Philippa Kaur

An AI tool developed by DeepMind can help historians restore ancient Greek texts with 72 per cent accuracy, and date inscriptions to within 30 years of their true age

Humans

9 March 2022

By Carissa Wong
The Celsus Library in the ancient city of Ephesus, TurkeyMazur Travel/Shutterstock
An artificial intelligence algorithm developed as part of a collaboration between historians and UK-based AI firm DeepMind can help restore ancient Greek texts with 72 per cent accuracy.
The AI can also predict where in the ancient Mediterranean world the texts were originally written with more than 70 per cent accuracy and date them to within a few decades of their agreed-upon date of creation. All of this marks an improvement upon an earlier version of the AI that could only restore ancient texts.
“Inscriptions provide evidence of the thought, language, society and history of past civilisations,” says Thea Sommerschield at Ca’ Foscari University of Venice in Italy. “But most surviving inscriptions have been damaged over the centuries, so their texts are now fragmentary or illegible. They may also have been moved or trafficked far from their original location.”Advertisement
Join us for a mind-blowing festival of ideas and experiences. New Scientist Live is going hybrid, with a live in-person event in Manchester, UK, that you can also enjoy from the comfort of your own home, from 12 to 14 March 2022. Find out more.
When recovering ancient texts, historians are usually interested in achieving three major goals: restoring the text, and working out exactly when and where it was written. To do this, they look for distinctive features and patterns in the style of writing and compare them to those of ancient texts that have already been found and dated.
“However, it’s really difficult for a human to harness all existing relevant data, and to discover underlying patterns every time,” says Sommerschield.
Sommerschield and her colleagues worked with researchers at DeepMind to get the machine-learning AI – called Ithaca after a Greek island that is famous for being the home of the legendary figure Odysseus – to carry out all three tasks.
To train Ithaca, the team used around 60,000 ancient Greek texts from across the Mediterranean that are already well-studied and known to have been written between 700 BC and AD 500. The team masked some of the characters in the texts and then compared Ithaca’s predictions for this “missing” text with the actual inscriptions.
Next, the team used a data set of nearly 8000 inscriptions – again, already well-studied and understood – to test Ithaca’s performance alone, or in combination with two ancient historians. On its own, Ithaca could restore texts with 62 per cent accuracy, while ancient historians alone restored text with around 25 per cent accuracy.
However, the most accurate reconstructions involved Ithaca and historians working together. When historians took Ithaca’s top 20 most likely reconstructions for a given text and used them to inform their own work, they could restore the text with an accuracy even greater than Ithaca alone.
“When historians used Ithaca, their performance on the text restoration task actually tripled, to 72 per cent,” says Sommerschield.

Ithaca could also predict where in the Mediterranean a text was written 71 per cent of the time and it could date the texts to within 30 years of their true date of creation, as previously established by historians.
“It is clear that the authors’ work is important and groundbreaking. The ‘ancient historian and Ithaca’ method produces startlingly significant improvements in outcomes over traditional human-only methods,” says Tom Elliott at New York University. However, further testing with more historians is needed and people will need training and technical support to use the tool, he adds.
The team says the feedback from historians so far has been positive.
“We hope that the way we’ve designed it, it’s going to be easy for an ancient historian to use, because they will just type in the text [to an online interface] and then they will get all these visualisations that they can use,” says Yannis Assael at DeepMind in the UK, and an author of the study.
Ithaca’s design should also make it easily applicable to any ancient language and any written medium, says Sommerschield.
Journal reference: Nature, DOI: https://doi.org/10.1038/s41586-022-04448-z

More on these topics: More

Scientific monitoring of the Chernobyl nuclear power plant is being affected by the Russian invasion of Ukraine

