ADAPT Radio The ADAPT Centre

Conversations about AI are often future focused, but these emerging technologies can also help us bring the past to life.

Today we find out how AI has been used in Beyond 2022, a flagship research project led by Trinity College Dublin which will digitally recreate seven centuries of historical records of the Public Record Office of Ireland destroyed by fire at Dublin’s Four Courts at the beginning of the Irish Civil War.

Joining us on ADAPT Radio to tell us more are Senior Researcher at the Virtual Treasury of Ireland of Trinity College Dublin, Dr. David Brown and Managing Director of Transkribus developer Read-Coop, Andy Stauder.

THINGS WE SPOKE ABOUT

01:08 What is Beyond 2022?
02:55 How the technology of Transkribus works
05:33 Processing data for historians
07:40 How AI can decipher handwriting
11:42 Recovering and digitising lost records and documents
15:35 Transkribus In Use
17:34 Using Chat GPT and other large language models to reverse cultural loss
20:27 Preventing bias in AI processing
24:17 Other potential applications of this technology
26:54 Futureproofing data accessibility

GUEST DETAILS

Dr David Brown is Senior Researcher at the Virtual Treasury of Ireland, a project based at Trinity College Dublin that aims to recreate a digital model of the Public Record Office of Ireland with its contents, destroyed in 1922. Starting in 2018, the Virtual Treasury of Ireland has developed a suite of deep-learning ‘models’, perfectly curated transcriptions, to train an AI system to read digital images of historical sources relating to Ireland and convert these into searchable text files. The automatic conversion of handwritten historical documents into searchable text is the latest major step in the digitisation of our written cultural heritage.

Website: https://virtualtreasury.ie/

Andy Stauder studied Translation Studies (German, Ital., Engl., Russ.; two Bachelor's, one Master's and one PhD degree) in Innsbruck, as well as Linguistics (Master's), Philosophy (Bachelor's) and Computer Science (not completed). His dissertation topic and research focus was the measurement of translation quality with combined machine- and human-based methods. In addition to his studies, he worked as a translator, IT manager and project collaborator in academic and EU-funded research projects and took over the technical management of the University of Innsbruck's subsidiary, innsbruck university innovations, in 2012, its management in 2016 and became managing director of READ-COOP SCE, the provider of the software Transkribus, in 2019.

Website: www.transkribus.org
Social Media: https://www.linkedin.com/in/andy-stauder-22a2b191/

MORE INFORMATION

Adapt Radio is produced by DustPod.io for the Adapt Centre
For more information about ADAPT visit www.adaptcentre.ie

QUOTES

I think it shows how far you can go with AI because we can't read it. So we're looking at this and going well, maybe we can verify the text and check any mistakes and so on. Actually, we can't read the stuff at all. So the AI is actually doing an amazing job. - David Brown

You can search for, say, your family name, and you'll see your family name coming up on an old handwritten document from 500 years ago. People find this really, really interesting and engaging and an awful lot of fun. - David Brown

Everybody knows Chat GPT at this point, the GPT model that's at the base of this is also transformer based. What the language models can do now is interact in a more natural way with the user. So you can ask the document itself what it contains. They are some new avenues that we are pursuing, currently. - Andy Stauder

KEYWORDS
#ai #handwriting #language #Transkribus #history

Since ChatGPT burst into the mainstream late last year, these conversational AI interfaces have been prompting huge discussion about how AI will shape our futures.

Today we’re learning how these AI chat bots work and where they can be used to the human advantage. Plus, we hear how they’re redefining intelligence, but not enough to overtake the human mind just yet.

Joining us to talk about this today on ADAPT Radio are Professor Vinny Wade, who holds the Professorial Chair of Computer Science and a Personal Chair in Artificial Intelligence at Trinity College Dublin and Professor John Kelleher, who is Professor of Computer Science at Maynooth University.

THINGS WE SPOKE ABOUT

01:18 What is a language model?
03:49 The quality of language being generated by AI
7:35 Different uses of generative chat models
13:11 A new learning experience for students and educators
20:09 Existential threat and current concerns
25:57 Redefining intelligence
28:01 Data sets and GDPR
34:30 Exciting prospects in research

GUEST DETAILS

Professor Vincent Wade is co-founder of the ADAPT Centre for Digital Content Technology and holds the Professorial Chair of Computer Science (Est. 1990) in School of Computer Science and Statistics, Trinity College Dublin as well as a Personal Chair in Artificial Intelligence. He is also co-director of the DREAL Centre for Research Training. His research focuses on intelligent systems, AI and Personalisation. He was awarded Fellowship of Trinity College for his contribution to research and has published over three hundred and fifty scientific papers in peer reviewed international journals and conferences. In 2018, he was awarded the Provost Innovation Award, the highest accolade the university can bestow for international research impact. As Director of ADAPT, Vincent heads a world leading research centre in digital media technology (text, video, speech, image, VR/AR) and AI.

----

Professor John Kelleher is a Hamilton Institute Professor of Computer Science at Maynooth University. John’s core research expertise is in the areas of machine/deep learning and natural language processing. He is the TU Dublin lead in the ADAPT centre and the scientific lead for the Digital Content Transformation Strand. Within the ADAPT centre he leads research projects on language modelling, lexical semantics, machine translation, novelty detection, image captioning, dialog systems, and making AI more environmentally sustainable. John has been the academic lead on numerous industry projects across a range of topics and domains, including: anomaly detection, transfer learning, customer segmentation and propensity modelling, dialog systems and chat bots, and information retrieval and natural language processing.

MORE INFORMATION

Adapt Radio is produced by DustPod.io for the Adapt Centre
For more information about ADAPT visit www.adaptcentre.ie

KEYWORDS

#ai #language #data #artificialintelligence #adapt

We are in a world where our every move and thought is documented and databases.

With that data, future search agents and AI will be able to deliver us deliver hyper-personalised results which are based on everything we have done in the past. It could even use this information to act on our behalf.

Joining us to discuss this future are two people who are living it now. Dr Cathal Guerin is an expert on personal data. For 15 years he has logged his life through wearing a personal camera and other data collection points. He is joined by Professor Gareth Jones who explores way to enhance search engines through conversational interactions and various media formats such as speech, video, and music.

TOPICS WE DISCUSSED INCLUDE

02:14 Their approach to life logging
03:14 What it’s like to wear a camera 24/7
07:54 Limitations of speech recognition
13:15 The problem of protecting privacy
19:05 Thoughts on Chat GPT & AI
23:46 AI Search agents and your hyper personal life log
27:56 How do we make the future safe?

GUEST DETAILS
Dr Cathal Gurrin is an Associate Professor at the DCU School of Computing and a co-investigator at the Insight Centre for Data Analytics and the Adapt Centre. He leads a group of researchers dedicated to developing assistive technologies using wearable sensors and data analytics. The highly interdisciplinary approach advanced by his group, "lifelogging", integrates computer science, cognitive science and data-driven healthcare analytics to generate next-generation digital records of the individual. Cathal is also co-author of ‘Lifelogging, Personal Big Data’, 2014.

https://www.linkedin.com/in/cgurrin/
@cathal on Twitter

Professor Gareth Jones is a Principal Investigator with the ADAPT Centre at DCU. His research focuses on topics related to information retrieval and search technologies. His work encompasses a broad range of areas including speech and multimedia search, multilingual search, personalisation in search, search for personal information archives, legal and patent search, and more generally information retrieval models and evaluation. He is currently collaborating with Spotify on a benchmark evaluation task on Podcast Search at TREC.

https://www.linkedin.com/in/gareth-j-f-jones-8337a3ab/

With more people living even faster lives, our cities need to be smarter, more sustainable and easier to live in.

Dublin City Council are running an experiment in D8, using technology to measure everything from traffic movements to rainfall and even how full the bins get. They are using this data to create a ‘digital twin’ of the area and virtualise how changes will improve the quality of life. For example, they’re talking about a virtual cycle path which you can test with a VR cycle ride.

However, this smart city program raises questions about citizen data and how to best involve residents.

Our guests, Jamie Cudden from Dublin City Council and Aphra Kerr from the Adapt Centre are both working directly on the project and share details of their progress.

TOPICS WE DISCUSSED INCLUDE

01:05 The invisible technology behind the scenes which use data to make decisions and deliver services more smartly

04:00 Good practice around data governance as ADAPT works with Dublin City Council on Smart Dublin projects

07:26 How they are engaging citizens in new and creative ways and getting their input

12:35 The importance of protecting privacy

18:51 Who is responsible for implementing the technology in smart cities?

23:30 How the smart city approach has been evolving from an emphasis on connectivity to pilots on health and wellness initiatives

32:39 Why pilot projects offer the flexibility to experiment and learn before scaling up

38:08 The benefits of when data gathering, engagement and impact all fall into place seamlessly


GUEST DETAILS

Jamie Cudden is the Smart City Programme Manager at Dublin City Council. He established the regional Smart Dublin program in 2015. He has a track record of bringing together partners from industry, start-ups, academia and public agencies to accelerate deployment of new ideas using the ‘city as a testbed’. He was acknowledged as ‘Civic Innovator of the Year Award in 2021 by Harvard’s Technology and Entrepreneurship Centre. This was for his role in scaling up the Smart Dublin programme. He was also recognised by Trinity College Dublin as Innovation Partner of the ear for his pioneering work on Smart Docklands.

https://ie.linkedin.com/in/jcudden
Twitter - @jcudden

Professor Aphra Kerr is a Professor in Sociology at Maynooth University, Ireland. She is a funded PI and science lead within the Transparent Digital Governance strand at the Science Foundation Ireland funded ADAPT Centre for Digital Content Technology. She is also Principal Investigator on two projects collaborating with Dublin City Council and Smart Dublin and on the YouGamSI project which examines gambling marketing. Her current research focuses on the governance and social impacts of AI across media and everyday smart technologies.

MU Website - https://www.maynoothuniversity.ie/faculty-social-sciences/our-people/aphra-kerr
Google Scholar - https://scholar.google.com/citations?user=3Uop6uoAAAAJ
Linkedin - https://www.linkedin.com/in/dr-aphra-kerr-74b630/
Twitter - @aphrak

Our everyday lives are surrounded with everyday objects which have the ability to monitor, track and record everything we do. But what are the roles that AI, machine learning and government play in this gathering of surveillance data.

Sharing his considerable knowledge in the area is Professor David Lyon, who presented a fascinating seminar on this topic at the Adapt Centre.

TOPICS HE DISCUSSED INCLUDE
Smartphone Tracking
IT and Society
Todays Surveillance Landscape
Emergence of Surveillance Studies
Surges in surveillance after 911 and Covid 19
Surveillance and Democracy
Why Surveillance is Important
Automated and other kinds of automated surveillance

GUEST DETAILS
Prof David Lyon has been studying surveillance since the mid-1980s.
A pioneer in the field of Surveillance Studies, he has produced a steady stream of books and articles, starting with The Electronic Eye (1994). The latest is Pandemic Surveillance (2022).

He has led several large collaborative research projects on surveillance, with research funding totalling almost $8 million.
His work has been recognised in Canada, Switzerland, the USA and the UK with a number of fellowships, prizes, awards and an honorary doctorate.

CONTACT

Linkedin: https://www.linkedin.com/in/david-lyon-4bb8b714/
Website: https://www.surveillance-studies.ca/

Our health is one of our greatest assets, and it has many interconnected parts, including our mental health and the environment in which we live. How can artificial intelligence enable us to engage more with our mental health and environment, and help researchers better understand how to protect both?

Helping us discover the answers is Claire Gillan, whose lab is developing new approaches to studying brain health, and Marcus Collier (TCD) whose research explores the complex interface between social and ecological systems in informal wild spaces in cities.

TOPICS WE DISCUSSED INCLUDE

(01.08) How technology such as mobile apps can encourage a broader diversity of people to take part in psychology research in their ‘real lives’, not just in the lab

(05.03) Many plants grow wild in cities, but we probably don’t pay much attention to them or the spaces in which they grow

(10.04) If you want people to use technology to contribute their data to science, you need to think carefully about how to engage with and protect them

(12.51) Artificial intelligence is helping psychology research to better predict what treatments would work best for people

(17.25) Artificial intelligence can help more people tell researchers what they think about the environment, so we hear more voices and perspectives

(23.28) AI might be part of the solution to help people with mental health issues – at the moment there just are not enough human therapists, so might a ‘chat bot’ style of interaction help?

(26.14) Our guests Marcus and Claire make plans to discuss working together.

GUEST DETAILS

Professor Claire Gillan is an Associate Professor at Trinity College Dublin, where her lab is interested in developing new approaches to studying brain health in psychiatric and ageing populations – a key goal is to develop objective tests that can be used to diagnose individuals and predict who will respond to which treatment. Her longitudinal, smartphone-based research (www.neureka.ie) in dementia and mental health is supported by funding from the Global Brain Health Institute, and Science Foundation Ireland, and she was awarded an ERC starting grant for €1.5M in 2020 to address gaps in our current understanding of how we make and break habits.

http://www.gillanlab.com/
Twitter: @clairegillanTCD

Professor Marcus Collier is an Associate Professor of Sustainability Science in the School of Natural Sciences at Trinity College Dublin and a Principal Investigator with ADAPT. His research explores the complex interface between social and ecological systems in informal wild spaces in cities, and the potential of nature-based solutions in an urbanising world. He was the lead investigator of the EU-funded projects Connecting Nature from 2017 to 2022 and the TURAS project from 2011 to 2016, for which he received a Champions of European Research Award. In 2020, Marcus was awarded a prestigious European Research Council Consolidator grant for NovelEco, a citizen science project.

https://naturalscience.tcd.ie/people/COLLIEMA
https://noveleco.eu/team-member/dr-marcus-collier/
https://www.linkedin.com/in/marcus-collier-12a0417/
Twitter: @marcus_collier