation will describe several forthcoming outcomes of the research,
including a methodological article; a machine learning model; and a dataset of
annotated images to encourage improvement of image-detection classifiers. Last,
the presentation will offer some reflections on the value of curated data within
AI workflows in cultural heritage, and the necessity of further curatorial
oversight of their outputs.

Einion Gruffudd (National Library of Wales)
Einion Gruffudd started his career as a video librarian at Barcud
television resources company in north Wales, before returning to Aberystwyth
in 1992 to work at the National Library of Wales where he has served in
the Manuscripts, IT and Unique Collections departments. His work has included
managing Library systems, business continuity, setting up NLW’s digital archive,
and successfully leading a HLF funded project to digitise all 1,200 tithe maps
of Wales. He has been managing the NLHF funded project to establish a Broadcast
Archive at NLW since 2017.

Title: Describing the Welsh National Broadcast Archive
Abstract: This talk will describe how the National Library of Wales
is establishing a National Broadcast Archive, a National Lottery Heritage Fund
supported project involving acquiring a large corpus of digitised audiovisual
material from Welsh broadcasters. This collection which will be made available
to the people of Wales for research purposes at various locations across the
The project includes a focus on making the collection more
discoverable, applying Artificial Intelligence technologies to Welsh
Language voice2text and keyword generation. These activities to improve how
the collection is described will include volunteer participation in
the correction of machine learning output, among many other activities
to promote the use of the archive. Issues raised by the ownership and clearance
of rights affect all activities including AI activities, and the project’s
approach to these obstacles will be explained.
The talk will examine how the location of the Broadcast Archive within NLW
brings different use cases for archive use, and opportunities to take advantage
of other digitisation activities and technologies developed at the Library. A
key focus for the end of the project, which will be described, is to develop a
"linked data experience" to help people understand the relationships between
broadcasting and other historical sources from the wide range of holdings at

John Stack (Science Museum)
John Stack is Digital Director of the Science Museum Group. The Science Museum
Group encompasses five museums: Science Museum, London; National Science and
Media Museum, Bradford; National Railway Museum, York; Science and Industry
Museum, Manchester; and Locomotion, Shildon. He joined in 2015 and is
responsible for setting and delivering the Group's digital strategy. He manages
the Digital department which encompasses the museums’ websites, digitised
collections, apps, games and on gallery digital media. Prior to joining the
Science Museum Group, he was Head of Digital at Tate for ten years.

Title: Machine Learning and Cultural Heritage: What Is It Good Enough For?
Abstract: Funded through the AHRC’s Towards a National Collection Programme, the
Science Museum Group (SMG) is collaborating with the V&A and School of Advanced
Study, University of London, on a two-year project entitled “Heritage Connector:
Transforming text into data to extract meaning and make connections”.
As with almost all data, museum collection catalogues are largely unstructured,
variable in consistency and overwhelmingly composed of thin records. The form of
these catalogues means that the potential for new forms of research, access and
scholarly enquiry that range across multiple collections and related datasets
remains dormant.
The Heritage Connector project is deploying a range of machine learning-based
techniques to extract information from the SMG collection catalogue, link it to
third-party sources – primarily Wikidata and the V&A’s collection – will then
create a set of prototypes that demonstrate and explore the affordances of the
resulting dataset.
Rather than attempting to deploy machine learning to create a perfect linked
data model, Heritage Connector asks what’s “good enough” to provide useful
functionality to different audiences.

María R. Estorino (Associate University Librarian for Special Collections at The
University of North Carolina at Chapel Hill Libraries)
María R. Estorino serves as Associate University Librarian for Special
Collections and Director of Wilson Library with the University of North Carolina
at Chapel Hill Libraries. With degrees in public history and library science,
she has spent 20 years in cultural heritage work, principally in academic
special collections and local history museums.

Title: Rabbit Heart: Archives + the Machine
Abstract: From fear to fluency: considering the role of the archivist/special
collections librarian in explorations of machine learning and artificial
intelligence in our work.

