Turing Lecture: Data science or data humanities? - Melissa Terras The Alan Turing Institute 2026-09-18 | Turing Lecture: Data science or data humanities? - Melissa Terras
AIM RSF Seminar: How dynamic artificial intelligence (DynAIRx) might be used to support... The Alan Turing Institute 2024-02-28 | Full title: AIM RSF Seminar: How dynamic artificial intelligence (DynAIRx) might be used to support prescribing to ensure efficient medication reviewsThe DynAIRx project aims to develop artificial intelligence (AI) tools to support medication reviews for patients with multimorbidity (people with ≥ 2 chronic conditions), targeting those at greatest risk of medicine-related harm. This session will explore how medication reviews are currently undertaken, the barriers and facilitators of implementing AI tools and the potential for AI tools to aid medication reviews for people with multimorbidity and/or polypharmacy.The slides can be accessed: zenodo.org/records/10666498
AIM RSF Seminar Translating Practice into Evidence Why multimorbidity research needs ethnography The Alan Turing Institute 2024-02-28 | Ethnography is small scale observational research conducted in naturalistic ‘everyday’ settings which seeks to describe, illuminate and interpret everyday phenomena. With its disciplinary roots in anthropology, ethnography remains underexploited in healthcare research but is particularly valuable as a way of exploring complex issues and offering insights that can inform efforts to improve the quality and safety of care.In her talk, Deborah Swinglehurst began by introducing ethnography as an orientation to research, explaining what it was, why it was valuable, and what kinds of methods might be employed in an ethnographic enquiry.Drawing on examples of recently completed and ongoing research projects conducted within her team, she highlighted the potential of ethnography to support new ways of thinking about the nature of some of our most complex contemporary healthcare challenges, such as multimorbidity and polypharmacy.Slides: zenodo.org/records/10534628Swinglehurst, D. (2024, January 19). Translating 'practice-into-evidence': Why multimorbidity research needs ethnography. Zenodo. doi.org/10.5281/zenodo.10534628
Ameya Jagtap Enhancing Scientific Computing Through Physics informed Neural Networks The Alan Turing Institute 2024-02-28 | Traditional approaches to scientific computation have made significant strides in progress. However, they still face challenges as they operate within strict constraints, requiring precise knowledge of underlying physical laws, boundaries, and/or initial conditions. Additionally, these approaches often involve time-consuming workflows, including mesh generation and lengthy simulations. Addressing high-dimensional problems governed by parameterized partial differential equations (PDEs) remains a formidable task. Furthermore, efficiently incorporating noisy data remains a challenge when solving inverse problems. Physics-informed deep learning (PIDL) has emerged as a promising alternative to address these issues. In this presentation, we will delve into a specific type of PIDL method known as physics-informed neural networks (PINNs). We will provide an overview of the current capabilities and limitations of PINNs, highlighting diverse applications where PINNs have proven to be highly effective compared to traditional approaches. Additionally, we will explore extensions of the PINN method, such as conservative PINNs (cPINNs) and eXtended PINNs (XPINNs), specifically designed for handling big data and/or large models. Throughout the discussion, we will also explore various adaptive activation functions that can enhance the convergence of deep and physics-informed neural networks.
Karan Taneja - Feature Encoded and Multi-Resolution Physics-Informed Machine Learning Approaches... The Alan Turing Institute 2024-02-28 | Full title: Karan Taneja - Feature Encoded and Multi-Resolution Physics-Informed Machine Learning Approaches for Musculoskeletal Digital Twin ApplicationsMachine Learning (ML) approaches offer effective tools to identify biological system properties from physiological measurements, e.g., motion data and raw surface electromyography (sEMG), providing opportunities to construct a subject-specific musculoskeletal (MSK) digital twin system for health condition assessment and motion prediction. While physics-informed ML approaches for dynamic systems offer learning capabilities that satisfy the conservation laws, physics-informed time-domain mapping of high-frequency muscle excitation signals to low-frequency joint motion remains challenging owing to the large variation in frequency contents between the activation signals (input) and motion data (output). In this work, we first developed a Feature-Encoded Physics Informed Parameter Identification Neural Network (FEPI-PINN) for the simultaneous prediction of motion and parameter identification of human MSK systems. Here, the features of high-dimensional noisy sEMG signals were projected onto a low-dimensional noise-filtered embedding space for effective forward dynamic training. This FEPI-PINN model can be trained to relate sEMG signals to joint motion and simultaneously identify the key MSK parameters. To enhance time-domain mapping, we then propose a Multi-Resolution Recurrent Neural Network (MR-RNN) learning algorithm. In this approach, the fast wavelet transform is applied to noisy sEMG signals, decomposing them into nested multi-scale signals. The prediction model is first trained with lower-resolution input signals using a gated recurrent unit (GRU), and the trained parameters are then transferred to the next higher-scale training. These training processes are repeated recursively until a full-scale training is achieved. Numerical examples demonstrate that the proposed framework can effectively identify subject-specific muscle parameters with noisy sEMG signals, and the trained physics-informed forward-dynamics surrogate yields accurate motion predictions of elbow flexion extension motion, which are in good agreement with the measured joint motion data.
Living with Machines: Computational methods, infrastructure and skills The Alan Turing Institute 2024-02-23 | Living with Machines' is both a research project and a bold proposal for a new research paradigm. In this ground-breaking partnership between The Alan Turing Institute, the British Library and universities of Cambridge, East Anglia, Exeter, Queen Mary University of London and King’s College London, historians, data scientists, geographers, computational linguists, library professionals and curators were brought together to examine the human impact of industrial revolution.This series of short videos in documentary form seeks to make visible the collaborative underpinnings of the project by highlighting the team’s experiences, research objectives, challenges, and lessons learnt.Living with Machines was funded by UK Research and Innovations (UKRI), via the Strategic Priorities Fund and was administered by the Arts and Humanities Research Council (AHRC).Find out more here: livingwithmachines.ac.uk0:28 The diversity of the sources and the LWM database2:28 Data Wrangling3:29 Building skills and communities of users
Living with Machines: The digitisation process The Alan Turing Institute 2024-02-23 | Living with Machines' is both a research project and a bold proposal for a new research paradigm. In this ground-breaking partnership between The Alan Turing Institute, the British Library and universities of Cambridge, East Anglia, Exeter, Queen Mary University of London and King’s College London, historians, data scientists, geographers, computational linguists, library professionals and curators were brought together to examine the human impact of industrial revolution.This series of short videos in documentary form seeks to make visible the collaborative underpinnings of the project by highlighting the team’s experiences, research objectives, challenges, and lessons learnt.Living with Machines was funded by UK Research and Innovations (UKRI), via the Strategic Priorities Fund and was administered by the Arts and Humanities Research Council (AHRC).Find out more here: livingwithmachines.ac.uk0:13 The digitised collections of the British Library open new horizons to research1:08 The challenges of digitisation2:52 Resolving copyright constrains and making the data widely accessible3:51 Making resources available to the community
How do we build equitable and inclusive AI Governance structures? The Alan Turing Institute 2024-02-07 | Speakers at the AI Fringe event “AI at a Turning Point: How Can We Create Equitable AI Governance Futures?” discuss the Turing's AI Ethics & Governance programme and building equitable and inclusive AI Governance structures.Find out more about the Public Policy Programme’s AI Ethics & Governance in Practice programme here:turing.ac.uk/research/research-projects/ai-ethics-and-governance-practice
Centring childrens rights in AI development. The Alan Turing Institute 2024-02-06 | Turing researchers and speakers at the AI Fringe event “AI at a Turning Point: How Can We Create Equitable AI Governance Futures?” discuss children’s rights and AI development.Find out more about the Public Policy Programme’s work on Children’s Rights & AI:turing.ac.uk/research/research-projects/exploring-childrens-rights-and-ai
What is Data Justice? The Alan Turing Institute 2024-02-06 | Turing researchers and partners discuss global data justice, filmed at the AI Fringe event “AI at a Turning Point: How Can We Create Equitable AI Governance Futures?”.Find out more about the Public Policy Programme’s work advancing data justice in practice: turing.ac.uk/research/research-projects/advancing-data-justice-research-and-practice
