I turned myself into an AI-generated deathbot - here's what I found If a loved-one died tomorrow, would you want to keep talking to them? Not through memories or saved messages, but through artificial intelligence - a chatbot that uses their texts, emails and voice notes, to reply in their tone and style. A growing number of technology companies now offer such services as part of the digital afterlife industry, which is worth more than £100bn, with some people using it as a way to deal with their grief. Cardiff University's Dr Jenny Kidd has led research on so-called deathbots, published in the Cambridge University Press journal Memory, Mind and Media, and described the results as both fascinating and unsettling. Attempts to communicate with the dead are not new.

Researchers concluded that LLMs were able to spot the structure of a pun but did not really get the joke. Researchers concluded that LLMs were able to spot the structure of a pun but did not really get the joke. Can't tech a joke: AI does not understand puns, study finds Researchers say results underline large language models' poor grasp of humour, empathy and cultural nuance Comedians who rely on clever wordplay and writers of witty headlines can rest a little easier, for the moment at least, research on AI suggests. Experts from universities in the UK and Italy have been investigating whether large language models (LLMs) understand puns - and found them wanting. The team from Cardiff University, in south Wales, and Ca' Foscari University of Venice concluded that LLMs were able to spot the structure of a pun but did not really get the joke.

Reliable uncertainty quantification is crucial for the trustworthiness of machine learning applications. Inductive Conformal Prediction (ICP) offers a distribution-free framework for generating prediction sets or intervals with user-specified confidence. However, standard ICP guarantees are marginal and typically require a fresh calibration set for each new prediction to maintain their validity. This paper addresses this practical limitation by demonstrating how e-conformal prediction, in conjunction with Hoeffding's inequality, can enable the repeated use of a single calibration set with a high probability of preserving the desired coverage. Through a case study on the CIFAR-10 dataset, we train a deep neural network and utilise a calibration set to estimate a Hoeffding correction. This correction allows us to apply a modified Markov's inequality, leading to the construction of prediction sets with quantifiable confidence. Our results illustrate the feasibility of maintaining provable performance in conformal prediction while enhancing its practicality by reducing the need for repeated calibration. The code for this work is publicly available.

A cutting-edge artificial intelligence (AI) system that can accurately predict the areas of an image where a person is most likely to look has been created by scientists at Cardiff University. Based on the mechanics of the human brain and its ability to distinguish between different parts of an image, the researchers say the novel system more accurately represents human vision than anything that has gone before. Applications of the new system range from robotics, multimedia communication and video surveillance to automated image editing and finding tumors in medical images. The Multimedia Computing Research Group at Cardiff University are now planning to test the system by helping radiologists to find lesions within medical images, with the overall goal of improving the speed, accuracy and sensitivity of medical diagnostics. The system has been presented in the journal Neurocomputing.

A predictive tool using artificial intelligence could provide hope for earlier diagnosis of type 1 diabetes in children across the UK, reducing the risk of potentially fatal diabetic ketoacidosis (DKA), early research presented at the Diabetes UK Professional Conference 2022 has revealed. Type 1 diabetes is a serious auto-immune condition that cannot yet be prevented, and the gradual destruction of insulin-making beta cells can start months or even years before being diagnosed. Symptoms usually start to appear much closer to diagnosis. Early diagnosis and awareness of the signs and symptoms of diabetes are crucial to ensure that both children and adults who develop it do not become critically ill. A quarter of children and young people (25%1) aren't diagnosed with type 1 diabetes until they are in DKA2, a life-threatening condition that can lead to coma or even death.

Scientists are using cutting-edge artificial intelligence to help extract complex information from large collections of museum specimens. A team from Cardiff University is using state-of-the-art techniques to automatically segment and capture information from museum specimens and perform important data quality improvement without the need of human input. They have been working with museums from across Europe, including the Natural History Museum, London, to refine and validate their new methods and contribute to the mammoth task of digitizing hundreds of millions of specimens. With more than 3 billion biological and geological specimens curated in natural history museums around the world, the digitization of museum specimens, in which physical information from a particular specimen is transformed into a digital format, has become an increasingly important task for museums as they adapt to an increasingly digital world. A treasure trove of digital information is invaluable for scientists trying to model the past, present and future of organisms and our planet, and could be key to tackling some of the biggest societal challenges our world faces today, from conserving biodiversity and tackling climate change to finding new ways to cope with emerging diseases like COVID-19.

A reliable early warning system to detect tsunamis could be a step closer thanks to research from Cardiff University. Researchers say their analysis of ocean soundwaves triggered by underwater earthquakes has enabled them to develop artificial intelligence (AI) that allow prediction of when a tsunami might occur. The results are published today in the journal Scientific Reports. It is hoped this technology could assist experts in gaining accurate real-time assessments of these geological events. Dr. Usama Kadri, from Cardiff University's School of Mathematics, said: "Tsunamis have a devastating impact on communities. Developing accurate methods to detect them quickly is key to saving lives. "Our findings show we are able to classify the type of earthquake and retrieve its main properties from acoustic signals, in near real time.

Volcanologists can predict when volcanos are going to erupt if they have a full detail of its eruptions. But for potentially apocalyptic supervolcanoes, such as the one bubbling under Yellowstone National Park, it's nearly impossible, given how varied their known eruptions have been, according to a new study. Researchers at Cardiff University noted there is not a'single model' that can help scientists understand how eruptions from supervolcanoes happen, making it difficult to understand when they might occur in the future. The researchers looked at geochemical and petrological evidence of 13 supereruptions that have happened over the past 2 million years, including the most recent one, Taupō volcano in New Zealand, which happened more than 24,000 years ago. Experts said there is not a'single model' that can help them understand how eruptions from supervolcanoes happen There was no'single, unified mode' that showed how each of the 13 played out, with some starting gradually over a period of weeks to months, while others exploded suddenly and violently. The researchers also found that the eruptions lasted for varying times, some as short as a period of days or weeks, while others lasted decades.

With the advancements in the field of artificial intelligence, many sectors have been in fear of losing human employees over this advanced technology. And with the rise of machines amid this crisis for business continuity, the fear has started looming in the journalism industry where media houses are publishing automated news for their publications. In fact, Bloomberg News, one of the leading media publishing houses, has claimed that the company has been using automated technology for publishing one-third of their news content on their platform. According to news reports: Editor-in-chief John Micklethwait of Bloomberg News stated in their company memo a few years back, "I think automation is crucial to the future of journalism in a much broader way than many of us realise. Bloomberg already uses automation for customised news and trending stories …" Also, in the recent news, Microsoft has announced laying off a considerable number of journalists from their MSN in order to replace them with artificial intelligence.

As artificial intelligence (AI) becomes more commonplace in newsrooms, Cardiff University will be introducing the technology into its syllabus to produce'industry-ready journalists'. As of September 2020, students on the MA International Journalism course will learn to use programmes like Dataminr, Chartbeat, TweetDeck and Google Analytics - the inclusion of Dataminr being a first for UK universities. It features on a new'emerging journalism' module where students will use AI-powered tools to identify and report on a breaking news story, as this is becoming standard practice in local newsrooms, like Reach plc. As well as these practical skills, the module also offers the theory of using this technology for problem-solving. He stressed that aspiring journalists need to come to terms with AI, or risk being left behind by its advancements.