Humans

8 March 2022

By Matthew Sparkes
Damage to infrastructure at the Kharkiv Institute of Physics and Technology neutron sourceState Nuclear Regulatory Inspectorate of Ukraine
Scientists monitoring radiation levels at Chernobyl are unable to access their laboratories and instruments because Russian troops control the plant, warns a worker who escaped the facility when it was captured by Russian forces on 24 February. Other staff still running the working power plants on the site are reportedly being held in poor conditions without the chance to take breaks away from the facility to rest.
“We continue scientific monitoring as much as possible,” says the nuclear expert from the National Academy of Sciences of Ukraine, who asked to remain anonymous for security reasons. “This is very far from [the] usual volume [of testing] because [my colleagues] have no access to our labs and instruments in Chernobyl, but we do our best in monitoring important values, sometimes by indirect data.”
The scientist tells New Scientist that all of his team were able to escape the facility and leave the Chernobyl Exclusion Zone on the first day of the invasion. Despite this, some of those staff are now caught in areas of intense fighting.Advertisement
Join us for a mind-blowing festival of ideas and experiences. New Scientist Live is going hybrid, with a live in-person event in Manchester, UK, that you can also enjoy from the comfort of your own home, from 12 to 14 March 2022. Find out more.
He says his team advises staff at the working part of the nuclear power plant and adds that they remain in contact. But it has now been almost two weeks since Russia seized the plant and the International Atomic Energy Agency’s (IAEA) director general Rafael Mariano Grossi says that the 210 staff on site have still not been able to leave for rest, something he stressed is important for them to carry out their jobs safely.
The anonymous scientist says that workers at the plant are “heroes” for continuing to ensure nuclear and radiation safety under those conditions.
The IAEA has now listed a series of incidents at nuclear power plants that it says present a risk to safety, although there are no signs or evidence of radiation leaks.
“We cannot go on like this, there has to be clear understandings, clear commitments not to go anywhere near a nuclear facility when it comes to military operations,” said Grossi at a press conference yesterday.
On the first day of the invasion there were radiation spikes at the Chernobyl plant which the State Inspectorate for Nuclear Regulation of Ukraine put down to Russian military vehicles stirring up radioactive dust.
On 26 February, an electrical transformer at a radioactive waste disposal facility near Kharkiv was damaged, and the following day missiles hit the site of a similar facility in Kyiv. No radiation leaks were detected after these attacks.
On 4 March a fire was started by missiles targeted at the Zaporizhzhia nuclear power plant. The blaze was later put out, but reports suggest that firefighters initially came under fire from Russian forces.
A neutron generator at the Kharkiv Institute of Physics and Technology used for scientific research has also been destroyed by shelling, says Grossi, and there are also concerns about a lack of communication from staff at an oncology centre in Mariupol that has radioactive materials.

More on these topics: More

Goose bones from Stone Age China suggest the birds were being domesticated there 7000 years ago, which could mean they were domesticated before chickens

Humans

7 March 2022

By Michael Marshall
Chinese geese (Anser cygnoides f. domestica)blickwinkel/AGAMI/M. Guyt/Alamy
Geese may have been domesticated as early as 7000 years ago in what is now China, according to a study of preserved goose bones. That may make them the first bird to be domesticated, before chickens – although the timing of chicken domestication is uncertain.
The finding extends the history of goose domestication and potentially the history of domestic poultry as a whole, says Masaki Eda at Hokkaido University Museum in Sapporo, Japan.
Eda is part of a team that has excavated an archaeological site in east China called Tianluoshan, which was a Stone Age village between about 7000 and 5500 years ago. Its inhabitants “were basically hunter-gatherers”, says Eda, but they also grew rice in paddy fields.Advertisement
Join us for a mind-blowing festival of ideas and experiences. New Scientist Live is going hybrid, with a live in-person event in Manchester, UK, that you can also enjoy from the comfort of your own home, from 12 to 14 March 2022. Find out more.
The researchers have identified 232 goose bones at Tianluoshan and say there are multiple lines of evidence that some of the geese were at least partially domesticated.
Four of the bones belonged to immature geese that were less than 16 weeks old, with the youngest probably less than eight weeks old. This implies they must have hatched at Tianluoshan, says Eda, because they were too young to have flown in from elsewhere. However, no wild geese breed in the area today and it is unlikely they did so 7000 years ago, he says.
Some of the adult geese also seem to have been locally bred, based on the chemical make-up of their bones, which reflects the water they drank. These locally bred birds were all roughly the same size, indicating captive breeding. Finally, the researchers carbon-dated the bones and found that the locally bred geese lived about 7000 years ago.
Taken together, the findings suggest the geese were at an early stage of domestication, says Eda.

“It’s a major study in our understanding of poultry domestication,” says Ophélie Lebrasseur at the Centre for Anthropobiology and Genomics of Toulouse in France. “They’ve been very thorough.”
“The main thing that stood out for me is the fact they actually did radiocarbon dating on the bird bones,” says Julia Best at Cardiff University in the UK. This makes the dating much more reliable than if they had simply dated the surrounding sediment.
If geese were domesticated 7000 years ago, that would make them the first bird to be domesticated, says Eda. The other candidate is chickens, but there has been a dispute over when and where this first happened.
Chickens were probably domesticated from wild birds called red junglefowl, which live in southern Asia. However, genetics has complicated the story, revealing that domestic chickens subsequently interbred with other birds like the grey junglefowl.
A study published in 2014 reported that chickens were domesticated in northern China as early as 10,000 years ago, based on DNA from bones. However, it isn’t clear that red junglefowl ever lived that far north, says Lebrasseur. Furthermore, the bones weren’t directly dated and “a lot of the things they claimed were chickens were pheasants”, says Best. Firm evidence of domestic chickens only appears from around 5000 years ago, she says.
This implies geese were domesticated before chickens, says Lebrasseur. “With the evidence we currently have, I think it is true,” she says. But she adds that bird domestications are understudied compared with those of mammals like dogs and cows, so the story could well change as more evidence emerges.
It is very difficult to say why the geese were domesticated, says Eda. Meat, eggs, feathers and bone tools are all possibilities, and they may have been used in ritual ceremonies. “One of the things we see with chickens is they’re often held in high esteem when they’re first domesticated,” says Best.
Journal reference: PNAS, DOI: 10.1073/pnas.2117064119