Using the Policy Priority Inference app | Step 1 | Indicators The Alan Turing Institute 2024-01-26 | The Policy Priority Inference (PPI) app is a responsive AI tool from The Alan Turing Institute's PPI research programme that models the complex causal link between government expenditure and multidimensional policy outcomes. It allows you to interpret and analyse your data through a simple, clear interface without the need for any coding, and providing interactive visualisations of the outputs.This is the first in a set of step-by-step animated tutorials on how to use the app, which can be found here: youtube.com/playlist?list=PLuD_SqLtxSdXMJpqEkBmwwJ2jzFhjqOkNPolicy Priority Inference (PPI) is a research programme from the Modelling for Policy theme in the Turing's Public Policy Programme, that aims at modelling the causal link between government expenditure and policy outcomes, while accounting for the multidimensionality and complexity of development. You can find out more about PPI here: turing.ac.uk/research/research-projects/policy-priority-inference
Using the Policy Priority Inference app | Step 5 | Analysis The Alan Turing Institute 2024-01-26 | The Policy Priority Inference (PPI) app is a responsive AI tool from The Alan Turing Institute's PPI research programme that models the complex causal link between government expenditure and multidimensional policy outcomes. It allows you to interpret and analyse your data through a simple, clear interface without the need for any coding, and providing interactive visualisations of the outputs.This is the fifth in a set of step-by-step animated tutorials on how to use the app, which can be found here: youtube.com/playlist?list=PLuD_SqLtxSdXMJpqEkBmwwJ2jzFhjqOkNPolicy Priority Inference (PPI) is a research programme from the Modelling for Policy theme in the Turing's Public Policy Programme, that aims at modelling the causal link between government expenditure and policy outcomes, while accounting for the multidimensionality and complexity of development. You can find out more about PPI here: turing.ac.uk/research/research-projects/policy-priority-inference
Using the Policy Priority Inference app | Step 4 | Calibration The Alan Turing Institute 2024-01-26 | The Policy Priority Inference (PPI) app is a responsive AI tool from The Alan Turing Institute's PPI research programme that models the complex causal link between government expenditure and multidimensional policy outcomes. It allows you to interpret and analyse your data through a simple, clear interface without the need for any coding, and providing interactive visualisations of the outputs.This is the fourth in a set of step-by-step animated tutorials on how to use the app, which can be found here: youtube.com/playlist?list=PLuD_SqLtxSdXMJpqEkBmwwJ2jzFhjqOkNPolicy Priority Inference (PPI) is a research programme from the Modelling for Policy theme in the Turing's Public Policy Programme, that aims at modelling the causal link between government expenditure and policy outcomes, while accounting for the multidimensionality and complexity of development. You can find out more about PPI here: turing.ac.uk/research/research-projects/policy-priority-inference
Using the Policy Priority Inference app | Step 3 | Interdependencies The Alan Turing Institute 2024-01-26 | The Policy Priority Inference (PPI) app is a responsive AI tool from The Alan Turing Institute's PPI research programme that models the complex causal link between government expenditure and multidimensional policy outcomes. It allows you to interpret and analyse your data through a simple, clear interface without the need for any coding, and providing interactive visualisations of the outputs.This is the third in a set of step-by-step animated tutorials on how to use the app, which can be found here: youtube.com/playlist?list=PLuD_SqLtxSdXMJpqEkBmwwJ2jzFhjqOkNPolicy Priority Inference (PPI) is a research programme from the Modelling for Policy theme in the Turing's Public Policy Programme, that aims at modelling the causal link between government expenditure and policy outcomes, while accounting for the multidimensionality and complexity of development. You can find out more about PPI here: turing.ac.uk/research/research-projects/policy-priority-inference
Using the Policy Priority Inference app | Step 2 | Budgets The Alan Turing Institute 2024-01-26 | The Policy Priority Inference (PPI) app is a responsive AI tool from The Alan Turing Institute's PPI research programme that models the complex causal link between government expenditure and multidimensional policy outcomes. It allows you to interpret and analyse your data through a simple, clear interface without the need for any coding, and providing interactive visualisations of the outputs.This is the second in a set of step-by-step animated tutorials on how to use the app, which can be found here: youtube.com/playlist?list=PLuD_SqLtxSdXMJpqEkBmwwJ2jzFhjqOkNPolicy Priority Inference (PPI) is a research programme from the Modelling for Policy theme in the Turing's Public Policy Programme, that aims at modelling the causal link between government expenditure and policy outcomes, while accounting for the multidimensionality and complexity of development. You can find out more about PPI here: turing.ac.uk/research/research-projects/policy-priority-inference
Introducing the Policy Priority Inference app The Alan Turing Institute 2024-01-26 | The Policy Priority Inference (PPI) app is a responsive AI tool from The Alan Turing Institute's PPI research programme that models the complex causal link between government expenditure and multidimensional policy outcomes. It allows you to interpret and analyse your data through a simple, clear interface without the need for any coding, and providing interactive visualisations of the outputs.We have developed a set of step-by-step animated tutorials on how to use the app, which can be found here: youtube.com/playlist?list=PLuD_SqLtxSdXMJpqEkBmwwJ2jzFhjqOkNPolicy Priority Inference (PPI) is a research programme from the Modelling for Policy theme in the Turing's Public Policy Programme, that aims at modelling the causal link between government expenditure and policy outcomes, while accounting for the multidimensionality and complexity of development. You can find out more about PPI here: turing.ac.uk/research/research-projects/policy-priority-inference
Hongbo Zhao - Learning Battery Physics from Images The Alan Turing Institute 2024-01-18 | With the development of advanced synchrotron light sources, we have an unprecedented capability to perform in-situ and in-operando measurement of the dynamics of lithium-ion battery materials, allowing us to observe lithium (de)insertion both in real and reciprocal space during charge and discharge. At the same time, advances in data-driven discovery of physics offer an unprecedented opportunity to extract hidden information from full experimental datasets. In this talk, I will share our work combining electrochemical models with partial-differential-equation (PDE)-constrained optimization and Bayesian inference to learn the physics of lithium intercalation materials from multiple data streams including 1. in-situ scanning transmission x-ray microscopy (STXM) images of reactive lithium iron phosphate (LFP) nanoparticles, 2. correlative imaging of lithium concentration and strain maps of relaxed LFP particle, and 3. in-operando X-ray diffraction data of lithium transition metal oxide. We extracted thermodynamic and reaction kinetic models, spatial heterogeneity, and chemo-mechanical coupling from these data streams, demonstrating the possibility of achieving full utilization of the dataset and extracting physical quantities that are otherwise difficult to measure. The results advance our understanding of battery physics and offer insights into engineering high-performance materials.
AI for Womens Health - December 2023 The Alan Turing Institute 2024-01-16 | The AI for women's health group is an interdisciplinary community (e.g. clinicians, academics, data scientists, designers, health care professionals, students etc) interested in advancing women's health and how to leverage AI to do so.The groups goals include:- Building an interdisciplinary and diverse community of individuals working or interested to work in women's health with (or without) AI.- Bringing visibility to research on women's health and how AI can help to do so.- Facilitating collaborations and bringing more people to the field.- Offering relevant events & trainings, and sharing resources with the community. This was the second meeting of the AI for Women's Health Group.In this meeting, we had two talks:- Michelle Oyen and Adrienne Scott from the Centre for Women's Health Engineering, Washington University, who are speaking about 'Digital twins and Machine Learning in Pregnancy Research'.- Sarah Homewood from the Department of Computer Science, University of Copenhagen, who will talk about 'Designing With Menstrual Cycle Data'.
Jan Olle - Discovering Quantum Error Correcting Codes with Reinforcement Learning The Alan Turing Institute 2024-01-16 | ...
Pint of science: Swipe right for privacy- preserving, fair and robust machine learning models The Alan Turing Institute 2024-01-16 | Swipe right for privacy-preserving, fair, and robust machine learning models - Aras SelviThe last five years have seen a massive shift in the way we view and understand our rights when it comes to data privacy. With countless social media giants hitting the headlines for questionable practices over the uses and misuses of our personal data, we’ve had to wise up, and quickly. Enter dating apps, and the promise of love with the perfect match throws all our good sense out the window, making us vulnerable to bad actors.Is it too much to expect private, fair and robust algorithms in the world of online dating? We say no!Pint of Science is a worldwide science festival which brings researchers to your local pub to share their scientific discoveries with you - no prior knowledge required.