John McQuaid (Frick Collection)
John McQuaid is Photoarchive Lead at the Frick Art Reference Library. He
received a BA in Art History and Classics from Case Western Reserve and a MA in
the History of Art from The Ohio State University.

Vardan Papyan (University of Toronto)
Vardan Papyan is an assistant professor in the department of Mathematics at the
University of Toronto, cross-appointed with the department of Computer Science.
He received his BSc, MSc, and PhD at the Technion and was a postdoctoral
researcher at Stanford University.

X.Y. Han (Cornell University)
X.Y. Han is a PhD student in the department of Operations Research and
Information Engineering at Cornell University. He received his BSE in Operations
Research and Financial Engineering from Princeton University, and MS in
Statistics from Stanford University.

Title: AI and the Photoarchive
Abstract: In this talk, we describe a collaborative project between art
historians and staff at the Frick Art Reference Library (FARL) and researchers
at Cornell, Stanford, and the University of Toronto to develop an algorithm that
will apply a local classification system based on visual elements to the
Library’s digitized Photoarchive—a study collection of 1.2 million reproductions
of works of art. We leverage state-of-the-art artificial intelligence (AI)
systems to develop a classifier for the automatic annotation of digitized but
not-yet-catalogued images in the FARL’s Photoarchive. This was achieved by
engineering the syntax of the classification system into the training and
predictive process of deep convolutional neural networks, the cornerstone of
modern AI advancements.
The classifier is integrated into a mobile and desktop application that allows
Photoarchive staff to quickly validate or correct the decisions of the networks.
We demonstrate promising performance metrics and offer informative scientific
insights that have the potential to create a valuable tool for metadata creation
and image retrieval. This project offers a useful model for effective
interdisciplinary interaction.

Thomas Padilla (Director of Information Systems and Technology Strategy at the
Center for Research Libraries)
Thomas Padilla is Director of Information Systems and Technology Strategy at the
Center for Research Libraries. He is the author of the library community
research agenda, Responsible Operations: Data Science Machine Learning, and AI
in Libraries, Principal Investigator of Collections as Data: Part to Whole, and
past Principal Investigator of Always Already Computational: Collections as
Data. Thomas is Vice Chair, ACRL Research and Scholarly Environment Committee;
Executive Committee Member, Association for Computers and the Humanities; and
Technical Advisory Board Member, Linked Infrastructure for Networked Cultural

Title: Keep True: Three Strategies to Guide AI Engagement
Abstract: Recurrent bouts of AI enthusiasm over decades suggest no sector is
immune to losing itself in the face of potential. In the archipelago of varied
sector actors implementing AI, GLAMs have an opportunity to distinguish
themselves. While the component parts of this community are quite different and
sometimes functionally opposed in approaches to similar work, we share in common
a set of contemporary commitments that seek to advance equity in the communities
we serve. In what follows I will present three strategies I believe strengthen
our ability to realize these commitments: nonscalability imperative, avoiding
neoliberal traps, and seeing maintenance as innovation.

        Date: 2021-05-07 08:51:39+00:00
        From: David Zeitlyn <>
        Subject: Re: [Humanist] 35.2: low-level nittygritty

Dear all

Tom Boellstorff gave an Oxdeg seminar about the following book which
should be relevant

Tom Boellstorff & Braxton Soderman

Intelligent Visions: The Intellivision System, Video Games, and Society

The talk (not recorded) was about suitably nittygritty issues of early
programming problems...


David Zeitlyn,

Professor of Social Anthropology (research). ORCID: 0000-0001-5853-7351

Institute of Social and Cultural Anthropology, School of Anthropology and Museum
University of Oxford, 51 Banbury Road, Oxford, OX2 6PF, UK. The Virtual Institute of Mambila Studies

2020 Monograph:
Mambila Divination: Framing Questions, Constructing Answers (Routledge Studies
in Anthropology)
London: Routledge.  ISBN 9780367199500

Vestiges: Traces of Record Open Access Journal

Unsubscribe at:
List posts to:
List info and archives at at:
Listmember interface at:
Subscribe at: