Interdisciplinary Approaches to Computational Moving Images
6-7 July 2023
King’s College London, Strand Campus
This workshop brings together a select group of researchers in the fields of digital and computational humanities, film, cultural history, informatics, computer vision, and digital art, with the purpose of exploring together emerging computational approaches to the study of moving images.
Participants include researchers from leading laboratories in Europe, including the Cultural Data Analytics Open Lab (CUDAN) at Tallinn University and the Cultural Analytics Lab (CANAL) at the University of Amsterdam, as well as archives and digital preservation experts from public UK institutions such as the BBC and the BFI. The workshop is hosted by the Computational Humanities Research Group in the Department of Digital Humanities at King’s College London.
Over two days, we will consider the modelling of moving images as computational artefacts, and reflect on the past, present, and future of computational moving image studies. We will then discuss and actively experiment with several ways of encoding the flows of moving images in time: from shot lengths measurements to high-dimensional representations, computational techniques that might afford new perspectives on the constitution and analysis of cinematic time.
The workshop is broadly split between a day of introductions and theory, and a second day of practical work and plans for future collaboration. The workshop will take place in the Embankment Room (MB -1.1.4), except the public panel on High-dimensional cinema, which will be at the Nash Lecture Theatre (K2.31). See programme in the next page.
CUDAN participants are supported partially via the CUDAN ERA Chair project, funded through the Horizon 2020 research and innovation program of the European Commission (Grant no. 810961).
DDH researchers are contributing to several public talks and events as part of the The King’s Festival of Artificial Intelligence (Bringing the Human to the Artificial, King’s Institute for Artificial Intelligence).
The festival brings together speakers, exhibits, performances, demos, and screenings in an exciting programme of events from 24th-28th May 2023. The events are open to the public, and provide an opportunity to gather with academics, students, alumni and King’s cultural and industry partners to find out more about developments in artificial intelligence technologies, and the challenges and opportunities that arise from them.
This event brings together experts from across a range of disciplines who work on visual culture and AI – from art to facial recognition systems – to explore its opportunities and challenges. How do we live well with AI and the visual? And how do we address its systemic inequalities around race, gender and ethnicity? Speakers include:
Can computers be creative? Do AI image generators such as DALL·E 2 mean the end of art? Looking at different examples of computational creativity enabled by machine learning, this talk by Joanna Zylinska will aim to cut through the smoke and mirror effect surrounding the current narratives about ‘creative AI’. But it will also demonstrate some practices of machinic co-creation, in which human artists and engineers draw on robotics and AI to produce work that is both visually interesting and thought-provoking. Through this, the talk will raise broader questions about the conditions of art making and creativity today.
How might AI art impact society and humanity’s self-conception? Attend this live discussion from the makers of the Art & AI podcast. Hear new perspectives, deep insights and crackling debate from a unique mix of scholars from King’s, the Courtauld Institute and the National Gallery.
As AI advances, our interactions with chatbots and robots are becoming increasingly common. But what are the potentials and pitfalls of fostering friendships and intimacy with computer software and hardware? This talk explores our emotional connections with AIs and robots, from their ability to provide support and companionship to fears of dehumanisation and the loss of authentic human connection. Technology has the power to bridge social and personal gaps in our lives, while also raising important ethical questions about individual and cultural impact. Join us as we explore the complex and fascinating world of human-AI relationships and consider the implications for our future interactions with technology.
Modeling Doubt, Coding Humility: A Speculative Syllabus
At a time of increasing artificial intelligence and proliferating conspiracy, faith in ubiquitous data capture and mistrust of public institutions, the ascendance of STEM and declining support for the arts and humanities, we might wonder what kind of epistemological world we’re creating. Prevalent ways of knowing have tended to weaponize uncertainty or ambiguity, as we’ve seen in relation to COVID vaccines, elections, climate, and myriad political scandals. In this talk I’ll sketch out a speculative syllabus for a future class about the place of humility and doubt in various fields of study and practice. We’ll examine how we might use a range of methods and tools — diverse writing styles, modes of visualization and sonification, ways of structuring virtual conversations, etc — to express uncertainty and invite more thoughtful, reflective engagement with our professional and public audiences and interlocutors.
Bio
Shannon Mattern is the Penn Presidential Compact Professor of Media Studies at Art History at the University of Pennsylvania. From 2004 to 2022, she served in the Department of Anthropology and the School of Media Studies at The New School in New York. Her writing and teaching focus on media architectures and infrastructures and spatial epistemologies. She has written books about libraries, maps, and urban intelligence, and she contributes a column about urban data and mediated spaces to Places Journal. You can find her at wordsinspace.net.
AI: Who’s Looking After Me?, (presented in collaboration with FutureEverything) is a free exhibition and public events programme, running from 21 June 2023 to 20 January 2024.
AI: Who’s Looking After Me?’ takes a questioning, surprising, playful look at the ways Artificial Intelligence (AI) is already shaping so many areas of our lives, and ask if we can really rely on these technologies for our wellbeing and happiness. Presented in collaboration with FutureEverything, we explore who holds the power, distributes the benefits, and bears the burden of existing AI systems.
Most of us know very little about what AI is or how it works, but so much of how we’re cared for in different aspects of our lives – be it love, justice or health – is undergoing transformative change. ‘AI: Who’s Looking After Me?’ fractures this singular, monolithic ‘AI’ apart, and looks at the range of ways it’s changing how we’re cared for.
“So many of our conversations about AI treat it as this distant, sleek, even magical thing; our attentions are daily directed towards the latest product or scandal. In all this hype and marketing, I think we’re losing sight of the human — both in how AI technologies are made, and the many ways they’re already woven into our lives. To be able to grasp and shape the course of AI’s journey, we need to grapple with its messy, multiple realities and I hope this exhibition can be an invitation to do that. It’s characteristic of what we’re trying to do as a gallery, to nurture unlikely, inventive collaborations and dialogues and be a home for the cultural work that emerges from them.”
Siddharth Khajuria, Director of Science Gallery London
Exhibited works from the programme include:
Cat Royale is the futurist utopia where cats are watched over lovingly by an AI robot arm, tending to their every need. The film and installation documenting cats’ experiences with an AI caregiver probe the future impact of new technologies on animal care… and the trade-offs involved. The work from internationally renowned artist collective Blast Theory, currently cultural ambassadors for the Trustworthy Autonomous Systems Hub, will be accompanied by live research from author and computer scientist Dr Kate Devlin, King’s Department of Digital Humanities.
Each Saturday throughout the season, Sentient Beings will invite visitors to question their relationship to security and privacy within the digital landscape of AI assistants. Featuring an immersive soundscape, the work sees artist Salomé Bazin collaborate with Dr Mark Cote from King’s Department for Digital Humanities, and Jose Such and William Seymour from the Department of Informatics.
You can read further details and register here. An excerpt from the event blurb is copied below.
What if racism, sexism, and ableism aren’t just glitches in mostly functional machinery—what if they’re coded into our technological systems? In this talk, data scientist and journalist Meredith Broussard explores why neutrality in tech is a myth and how algorithms can be held accountable.
Broussard, one of the few Black female researchers in artificial intelligence, explores a range of examples: from facial recognition technology trained only to recognize lighter skin tones, to mortgage-approval algorithms that encourage discriminatory lending, to the dangerous feedback loops that arise when medical diagnostic algorithms are trained on insufficiently diverse data. Even when such technologies are designed with good intentions, Broussard shows, fallible humans develop programs that can result in devastating consequences.
Broussard argues that the solution isn’t to make omnipresent tech more inclusive, but to root out the algorithms that target certain demographics as “other” to begin with. She explores practical strategies to detect when technology reinforces inequality, and offers ideas for redesigning our systems to create a more equitable world.
Folgert Karsdorp (Royal Netherlands Academy of Arts & Sciences (KNAW), Amsterdam) and Mike Kestemont (University of Antwerp, Belgium), Forgotten knights, unseen sailors, and unapprehended criminals: applying unseen species models to the survival of culture
Abstract
Researchers of the past — whether historians, literary scholars or archaeologists — depend on the sources that have stood the test of time. That sample of history is usually far from complete, however. There are numerous reasons for this, such as natural causes (e.g., fires or floods), decisions at the level of archival policy (what do we preserve and what do we not?), and biases in the formation of the archives themselves. Data representing lower classes were long considered less relevant, for example, and thus socioeconomic factors likewise play a role in the survival of sources. In a series of recent experiments, we have explored how statistical methods from ecology can help us identify such gaps and biases in our knowledge. Those methods all find their basis in “Unseen Species Models,” which were were originally developed to estimate the number of unique species in an environment. Just as ecologists try to estimate biodiversity from an incomplete sample, we apply the models to incomplete historical archives to measure the actual cultural diversity. In this talk, we apply unseen species models to three cases. First, we show how these methods can tell us something about the forgotten medieval chivalric literature in Western Europe. We then apply an extension of the method to the historical archives of the Dutch East India Company, to map out the size of its workforce. Finally, we explore a generalization of the unseen species model with which co-variates of loss or absence can be mapped. We apply this extension to a dataset from historical criminology: the police registers of the Amigo prison (1879-1880) in Brussels, and show how the models can give us an estimate of the “dark number” of unapprehended perpetrators as well as the demographic composition of this group.
Bios
Mike Kestemont, PhD, is a full professor in the department of Literature at the University of Antwerp (Belgium). He specializes in computational text analysis for the Digital Humanities. Whereas his work has a strong focus on historic literature, his previous research has covered a wide range of topics in literary history, including classical, medieval, early modern and modernist texts. Together with Folgert Karsdorp and Allen Riddell he has written a textbook on data science for the Humanities. The persistence of cultural information over long stretches of time is his key research topic at the moment. In the new framework of Cultural Ecology, empirical methods are imported from ecology and biostatistics to provide innovative quantitative models of cultural change and survival. Together with his Polish colleagues Maciej Eder and Jan Rybicki he is involved in the Computational Stylistics Group. Mike lives in Brussels (http://mikekestemont.github.io/), tweets in English (@Mike Kestemont) and codes in Python (https://github.com/mikekestemont).
Folgert Karsdorp, PhD, is a senior researcher in Computational Humanities and Cultural Evolution at the Meertens Institute of the Royal Netherlands Academy of Arts and Sciences (Amsterdam, the Netherlands). His research focuses on modelling cultural change from an evolutionary perspective (e.g., why some cultural phenomena are adopted and persist through time, while others change or disappear). Additionally, he is interested in measuring cultural diversity and compositional complexity, and how we can account for biases in our estimations of diversity. To do that, he employs computational models from Machine Learning, Cultural Evolution, and Ecology. Besides cultural change and diversity, Karsdorp is also interested in teaching about computer programming in the context of the Humanities. Together with Mike Kestemont and Allen Riddell, he published a text book with Princeton University Press about using Python for Humanities data analysis. For more information see his website (https://www.karsdorp.io) and his projects on GitHub (https://github.com/fbkarsdorp).
Thea Sommerschield (Ca’ Foscari University of Venice, Italy), Restoring, dating and placing Greek inscriptions with machine learning: the Ithaca project
Abstract
Ithaca is the first deep neural network for the textual restoration, geographical attribution and chronological attribution of ancient Greek inscriptions. This AI model is designed to assist and expand the historian’s workflow, focusing on collaboration, decision support and interpretability. In this presentation, I will introduce Ithaca and guide you through the model’s architecture, design decisions and visualisation aids. I will also offer a demo of how to use Ithaca for your personal research.
Bio
Thea Sommerschield is a Marie Skłodowska-Curie fellow at Ca’ Foscari University of Venice. Her research uses machine learning to study the epigraphic cultures of the ancient Mediterranean world. Since obtaining her DPhil in Ancient History (University of Oxford), she has been the Ralegh Radford Rome Awardee at the British School at Rome, Fellow in Hellenic Studies at Harvard’s CHS and Research Innovator at Google Cloud. She co-led the Pythia (2019) and Ithaca (2022) projects, and has worked extensively on Sicilian epigraphy.
Piroska Lendvai and Claudia Wick (Bavarian Academy of Sciences and Humanities, Germany), Finetuning Latin BERT for Word Sense Disambiguation on the Thesaurus Linguae Latinae
Abstract
The Thesaurus Linguae Latinae (TLL) is a comprehensive monolingual dictionary that records contextualized meanings and usages of Latin words in antique sources at an unprecedented scale. We created a new dataset based on a subset of sense representations in the TLL, with which we finetuned the Latin BERT neural language model (Bamman and Burns, 2020) on a supervised Word Sense Disambiguation task. We observe that the contextualized BERT representations finetuned on TLL data score better than static embeddings used in a bidirectional LSTM classifier on the same dataset, and that our per-lemma BERT models achieve higher and more robust performance than reported by Bamman and Burns (2020) based on data from a bilingual Latin dictionary. We discuss the differences in sense organizational principles between these two lexical resources, and report about our dataset construction and improved evaluation methodology.
Bios
Piroska Lendvai (PhD) works at the Digital Humanities R&D Department of the Bavarian Academy of Sciences and Humanities (Munich, Germany), where she supports research in Humanities and Social Sciences via tools and approaches from language technology.
Claudia Wick (PhD) works as a lexicographer in the Thesaurus linguae Latinae project at the Bavarian Academy of Sciences and Humanities (Munich, Germany) that targets the compilation of a comprehensive dictionary for ancient Latin. In her spare time she pursues programming. To register to this seminar, please email Barbara McGillivray at Barbara.mcgillivray@kcl.ac.uk
Enrique Manjavacas (Leiden University, The Netherlands), Historical Language Models and their application to Word Sense Disambiguation
Abstract
Large Language Models (LLMs) have become the cornerstone of current methods in Computational Linguistics. As the Humanities look towards computational methods in order to analyse large quantities of text, the question arises as to how these models are best developed and applied to the specificities of their domains. In this talk, I will address the application of LLMs to Historical Languages, following up on the MacBERTh project. In the context of the development of LLMs for Historical Languages, I will address how they can be specifically fine-tuned with efficiency to tackle the problem of Word Sense Disambiguation. In a series of experiments relying on data from the Oxford English Dictionary, I will highlight how non-parametric and metric learning approaches can be an interesting alternative to traditional fine-tuning methods that rely on classifiers that learn to disambiguate specific lemmas.
Bio
Enrique Manjavacas Arevalo is currently a post-doc at the University of Leiden, working in the MacBERTh project developing Large Language Models for Historical Languages. He obtained a PhD at the University of Antwerp (2021) with a dissertation on computational approaches to text reuse detection.
Organisers: Andrea Ballatore (KCL), Jamie Larkin (Chapman Uni.), with the support of Zhi Ye (Nina) (KCL)
Registration for participants: if you want to attend, please register at this <form>.
Programme
⏱️ When: Thursday 18 May 2023, 10:00 am – 5:30 pm (UK time)
📍 Where (hybrid): King’s College London, Bush House, (S) 1.01 (lecture theatre 1). For registration, please go to the North Entrance (see this map and this photo of the entrance). The link for online participation will be circulated by email before the event.
9:30 – 10:00 • Coffee/registration 10:00 – 10.15 • Introduction by organizers 10:15 – 11:30 • Paper session 1: Data Analytics for Museum Management 11.30 – 11.45 • Coffee break 11:45 – 13:00 • Paper session 2: New Media Ecologies 13.00 – 14.00 • Lunch break 14.00 – 15.15 • Paper session 3: Museum Data Practices 15.15 – 15.30 • Coffee break 15:30 – 16:45 • Paper session 4: Museum Data Infrastructures 16.45 – 17.00 • Comfort break 17.00 – 17:30 • Discussion and activity 17:30 – 17:40 • Closing remarks 17.40 – 18:00 Networking opportunity
Paper session 1: Data Analytics for Museum Management
Chair: Dr Mark Liebenrood
‘How Italian museums are facing the digital challenge’ Mauro De Bari (University of Bari, Italy)
‘Challenges and opportunities for data based museums’ Pille Runnel, Pille Pruulmann-Vengerfeld & Agnes Aljas (Estonia National Museum, Estonia, & Malmö University, Sweden)
‘What a Wonderful World: designing a media ecosystem at MAXXI museum in Rome’ Carmen Guarino, Stefano Capezzuto & Daniele Bucci (University of Palermo, Italy & Ecosistemica)
Paper session 2: New Media Ecologies
Chair: Dr Chiara Zuanni
‘Sensory Experience as Interaction Data in Exhibition Spaces’ Izabel Derda (Erasmus University, The Netherlands)
‘Worlding ontologies. Towards more ethical museum databases’ Mariel Hidalgo-Urbaneja, Athanasios Velios & Paul Goodwin (University of the Arts London, UK)
‘Virtual Acoustic Objects in Museum Environments: Bridging the Inaccessibility of Interactive Cultural Objects’ Dominik Ukolov (Leipzig University, Germany)
‘Samplebar Kenya and Beyond: On Interactive Ways to Digitise Traditional Music’ Kahithe Kiiru & Hakan Libdo (Bombas of Kenya & Libido Music AB)
Paper session 3: Museum Data Practices
Chair: Dr Serena Iervolino
‘Counting the small majority: correcting sampling bias in online museum studies’ Ellen Charlesworth (Durham University, UK)
‘Biases in data: the case of online museum catalogues’ Anna-Maria Sichani (School of Advanced Study, University of London, UK)
‘Enriching Exhibition Scholarship’ Clare Llewellyn (University of Edinburgh, UK)
‘Unlocking Collection Histories: Provenance Data and Agency’ Fabio Mariani (Lynn Rother & Max Koss, Leuphana University, Germany)
Paper session 4: Museum Data Infrastructures
Chair: Prof Alex Poulovassilis
‘Humanities Research, Museums, and Linked Data’ Toby Burrows, Deb Verhoeven & Mike Jones (University of Western Australia, University of Alberta, Canada, & Australia National University)
“Sloane Lab: Domain Vocabularies for Semantic Interoperability of Museum Collections” Andreas Vlachidis & Daniele Metilli (UCL, UK)
‘Generative AI and the Museum: Working with Collections as Training Data’ Joel McKim & Alessandro Provetti (Birkbeck, University of London, UK)
‘Unlocking the connective potential of Oral History through Natural Language Processing’ Stefano De Sabbata, Stefania Zardini Lacedelli, Alex Butterworth, Colin Hyde, Sally Horrocks, Neslihan Suzen, (University of Leicester, UK & Science Museum Group)
🍻 Conference pub (from 6 PM): Museum Tavern, 49 Great Russell St, London WC1B 3BA
🍽️ Conference dinner (self-funded): please contact organisers for information