Pint of science: Can AI save our planet? The Alan Turing Institute 2024-01-16 | Can AI save the planet? - Evangeline Corcoran The climate crisis is already having profound effects on our everyday lives. From longer droughts to more severe storms, the future of our planet is uncertain. Finding solutions requires reliable and effective methods for measuring and predicting how animals, plants, and ecosystems will respond to future conditions.AI may provide hope, with the promise of more accurate and rapid methods to monitor and predict environmental responses to possible futures.But the question is, how best to apply these methods to make sense of our ever-changing, unpredictable natural world?Pint of Science is a worldwide science festival which brings researchers to your local pub to share their scientific discoveries with you - no prior knowledge required.
Series wrap up: Listeners questions answered and the stories that didn’t make the cut The Alan Turing Institute 2023-12-27 | In this, the final episode of this season of tldr, Smera and Jonah delve into their post bag and attempt answer some listeners’ questions. Also, Jonah asks Smera about some of the stories that didn't make the cut for this series. With new developments in artificial intelligence happening at lightning speed, it can be hard to keep up. That's where too long didn't read comes in. We know you're busy, so we'll get right to the main points on how breakthroughs in machine learning, neural networks, and other AI systems could impact your life. No AI degree required! We're not here for hype or fear mongering. Just balanced coverage to keep you up-to-date. Think of us as your shortcut to AI literacy. Join us as we read between the lines on artificial intelligence every week right here on too long didn't read. Episode Sources Mike Wooldridge Turing Lectureyoutu.be/2kSl0xkq2lM?si=tzAi0sAqftQICLdq Julia Shaw on Life Scientific bbc.co.uk/programmes/m0014p73 Speaker knows you’re drunk thetimes.co.uk/article/smart-tech-can-use-tongue-twisters-to-tell-if-youre-drunk-ttf6203k7 AI Headteacher interestingengineering.com/innovation/abigail-bailey-schools-ai-principle-headteacher Less sperm news-medical.net/news/20231102/Study-links-heavy-mobile-phone-use-to-lower-sperm-count.aspx#:~:text=Men%20using%20mobile%20phones%20at,in%20this%20group%20of%20men.
The Turing Lectures: The future of generative AI The Alan Turing Institute 2023-12-21 | With their ability to generate human-like language and complete a variety of tasks, generative AI has the potential to revolutionise the way we communicate, learn and work. But what other doors will this technology open for us, and how can we harness it to make great leaps in technology innovation? Have we finally done it? Have we cracked AI?Join Professor Michael Wooldridge for a fascinating discussion on the possibilities and challenges of generative AI models, and their potential impact on societies of the future. Michael Wooldridge is Director of Foundational AI Research and Turing AI World-Leading Researcher Fellow at The Alan Turing Institute. His work focuses on multi-agent systems and developing techniques for understanding the dynamics of multi-agent systems. His research draws on ideas from game theory, logic, computational complexity, and agent-based modelling. He has been an AI researcher for more than 30 years and has published over 400 scientific articles on the subject.This lecture is part of a series of events - How AI broke the internet - that explores the various angles of large-language models and generative AI in the public eye.This series of Turing Lectures is organised in collaboration with The Royal Institution of Great Britain.
How Santa could implement AI and the career journey of our AI professional The Alan Turing Institute 2023-12-20 | In this special festive episode Smera and Jonah are recruited by Santa and his elves to discuss how AI could increase productivity for the festive friends at this busy time of year. Jonah also asks Smera all about her professional journey that led to her becoming a researcher in Data Justice and Global Ethical Futures. With new developments in artificial intelligence happening at lightning speed, it can be hard to keep up. That's where too long didn't read comes in. We know you're busy, so we'll get right to the main points on how breakthroughs in machine learning, neural networks, and other AI systems could impact your life. No AI degree required! We're not here for hype or fear mongering. Just balanced coverage to keep you up-to-date. Think of us as your shortcut to AI literacy. Join us as we read between the lines on artificial intelligence every week right here on too long didn't read. Episode Sourceshttps://www.telstra.com.au/exchange/hello-christmas nytimes.com/2023/12/05/technology/calm-jimmy-stewart-ai.htmlbbc.com/future/article/20181219-what-to-do-with-your-leftover-christmas-food
AI gods, child protection and Britcoin The Alan Turing Institute 2023-12-15 | In this week’s episode, Jessie stands in for Jonah and alongside Smera, the pair explore new AI-worshipping religions, the ever-growing concerns around children and online pornography, and the emergence of a new kind of digital currency. With new developments in artificial intelligence happening at lightning speed, it can be hard to keep up. That's where too long didn't read comes in. We know you're busy, so we'll get right to the main points on how breakthroughs in machine learning, neural networks, and other AI systems could impact your life. No AI degree required! We're not here for hype or fear mongering. Just balanced coverage to keep you up-to-date. Think of us as your shortcut to AI literacy.Join us as we read between the lines on artificial intelligence every week right here on too long didn't read. Episode sourcesReligion Gods in the machine? The rise of artificial intelligence may result in new religions (theconversation.com) (theconversation.com/gods-in-the-machine-the-rise-of-artificial-intelligence-may-result-in-new-religions-201068) Neil McArthur, AI Worship as a New Form of Religion - PhilPapers (philpapers.org/rec/MCAAWA) ‘Way of the Future’: Former Google engineer Anthony Levandowski relaunches AI church - Technology News (wionews.com) (wionews.com/technology/way-of-the-future-former-google-engineer-anthony-levandowski-relaunches-ai-church-663284) Inside Artificial Intelligence's First Church | WIRED (wired.com/story/anthony-levandowski-artificial-intelligence-religion) europeanacademyofreligionandsociety.com/news/the-way-of-the-future-is-now-a-thing-of-the-past/ Porn UK porn watchers could have faces scanned - BBC News (bbc.co.uk/news/technology-67615719) AI Fake Nudes Are Out of Control. These Victims Are Fighting Back - Bloomberg (bloomberg.com/news/features/2023-11-29/deepfake-porn-victims-learn-us-has-no-federal-laws-to-fight-it) arxiv.org/abs/2110.01963 Turing Lecture: Regulating Unreality (Deepfakes, revenge-pornography & fake news) - YouTube (youtube.com/watch?v=qfZKbffgvLQ&t=1s) Growing up with pornography: advice for parents and schools | Children's Commissioner for England (childrenscommissioner.gov.uk) (childrenscommissioner.gov.uk/blog/growing-up-with-pornography-advice-for-parents-and-schools) Implementing the Online Safety Act: Protecting children from online pornography - Ofcom (ofcom.org.uk/news-centre/2023/implementing-the-online-safety-act-protecting-children) Finance Digital pound plans should proceed with caution, say MPs - BBC News (bbc.co.uk/news/technology-67590468) The digital pound | Bank of England (bankofengland.co.uk/the-digital-pound) cfr.org/backgrounder/cryptocurrencies-digital-dollars-and-future-money omfif.org/2023/04/are-cbdcs-a-ticking-timebomb-for-commercial-banks/ atlanticcouncil.org/cbdctracker/ youtu.be/fpb-qJv6dBs?si=uW-FpfJ24ie6gNqX youtu.be/8u8qZT33DX4?si=GO2dtgLD98Yimz_9 youtu.be/fpb-qJv6dBsGood news (for some) dailymail.co.uk/sciencetech/article-12716977/ai-chatgpt-predicts-world-vegan-gen-z-millennials.html
Internet shutdowns, Pixel perfect art analysis and the AI will see you now The Alan Turing Institute 2023-12-07 | In this week’s episode, Jonah and Smera tackle internet shutdowns, drawing from recent global events to understand the wide-reaching impacts of disconnectivity; Jonah gets philosophical (again) discussing the latest technology being used to analyse art and Smera warns of the dangers of trusting a chatbot with our mental health. This show comes with a trigger warning, as the pair discuss mental health and what treatment could look like in an AI-driven society. With new developments in artificial intelligence happening at lightning speed, it can be hard to keep up. That's where too long didn't read comes in. We know you're busy, so we'll get right to the main points on how breakthroughs in machine learning, neural networks, and other AI systems could impact your life. No AI degree required! We're not here for hype or fear mongering. Just balanced coverage to keep you up-to-date. Think of us as your shortcut to AI literacy.Join us as we read between the lines on artificial intelligence every week right here on too long didn't read. Email us podcast@turing.ac.uk (mailto:podcast@turing.ac.uk) and find us on Instagram @theturinginst Episode sources STORY 1 - Internet Shutdowns ifj.org/media-centre/news/detail/category/press-releases/article/israel-government-to-shut-down-critical-media-alleged-to-undermine-national-security google.com/url?sa=t&rct=j&q=&esrc=s&source=newssearch&cd=&cad=rja&uact=8&ved=2ahUKEwiJm_fp-8WCAxUMQkEAHfNKAIIQxfQBKAB6BAgPEAE&url=https%3A%2F%2Fwww.wired.com%2Fstory%2Fisrael-gaza-internet-blackouts-weapon%2F&usg=AOvVaw1G8PO_rwUgVvGrB86RmxEE&opi=89978449 theguardian.com/global-development/2023/feb/28/internet-shutdowns-record-number-countries-2022-report google.com/url?sa=t&rct=j&q=&esrc=s&source=newssearch&cd=&cad=rja&uact=8&ved=2ahUKEwiop_HQ-8WCAxWBVUEAHYAHCfkQxfQBKAB6BAgKEAE&url=https%3A%2F%2Fwww.article19.org%2Fresources%2Ftechtonic-internet-shutdowns-as-a-tool-of-control%2F&usg=AOvVaw1IO1xc5_ye6PC5Y4cPzdNp&opi=89978449 accessnow.org/internet-shutdowns-2022 chat.openai.com/share/71bd3d50-086c-475c-8818-fb51fcc6f047 technologyreview.com/2021/09/09/1035237/internet-shutdowns-censorship-exponential-jigsaw-google/ amnestyusa.org/updates/is-internet-access-a-human-right/ ohchr.org/en/instruments-mechanisms/instruments/international-covenant-civil-and-political-rights theguardian.com/technology/2023/oct/18/instagram-palestine-posts-censorship-accusations accessnow.org/press-release/how-israel-is-shutting-down-the-internet-in-gaza/#:~:text=Access%20Now's%20new%20report%2C%20Palestine,to%20information%20is%20most%20vital. hrw.org/report/2023/06/14/no-internet-means-no-work-no-pay-no-food/internet-shutdowns-deny-access-basic techwontsave.us/episode/194_the_information_war_in_gaza_w_marwa_fatafta Story 2 ART BIMM independent.co.uk/news/uk/sagaftra-david-jones-space-writers-guild-of-america-hollywood-b2455985.html bimm.ac.uk/ai-labs/ soundspheremag.com/news/bimm-university-launch-ai_labs-new-course-preparing-students-for-ai-in-the-creative-industries/ SCROLL theguardian.com/science/2023/oct/12/researchers-use-ai-to-read-word-on-ancient-scroll-burned-by-vesuvius BRUSH STROKES nature.com/articles/d41586-023-03604-3 The Book of Trespass theguardian.com/books/2020/aug/10/the-book-of-trespass-by-nick-hayes-review-a-trespassers-radical-manifesto INDUS CIVILIZATIONworldhistory.org/india/ restofworld.org/2022/indus-translation-ai-code-scriptwired.com/2009/04/indusscript Story 3 – Health Samartians Radar samaritans.org/about-samaritans/research-policy/internet-suicide/samaritans-radar/#:~:text=This%20feedback%20included%20concerns%20about,and%20seeking%20support%20and%20community. cambridge.org/core/journals/european-psychiatry/article/artificial-intelligence-and-suicide-prevention-a-systematic-review/4AEF310A4924FCED128DEEBA63E349F9 BACP warning dailymail.co.uk/health/article-12765337/Fears-drive-replace-mental-health-counsellors-AI-clear-waiting-list-MILLION-patients.html rcpsych.ac.uk/news-and-features/latest-news/detail/2022/10/10/hidden-waits-force-more-than-three-quarters-of-mental-health-patients-to-seek-help-from-emergency-services AI Therapists psychologytoday.com/ca/blog/urban-survival/202301/are-ai-chatbots-the-therapists-of-the-future?amp theconversation.com/move-over-agony-aunt-study-finds-chatgpt-gives-better-advice-than-professional-columnists-214274 NHS stats https://nhsproviders.org/news-blogs/news/government-urged-to-get-a-grip-on-youngsters-mental-h...
In-effective altruism, a brief history of cyber security and can AI read minds? The Alan Turing Institute 2023-11-29 | In this week's episode, Jonah and Smera talk celebs, CEO's and Open AI's rollercoaster week, the pair look back at some of recent history's worst cybercrime attacks and we learn about the incredible potential of ChatGPT in neuroscience. Plus, the cutting-edge AI that can predict heart attacks up to ten years before they happen. With new developments in artificial intelligence happening at lightning speed, it can be hard to keep up. That's where too long didn't read comes in. We know you're busy, so we'll get right to the main points on how breakthroughs in machine learning, neural networks, and other AI systems could impact your life. No AI degree required! We're not here for hype or fear mongering. Just balanced coverage to keep you up-to-date. Think of us as your shortcut to AI literacy.Join us as we read between the lines on artificial intelligence every week right here on too long didn't read. Episode sources AI Business vox.com/future-perfect/2023/11/21/23971765/openai-sam-altman-microsoft theguardian.com/news/audio/2023/nov/24/sam-altman-chaos-heart-ai-industry-podcast diginomica.com/openais-meltdown-prompts-further-questions-around-future-ai-safety-surveillance bbc.co.uk/news/business-67494165 Cyber security theguardian.com/technology/2023/nov/22/personal-data-stolen-in-british-library-cyber-attack-appears-for-sale-online British Library: Employee data leaked in cyber attack - BBC News (bbc.co.uk/news/entertainment-arts-67484639) researchbriefings.files.parliament.uk/documents/CBP-9184/CBP-9184.pdf gov.uk/government/publications/defence-artificial-intelligence-strategy/defence-artificial-intelligence-strategy - Explain Estonia 2007 (bbc.co.uk/news/39655415) bbc.co.uk/news/world-middle-east-56722181 washingtonpost.com/world/2021/04/12/faq-natanz-nuclear-site-attack-israel/ theguardian.com/world/2021/apr/12/iran-blames-israel-attack-natanz-nuclear-plant talwork.net/wannacry cetas.turing.ac.uk/ Media literacy informationgov.uk/guidance/online-media-literacy-resources ofcom.org.uk/news-centre/2023/ofcom-supports-organisations-boosting-online-literacy-skills-in-local-communities Mind reading edition.cnn.com/2023/05/23/tech/chatgpt-mind-reading/index.html thephilosophyforum.com/discussion/14781/mind-blowing-mind-reading-technology medium.com/@affiliatemoneymaster2023/meta-just-achieved-mind-reading-with-ai-a-breakthrough-in-brain-computer-interface-technology-98cf7deb6858 theguardian.com/technology/2023/may/01/ai-makes-non-invasive-mind-reading-possible-by-turning-thoughts-into-text wired.com/story/ai-thought-decoder-mind-philosophy/ Heart attack predictions theguardian.com/society/2023/nov/13/ai-could-predict-heart-attack-risk-up-to-10-years-in-the-future-finds-oxford-study
Nathan Kutz - The Dynamic Mode Decomposition - A Data-Driven Algorithm The Alan Turing Institute 2023-11-24 | Full title - The Dynamic Mode Decomposition - A Data-Driven Algorithm for the Analysis of Complex SystemsThe dynamic mode decomposition (DMD) is a powerful data-driven modeling technique that reveals coherent spatiotemporal structures and produces reconstructions and future-state predictions from data. The method's simple linear algebra-based formulation additionally allows for a variety of optimizations and extensions that make the algorithm more practical and viable for analyzing real-world data sets. As a result, DMD has grown to become a leading method for equation-free system analysis across multiple scientific disciplines. PyDMD is a Python package that implements DMD and several of its major variants. In this talk, I will go over the underlying DMD theory and show the use of the PyDMD package which is specifically designed to handle dynamics that are noisy, multiscale, parameterized, prohibitively high-dimensional, or even strongly nonlinear. I will additionally provide a complete overview of the features currently available in PyDMD, along with a brief overview of the theory behind the DMD algorithm, tips regarding practical DMD usage, information for developers, and coding examples.
Cristian Axenie - Physics-informed Machine Learning for Robust Pedestrian Detection The Alan Turing Institute 2023-11-24 | Full title - Physics-informed Machine Learning for Robust Pedestrian Detection in Embedded Low-power Systems.For urban and metropolitan areas, population growth is expected to continue in the coming years. Cities and municipalities will continue to strive to provide residents with a sense of security and freedom. Camera-based surveillance systems are currently the most mature solution for urban surveillance, for example in monitoring road traffic and, or course, detecting abnormal pedestrian motion and activity at road side. However, most existing and established algorithms and systems for urban surveillance require expensive and power-hungry computer hardware. Neuromorphic sensors and computing systems are a promising alternative. The key advantages of such systems are high energy efficiency, fast and representative (relevant) data acquisition, fast local processing, improved data/identity protection, and rational budgeted use of resources. At town-scale this can provide high benefits. In this talk, we will focus on a novel and efficient real-time machine learning algorithm infused with physical motion models that is able to accurately detect and track pedestrians and bikers both day and night in urban scenarios.
Chris Rackauckas - Generalizing Scientific Machine Learning and Differentiable Simulation The Alan Turing Institute 2023-11-24 | Full Title - Generalizing Scientific Machine Learning and Differentiable Simulation Beyond Continuous ModelsThe combination of scientific models into deep learning structures, commonly referred to as scientific machine learning (SciML), has made great strides in the last few years in incorporating models such as ODEs and PDEs into deep learning through differentiable simulation. However, the vast space of scientific simulation also includes models like jump diffusions, agent-based models, and more. Is SciML constrained to the simple continuous cases or is there a way to generalize to more advanced model forms? This talk will dive into the mathematical aspects of generalizing differentiable simulation to discuss cases like chaotic simulations, differentiating stochastic simulations like particle filters and agent-based models, and solving inverse problems of Bayesian inverse problems (i.e. differentiation of Markov Chain Monte Carlo methods). We will then discuss the evolving numerical stability issues, implementation issues, and other interesting mathematical tidbits that are coming to light as these differentiable programming capabilities are being adopted.
Will Plutocrats be the first to mine in space? Not really artificial, not especially intelligent ... The Alan Turing Institute 2023-11-22 | This week Smera and Jonah go outer space to discuss mineral sourcing, they explore how the terms we use to describe machine learning, need to be properly understood and they ask if AI in agriculture is a muddy step in the right direction. Too long didn't read, brought to you by The Alan Turing Institute, the UK's national institute for data science and AI.The weekly podcast that reads the week's big AI stories so you don't have to. With new developments in artificial intelligence happening at lightning speed, it can be hard to keep up. That's where too long didn't read comes in. We know you're busy, so we'll get right to the main points on how breakthroughs in machine learning, neural networks, and other AI systems could impact your life. No AI degree required! We're not here for hype or fear mongering. Just balanced coverage to keep you up-to-date. Think of us as your shortcut to AI literacy.Join us as we read between the lines on artificial intelligence every week right here on too long didn't read. Episode sourcesSpace AI in space – how Artificial Intelligence is now helping us find untapped natural resources from outer space (cryptoslate.com/ai-in-space-how-artificial-intelligence-is-now-helping-us-find-untapped-natural-resources-from-outer-space) - Nov 14th 2023 First supernova detected, confirmed, classified and shared by AI (phys.org/news/2023-10-supernova-ai.amp) – Oct 13th 2023 GenMat (https://www.genmat.xyz/) and Propriety AI –based physics platform (globenewswire.com/news-release/2023/07/26/2711475/0/en/Research-Partnership-Extends-Capabilities-Of-Proprietary-AI-based-Physics-Platform.html) HRSI (uk.finance.yahoo.com/news/genmat-launches-pioneering-space-based-110000236.html) and Hyperspectral imaging (sciencedirect.com/topics/medicine-and-dentistry/hyperspectral-imaging#:~:text=Hyperspectral%20imaging%20(HSI)%20is%20a,information%20on%20what%20is%20imaged.) Moon Treaty (unoosa.org/oosa/en/ourwork/spacelaw/treaties/intromoon-agreement.html) How many exoplanets discovered by Machine Learning (https://newsroom.usra.edu/discovery-of-69-new-exoplanets-using-machine-learning/%22%20/l%20%22:~:text=The%20discovery%20of%2069%20new,our%20place%20in%20the%20cosmos.) A lot of exoplanets, like, all of the time (https://exoplanetarchive.ipac.caltech.edu/cgi-bin/TblView/nph-tblView?app=ExoTbls&config=TD) 300 in 2021 (nasa.gov/missions/kepler/new-deep-learning-method-adds-301-planets-to-keplers-total-count) by ExoMiner Astronomers are using AI to discover fledgling planets (astronomy.com/science/astronomers-are-using-ai-to-discover-fledgling-planets) Mining the moon and outer space (rand.org/pubs/commentary/2022/11/governance-in-space-mining-the-moon-and-beyond.html) AGI and definitions Define AI (technologyreview.com/2023/11/16/1083498/google-deepmind-what-is-artificial-general-intelligence-agi/?truid=&utm_source=the_download&utm_medium=email&utm_campaign=the_download.unpaid.engagement&utm_term=&utm_content=11-17-2023&mc_cid=baf487d12f&mc_eid=5aab967041) Hallucinate 40% of new AI companies don't use AI at all (technologyreview.com/2019/03/05/65990/about-40-of-europes-ai-companies-dont-actually-use-any-ai-at-all) Difference between AI and machine learning (zdnet.com/article/dont-be-alarmed-but-youre-probably-using-the-term-ai-wrong) Dangerous to think of AI as more than a tool (newyorker.com/science/annals-of-artificial-intelligence/there-is-no-ai) New taxonomy of AGI (arxiv.org/pdf/2311.02462.pdf) AGI is here (noemamag.com/artificial-general-intelligence-is-already-here) AGI definition (adalovelaceinstitute.org/resource/foundation-models-explainer) Farming theverge.com/2023/11/14/23950666/ai-sustainable-farming-machine-learning-agriculture zordi.com/ microsoft.com/en-us/research/project/farmbeats-iot-agriculture/?ranMID=24542&ranEAID=nOD/rLJHOac&ranSiteID=nOD_rLJHOac-HRyi6dSeSC_Ru0Ivu9iQOA&epi=nOD_rLJHOac-HRyi6dSeSC_Ru0Ivu9iQOA&irgwc=1&OCID=AIDcmm549zy227_aff_7593_1243925&tduid=%28ir__ebmjgu2dmkkfdmmfqnxkjvbek32xbqwc3q0r6qyd00%29%287593%29%281243925%29%28nOD_rLJHOac-HRyi6dSeSC_Ru0Ivu9iQOA%29%28%29&irclickid=_ebmjgu2dmkkfdmmfqnxkjvbek32xbqwc3q0r6qyd00 Turing underground farms (turing.ac.uk/about-us/impact/optimising-worlds-first-underground-farm)
too long didnt read - A weekly podcast from The Alan Turing Institute The Alan Turing Institute 2023-11-16 | Too long didn’t read, brought to you by The Alan Turing Institute, the national institute for data science and AI. The weekly podcast that reads the week’s big AI stories so you don’t have to. With new developments in artificial intelligence happening at lightning speed, it can be hard to keep up. That’s where too long didn’t read comes in. We know you’re busy, so we’ll get right to the main points on how breakthroughs in machine learning, neural networks, and other AI systems could impact your life. No AI degree required! We’re not here for hype or fear mongering. Just balanced coverage to keep you up-to-date. Think of us as your shortcut to AI literacy. Join us as we read between the lines on artificial intelligence every week right here on too long didn’t read.Listen wherever you get your podcasts, including here on YouTube
Hot Chips, Exploitation in AI and ’Is it a bird, is it a plane? No it’s Supercomputer!’ The Alan Turing Institute 2023-11-16 | This week Smera explains to Jonah some of the intricacies of the global 'Chip War'. They explore claims of labour exploitation in the name of AI safety, and discuss how many whales you can fit inside a supercomputer.Smera tells us how deepfakes are being used for good during Diwali and the team enjoy a brand new DJ GEF inspired remix of the intro music. Too long didn't read, brought to you by The Alan Turing Institute, the UK's national institute for data science and AI.The weekly podcast that reads the week's big AI stories so you don't have to. With new developments in artificial intelligence happening at lightning speed, it can be hard to keep up. That's where too long didn't read comes in. We know you're busy, so we'll get right to the main points on how breakthroughs in machine learning, neural networks, and other AI systems could impact your life. No AI degree required! We're not here for hype or fear mongering. Just balanced coverage to keep you up-to-date. Think of us as your shortcut to AI literacy.Join us as we read between the lines on artificial intelligence every week right here on too long didn't read. Episode sources SOURCES CHIPS US orders immediate halt to some AI chip exports to China, Nvidia says - BBC News (bbc.co.uk/news/business-67213134) Nvidia and iPhone maker Foxconn tie-up to build ‘AI factories’ (ft.com) (ft.com/content/56e99b7c-efc0-49c4-8d58-c0702183e791) (from last week) US-China chip war: Beijing unhappy at latest wave of US restrictions - BBC News (bbc.co.uk/news/business-67141987) (from last week) Exclusive: ChatGPT-owner OpenAI is exploring making its own AI chips | Reuters (reuters.com/technology/chatgpt-owner-openai-is-exploring-making-its-own-ai-chips-sources-2023-10-06) Baidu order AI chips from Huawei (twitter.com/Reuters/status/1722136869910270451?utm_source=theaibreakdown.beehiiv.com&utm_medium=newsletter&utm_campaign=the-ai-phone-wars-heat-up-with-samsung-gauss) EXPLOITATION Mhairi Aitken’s Turing Lecture (youtube.com/watch?v=UYdx74st9O4) OpenAI Used Kenyan Workers on Less Than $2 Per Hour: Exclusive | Time (time.com/6247678/openai-chatgpt-kenya-workers) - Jan 23 The Exploited Labor Behind Artificial Intelligence (noemamag.com) (noemamag.com/the-exploited-labor-behind-artificial-intelligence) - Oct 22 restofworld.org/2021/refugees-machine-learning-big-tech/ theverge.com/2019/3/28/18285572/prison-labor-finland-artificial-intelligence-data-tagging-vainu srinstitute.utoronto.ca/news/the-data-production-dispositif media.nature.com/original/magazine-assets/d41586-020-02003-2/d41586-020-02003-2.pdf Radical AI (radicalai.net/work) Alphabet Union (alphabetworkersunion.org) The trauma floor (theverge.com/2019/2/25/18229714/cognizant-facebook-content-moderator-interviews-trauma-working-conditions-arizona) - Feb 2019 SUPERCOMPUTER Bristol and Cambs to host supercomputers (computerweekly.com/news/366558093/Bristol-and-Cambridge-to-host-2024-AI-supercomputers) What is a supercomputer (bbc.co.uk/news/uk-england-bristol-67296381) Cooling of Isambard (techspark.co/blog/2023/11/02/isambard-ai-supercomputer) Exec order (whitehouse.gov/briefing-room/statements-releases/2023/10/30/fact-sheet-president-biden-issues-executive-order-on-safe-secure-and-trustworthy-artificial-intelligence) Positive news outlookindia.com/website/story/india-news-how-the-notjustacadburyad-campaign-us-using-ai-to-bring-joy-to-local-c/398614
too long didn’t read trailer The Alan Turing Institute 2023-11-16 | Too long didn't read, brought to you by The Alan Turing Institute, the UK's national institute for data science and AI.The weekly podcast that reads the week's big AI stories so you don't have to.
Living with Machines: On Collaboration The Alan Turing Institute 2023-11-10 | Living with Machines' is both a research project and a bold proposal for a new research paradigm. In this ground-breaking partnership between The Alan Turing Institute, the British Library and universities of Cambridge, East Anglia, Exeter, Queen Mary University of London and King’s College London, historians, data scientists, geographers, computational linguists, library professionals and curators were brought together to examine the human impact of industrial revolution.This series of short videos in documentary form seeks to make visible the collaborative underpinnings of the project by highlighting the team’s experiences, research objectives, challenges, and lessons learnt.Living with Machines was funded by UK Research and Innovations (UKRI), via the Strategic Priorities Fund and was administered by the Arts and Humanities Research Council (AHRC).Find out more here: livingwithmachines.ac.uk00:20 A new collaboration paradigm05:00 A Handbook on Collaboration07:00 We are Living with Machines today
The Turing Lectures: Addressing the risks of generative AI The Alan Turing Institute 2023-11-09 | As generative AI becomes increasingly sophisticated, its potential to revolutionise the way we interact with data is clear. It has already shown its ability to assist with tasks such as image and video synthesis, text and speech generation, and music composition. However, the rapid development of generative AI models has also raised concerns about their misuse, particularly in the context of disinformation campaigns, deepfakes and online harassment. Are generative AI models moving too fast for regulation to keep up? Will the development of generative AI outpace our ability to ensure their responsible use?In this lecture, Dr Mhairi Aitken will examine what this means for online and offline safety and discuss how society might be able to mitigate these risks. Mhairi Aitken is an Ethics Fellow in the Public Policy Programme at The Alan Turing Institute, and an Honorary Senior Fellow at Australian Centre for Health Engagement, Evidence and Values (ACHEEV) at the University of Wollongong in Australia. She is a Sociologist whose research examines social and ethical dimensions of digital innovation particularly relating to uses of data and AI. She was included in the 2023 international list of “100 Brilliant Women in AI Ethics”.This lecture is part of a series of events - How AI broke the internet - that explores the various angles of large-language models and generative AI in the public eye. This series of Turing Lectures is organised in collaboration with The Royal Institution.
The Summit of Safety, Classroom cheatbots and Ladybird subterfuge The Alan Turing Institute 2023-11-08 | This week Jonah asks Smera the real take-aways from the UK's AI Safety Summit have been, what are the risks, regs and ethics behind using large language models for homework and lesson planning and the technical secrets behind the new Beatles track. Too long didn't read, brought to you by The Alan Turing Institute, the UKs national institute for data science and AI.The weekly podcast that reads the week's big AI stories so you don't have to. With new developments in artificial intelligence happening at lightning speed, it can be hard to keep up. That's where too long didn't read comes in. We know you're busy, so we'll get right to the main points on how breakthroughs in machine learning, neural networks, and other AI systems could impact your life. No AI degree required! We're not here for hype or fear mongering. Just balanced coverage to keep you up-to-date. Think of us as your shortcut to AI literacy.Join us as we read between the lines on artificial intelligence every week right here on too long didn't read. Episode 4 sources AI Safety Summit: theguardian.com/technology/2023/nov/03/rishi-sunak-elon-musk-ai-summit-what-we-learned news.sky.com/story/joe-biden-to-unveil-sweeping-ai-regulations-days-before-skipping-rishi-sunaks-safety-summit-12996617 newscientist.com/article/2400834-what-did-the-uks-ai-safety-summit-actually-achieve/ adalovelaceinstitute.org/resource/foundation-models-explainer/ bbc.co.uk/news/technology-67172229 AI in academia and childrens rights - nature.com/articles/s42254-023-00581-4 turing.ac.uk/research/research-projects/exploring-childrens-rights-and-ai bbc.co.uk/news/education-67236732 theguardian.com/technology/2023/oct/31/educators-teachers-ai-learning-classrooms-misuse newscientist.com/article/mg26034613-400-lets-use-ai-to-rethink-education-instead-of-panicking-about-cheating/ The Beatles and chatbot relationships bbc.co.uk/news/entertainment-arts-65881813 bbc.co.uk/news/entertainment-arts-63226914 theguardian.com/music/2023/oct/26/the-beatles-final-song-now-and-then-ai-technology#:~:text=In%20June%2C%20McCartney%20told%20BBC,you%20some%20sort%20of%20leeway.%E2%80%9D thebeatles.com/today-learn-full-story-behind-now-and-then-watch-documentary-730pm-gmt-330pm-edt-1230pm-pdt washingtonpost.com/technology/2023/03/30/replika-ai-chatbot-update/ vice.com/en/article/y3py9j/ai-companion-replika-erotic-roleplay-updates makeuseof.com/how-does-replika-chatbot-work/ Optimistic about bbc.co.uk/news/health-67264350 Word of Year - bbc.co.uk/news/entertainment-arts-67271252
The Turing Lectures: What is generative AI? The Alan Turing Institute 2023-11-08 | Generative AI refers to a type of artificial intelligence that involves creating new and original data or content. Unlike traditional AI models that rely on large datasets and algorithms to classify or predict outcomes, generative AI models are designed to learn the underlying patterns and structure of the data and generate novel outputs that mimic human creativity. ChatGPT is perhaps the most well-known example, but the field is far larger and more varied than text generation. Other applications of generative AI include image and video synthesis, speech generation, music composition, and virtual reality.In this lecture, Professor Mirella Lapata will present an overview of this exciting—sometimes controversial—and rapidly evolving field. Mirella Lapata is professor of natural language processing in the School of Informatics at the University of Edinburgh. Her research focuses on getting computers to understand, reason with, and generate natural language. She is the first recipient (2009) of the British Computer Society and Information Retrieval Specialist Group (BCS/IRSG) Karen Sparck Jones award and a Fellow of the Royal Society of Edinburgh, the ACL, and Academia Europaea.This lecture is part of a series of events - How AI broke the internet - that explores the various angles of large-language models and generative AI in the public eye.This series of Turing Lectures is organised in collaboration with The Royal Institution
Who’s doing what with my face? AI Safety Summit and Lost in Translation The Alan Turing Institute 2023-11-01 | Joining Smera this episode is Luca to discuss Google's new controversial face-tuning feature, a discussion on the up-coming AI Safety Summit happening in the UK and how ChatGPT isn't quite ready to take over the entire world yet. Too long didn't read, brought to you by The Alan Turing Institute, the national institute for data science and AI.The weekly podcast that reads the week's big AI stories so you don't have to. With new developments in artificial intelligence happening at lightning speed, it can be hard to keep up. That's where too long didn't read comes in. We know you're busy, so we'll get right to the main points on how breakthroughs in machine learning, neural networks, and other AI systems could impact your life. No AI degree required! We're not here for hype or fear mongering. Just balanced coverage to keep you up-to-date. Think of us as your shortcut to AI literacy.Join us as we read between the lines on artificial intelligence every week right here on too long didn't read.
Mobilising for Data Justice The Alan Turing Institute 2023-10-26 | The path to data justice has many complex and systemic challenges. Yet individuals and communities around the world are paving a hopeful path.In this final episode of the Advancing Data Justice series, we hear from our global partners on how they mobilise for data justice through their transformative activism and advocacy. They shed light on the actions we need to take now to disrupt longstanding structures of inequity.Read more about the Advancing Data Justice: Research and Practice project here: turing.ac.uk/research/research-projects/advancing-data-justice-research-and-practiceTimestamps:0:00 Introduction01:36 How can we mobilise for data justice?03:07 Raising awareness05:58 Amplifying lived experiences09:26 Meaningful participation in decision-making processes12:16 Partner perspective: Digital Rights Foundation with HOPE – Project Adal17:43 Partner perspective: Digital Natives Academy – Māori Data Rangatiratanga25:30 Envisioning the future28:17 Credits
Dr eVTOL, Burnt Laps and Who’s got my face? The Alan Turing Institute 2023-10-24 | In this episode, Smera and Jonah talk flying cars, water usage in AI technologies and why you should think twice about engaging with face filter trends. Plus, more questionable dad-jokes from Jonah and we learn about Smera’s secret DJ backup plan. SourcesPredictive Policing: • propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing • zenodo.org/record/4050457#.YDP7k-j7TuT%20 • post.parliament.uk/ai-in-policing-and-security/ • propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing Clearview: • https://www.politico.eu/article/ai-ruling-obstruct-british-efforts-protect-citizens-images-us-data-harvesting/ • ico.org.uk/about-the-ico/media-centre/news-and-blogs/2022/05/ico-fines-facial-recognition-database-company-clearview-ai-inc/ • An AI firm harvested billions of photos without consent. Britain is powerless to act (https://www.politico.eu/article/ai-ruling-obstruct-british-efforts-protect-citizens-images-us-data-harvesting/) EVTOLS: • aviationtoday.com/2023/09/14/piloting-the-future-ai-evtols-and-sustainability/ • researchgate.net/profile/Jasenka-Rakas/publication/341074157_Urban_air_mobi[…]ban-air-mobility-and-manned-eVTOLs-safety-implications.pdf (researchgate.net/profile/Jasenka-Rakas/publication/341074157_Urban_air_mobility_and_manned_eVTOLs_safety_implications/links/5ebde82f92851c11a867ca06/Urban-air-mobility-and-manned-eVTOLs-safety-implications.pdf) • Piloting the Future: AI, eVTOLs, and Sustainability - Avionics International (aviationtoday.com/2023/09/14/piloting-the-future-ai-evtols-and-sustainability) Water consumption • forbes.com/sites/forbestechcouncil/2023/09/11/ai-liquid-cooling-and-the-data-center-of-the-future
Morten Mattrup Smedskjær - Cracking the Code The Alan Turing Institute 2023-10-20 | Full Title - Cracking the Code: Using Data Analysis to Engineer Stronger and Tougher GlassesGlasses are non-equilibrium materials that exhibit a glass transition and have a disordered (non-crystalline) structure. Glass materials can therefore be found across a variety of chemical compositions, from oxides to metal-organic frameworks. They feature heterogeneity and exhibit varying degree of structural disorder on different length scales, which has profound consequences for their properties and enables many applications. However, as known by its users, glass still breaks. Traditionally, new glass compositions have been developed through time-consuming trial-and-error experimentation. In this talk, I will discuss how we attempt to decipher their structure-property relations to accelerate the discovery process. Our work in this area combines experimental and simulation data with topological data analysis and various machine learning methods. I will also highlight how we use this knowledge to design more fracture-resistant glasses.
Disinformation, One Ring to Rule Them and Artificial IntelligDANCE The Alan Turing Institute 2023-10-18 | Too long didn't read, brought to you by The Alan Turing Institute, the national institute for data science and AI.The weekly podcast that reads the week's big AI stories so you don't have to.With new developments in artificial intelligence happening at lightning speed, it can be hard to keep up. That's where too long didn't read comes in. We know you're busy, so we'll get right to the main points on how breakthroughs in machine learning, neural networks, and other AI systems could impact your life. No AI degree required! We're not here for hype or fear mongering. Just balanced coverage to keep you up-to-date. Think of us as your shortcut to AI literacy. Join us as we read between the lines on artificial intelligence every week right here on too long didn't read.
Arnaud Vadeboncoeur - On Random Grid Neural Processes for Solving Forward and Inverse Problems The Alan Turing Institute 2023-10-12 | Full Title - On Random Grid Neural Processes for Solving Forward and Inverse Problems in Parametric PDEsWe introduce a new class of spatially stochastic physics and data informed deep latent models for parametric partial differential equations (PDEs) which operate through scalable variational neural processes. We achieve this by assigning probability measures to the spatial domain, which allows us to treat collocation grids probabilistically as random variables to be marginalised out. The implementation of these random grids poses a unique set of challenges for inverse physics informed deep learning frameworks and we propose a new architecture called Grid Invariant Convolutional Networks (GICNets) to overcome these challenges. We further show how to incorporate noisy data in a principled manner into our physics informed model to improve predictions for problems where data may be available but whose measurement location does not coincide with any fixed mesh or grid. We test our method on a nonlinear Poisson problem, Burgers, and Navier-Stokes equations.
Maldon Goodridge - A rare event study of frequency regulation and contingency services... The Alan Turing Institute 2023-10-12 | Full Title - A rare event study of frequency regulation and contingency services from grid-scale batteriesWe perform a rare-event study on a simulated power system in which grid-scale batteries provide both regulation and emergency frequency control ancillary services. Using a model of random power disturbances at each bus, we employ the skipping sampler, a Markov Chain Monte Carlo algorithm for rare-event sampling, to build conditional distributions of the power disturbances leading to two kinds of instability: frequency excursions outside the normal operating band, and load shedding. Potential saturation in the benefits, and competition between the two services, are explored as the battery maximum power output increases.
Jonathan Smith - HypoSVI: Earthquake hypocentre inversion with Stein variational inference... The Alan Turing Institute 2023-10-12 | Full Title - Jonathan Smith - HypoSVI: Earthquake hypocentre inversion with Stein variational inference and physics informed neural networksPhysics informed deep learning techniques have revolutionised the ways we approach environmental and physical sciences. In this talk I outline the scientific development we have made in common themed topics of: Physics Informed Neural Network (PINNs) for the Eikonal Equation; Probabilistic inversion in PINNs – Earthquake Location application; and Neural Operators for In-Ice Navigation. Outlined below is more information for each of these topics, with associated publications: Physics Informed Neural Network (PINNs) for the Eikonal Equation – In this topic we outline the use of Physics Informed Neural Networks to solving the partial differential equation for the high-frequency approximation to the wave equation, the Eikonal Equation, defining the network as EikoNet. The Eikonal equation represents the quickest travel-time between a source location and receiver location for a given speed field. Both toy-problems and sub-surface geological models are used to demonstrate the validity of this method. The training of EikoNet is achieved in a non-gridded representation of the solution, with the additional feature of the network allowing higher-order differentiability for downstream applications as defined in the next topic. Probabilistic inversion in PINNs – Earthquake Location application – In this topic we leverage the travel-time model EikoNet and outline how the differentiability of PINNs allow new variational inference technique of ‘Stein Variational Gradient Descent’ to be leverage to supply the earthquake location posterior distribution from observational data in seconds, scaling linearly with observational number. Neural Operators for In-Ice Navigation – In this topic we outline our preliminary work into using a Neural Operator (NOs) to determine the optimal travel-time within different sea-ice conditions to minimise fuel usage and risk. This technique leverages the transferability of NOs to train and validate on gaussian random field speed maps, with corresponding optimal travel-time field, and test on satellite derived speed maps. Although this work is very initial, we outline the next stages in the project.
Alice Cicirello - Physics-enhancing machine learning strategies in applied solid mechanics... The Alan Turing Institute 2023-10-12 | Full Title - Physics-enhancing machine learning strategies in applied solid mechanics: recent advancesMachine Learning based techniques are widely used in applied solid mechanics for reducing computational costs, improving modelling and forecasting, and enabling efficient and accurate information extraction. However, we are often confronted with the challenges of (i) having access to a small volume of informative data, (ii) the need of embedding physics-based knowledge to enable generalization in the small/medium data cases, and (iii) to enforce physics-constraints with large data to ensure physically consistent predictions. Physics-informed Machine Learning (aka Scientific Machine Learning) poses an integration challenge that goes beyond data and physics-based models, and includes the quantification of uncertainty, interpretability and explainability of the results, and dealing with small heterogeneous, gappy and noisy data. This seminar will give an overview of the open challenges and of recent research work carried out within the Data, Vibration and Uncertainty Group for developing enhanced strategies in applied solid mechanics, with particular focus on structural health monitoring and friction force evaluation.
Arka Daw - Uncertainty Quantification with Physics-informed Machine Learning The Alan Turing Institute 2023-10-12 | As applications in deep learning (DL) continue to seep into critical scientific use-cases, the importance of performing uncertainty quantification (UQ) with deep learning has become more pressing than ever before. In scientific applications, it is also important to inform the learning of deep learning models with the knowledge of physics of the problem to produce physically consistent and generalized solutions. This is referred to as the emerging field of Physics-informed Machine Learning (PIML). We present two distinct schools of thought for performing uncertainty quantification with physics-informed machine learning: (1) Physics-informed Architecture (PIA), where the physics-knowledge is hard-encoded into the neural network architecture, thereby producing meaning uncertainty estimates, (2) Physics-informed learning (PIL), where the additional physical knowledge is enforced as a soft constraint. We will demonstrate a Physics-informed Architecture approach for a special case of enforcing monotonicity constraints in the context of lake temperature modeling. We would also present a more generic Physics-informed learning (PIL) approach paired with generative adversarial networks (PID-GAN) to perform uncertainty quantifications when the physics-knowledge is found in the form of closed form equations/partial differential equations. We also provide an extension of PID-GAN for applications where the physics-equations available to us are derived using simplistic assumptions of complex real-world phenomena, i.e., the available knowledge is imperfect.
Burigede Liu - Learning-based multiscale modelling: computing, data science... The Alan Turing Institute 2023-10-12 | Full Title - Learning-based multiscale modelling: computing, data science, and uncertainty quantificationThe macroscopic properties of materials that we observe and exploit in engineering application result from complex interactions between physics at multiple lengths and time scales: electronic, atomistic, defects, domains etc. Multiscale modeling seeks to understand these interactions by exploiting the inherent hierarchy where the behavior at a coarser scale regulates and averages the behavior at a finer scale. This requires the repeated solution of computationally expensive finer-scale models, and often a priori knowledge of those aspects of the finer-scale behavior that affect the coarser scale (order parameters, state variables, descriptors, etc.). This talk reviews a number of machine learning frameworks that can be used to address the challenges in multi-scale modeling. First, we demonstrate the use of Fourier neural operators (FNOs) to accelerate the solution of governing partial differential equations of fine-scale models. We then demonstrate the use of recurrent neural operators (RNOs) to bridge the scales that is capable of providing insights into the history dependence and the macroscopic internal variables that govern the overall response. We end the talk with a discussion on how one can quantify the propagation of uncertainties through the length scales.
Cristian Axenie - Fusing Biophysics and Machine Learning for Computational Oncology The Alan Turing Institute 2023-10-12 | For decades, researchers have used the concepts of rate of change and differential equations to model and forecast neoplastic processes. This expressive mathematical apparatus brought significant insights in oncology by describing the unregulated proliferation and host interactions of cancer cells, as well as their response to treatments. Now, these theories have been given a new life and found new applications. With the advent of routine cancer genome sequencing and the resulting abundance of data, oncology now builds an “arsenal” of new modelling and analysis tools. Models describing the governing physical laws of tumour–host–drug interactions can be now fused with biological data to make predictions about cancer progression. Our study joins the efforts of the mathematical and computational oncology community by introducing a novel biophysics-informed machine learning system for the extraction of disease dynamics in oncology. The system utilizes computational mechanisms such as competition, cooperation, and adaptation in neural networks to simultaneously learn statistics and the governing relations between multiple clinical data covariates. Targeting an easy adoption in clinical oncology, the solutions of our system reveal human-understandable properties and features hidden in the data. As our experiments demonstrate, our system can describe: 1) nonlinear conservation laws in cancer kinetics and growth curves, 2) symmetries in tumour’s phenotypic staging transitions, 3) the preoperative spatial tumour distribution, and up to the 3) nonlinear intracellular and extracellular pharmacokinetics of neoadjuvant therapies. The primary goal of our work is to enhance the mechanistic understanding of cancer dynamics by exploiting heterogeneous clinical data and machine learning.
Cristian Axenie - Combining Mechanistic Modelling, Nonlinear Control, and Neuronal Learning... The Alan Turing Institute 2023-10-12 | Full Title - Cristian Axenie - Combining Mechanistic Modelling, Nonlinear Control, and Neuronal Learning Algorithms for Road Traffic OptimisationFor decades, researchers have used the concepts of rate of change and differential equations to model and forecast neoplastic processes. This expressive mathematical apparatus brought significant insights in oncology by describing the unregulated proliferation and host interactions of cancer cells, as well as their response to treatments. Now, these theories have been given a new life and found new applications. With the advent of routine cancer genome sequencing and the resulting abundance of data, oncology now builds an “arsenal” of new modelling and analysis tools. Models describing the governing physical laws of tumour–host–drug interactions can be now fused with biological data to make predictions about cancer progression. Our study joins the efforts of the mathematical and computational oncology community by introducing a novel biophysics-informed machine learning system for the extraction of disease dynamics in oncology. The system utilizes computational mechanisms such as competition, cooperation, and adaptation in neural networks to simultaneously learn statistics and the governing relations between multiple clinical data covariates. Targeting an easy adoption in clinical oncology, the solutions of our system reveal human-understandable properties and features hidden in the data. As our experiments demonstrate, our system can describe: 1) nonlinear conservation laws in cancer kinetics and growth curves, 2) symmetries in tumour’s phenotypic staging transitions, 3) the preoperative spatial tumour distribution, and up to the 3) nonlinear intracellular and extracellular pharmacokinetics of neoadjuvant therapies. The primary goal of our work is to enhance the mechanistic understanding of cancer dynamics by exploiting heterogeneous clinical data and machine learning.