More on these topics: More

In Dying Light 2, a variant of a virus has turned people into zombiesTechland
Dying Light 2 Stay Human
Techland
PC, PlayStation 4 and 5, Xbox One and Series X/SAdvertisement
IN APRIL 2020, soon after the UK entered its first lockdown, I reviewed the zombie-packed Resident Evil 3, describing it as noticeably “pre-pandemic fiction”. Two years on, the pandemic is still going, and I am still playing zombie games. This time, it is Dying Light 2 Stay Human, and it is interesting to look at the game as a work of post-pandemic (mid-pandemic?) fiction.
It is a sequel to the 2015 game Dying Light, which saw a viral outbreak in the fictional Middle Eastern city of Harran turn people into zombies. The end of the game promised a cure to the disease, but as the introduction of Dying Light 2 explains – and stop me if you have heard this before – a new variant of the virus emerged in 2021 and spread rapidly. The zombies took over and civilisation collapsed. Cheery stuff.
The game picks up the story in 2036, where you play as a survivor called Aiden Caldwell. After being bitten by a zombie, you enter one of the last remaining outposts of society, known only as the City. There, you discover that all of the other survivors are also infected, but use a variety of tools to avoid zombification – hence the “Stay Human” part of the game’s title.
Full zombies can’t survive in sunlight, so City folk have set up ultraviolet lamps to hold back the infection. One of your early goals in the game is to acquire a wristband that provides an alert when you need a top-up of UV. Owning one of these wristbands is a condition of living in the City, perhaps a nod to the various covid passes that have been implemented around the world.
“Aiden has expert parkour skills that allow him to scale buildings and dodge undesirable characters”
With a wristband secured, the game settles into a rhythm. By day, you are more or less safe from zombies outside (though not from roving bandits), although it is risky to enter derelict buildings, where the undead tend to gather. Then, at night, the zombies hit the streets, so it is tricky to get around outside, but easier to explore within. Dodging zombies has its rewards: you get bonus experience points, which you can use to upgrade your abilities, handy for venturing out at night and for surviving a zombie chase.
For reasons that are never properly explained, Aiden has expert parkour skills that allow him to scale buildings, jump across rooftops and generally dodge undesirable characters. In a strange game design decision, features that would usually be part of the basic move set in this kind of game (such as the ability to slide) require unlocking upgrades, so it takes a while to accumulate the full set of skills.
That is a shame, because this freedom of movement is probably the best thing about the game. I had great fun racing through the city, but beyond the obvious covid-19 links, the meat of the game is nothing you haven’t seen before. Everything boils down to: go here, get this thing, kill these zombies, repeat.
As you explore the city, you get the opportunity to claim various locations, such as a water tower, for one of three factions: the slightly fascist Peacekeepers, the anarchic Renegades or the ordinary survivors. You get to pick a side, and Techland, the game’s developer, goes big on the idea that which you choose matters to the (entirely forgettable) storyline. But two years into the pandemic, I was more inclined to stick with the ordinary survivors. It is hard not to sympathise with people who have lived through a world-altering disaster and are just trying their best to carry on existing.
Jacob also recommends… More

Despite AI’s impressive feats at driving cars and playing games, a new book by psychologist Gerd Gigerenzer argues that our brains have plenty to offer that AI will never match

Humans

2 March 2022

By Chen Ly

IN THE 1950s, Herbert Simon – a political scientist and one of the founders of AI – declared that, once a computer could beat the best chess player in the world, machines would have reached the pinnacle of human intelligence. Just a few decades later, in 1997, the chess-playing computer Deep Blue beat world champion Garry Kasparov.
It was an impressive feat, but according to Gerd Gigerenzer, a psychologist at the Max Planck Institute for Human Development in Berlin, human minds don’t need to worry just yet. In How to Stay Smart in a Smart World, he unpacks humanity’s complicated relationship with artificial intelligence and digital technology. In an age where self-driving cars have been let loose on the roads, smart homes can anticipate and cater for our every need and websites seem to know our preferences better than we do, people tend to “assume the near-omniscience of artificial intelligence”, says Gigerenzer. But, he argues, AIs aren’t as clever as you might think.
A 2015 study, for example, showed that even the smartest object-recognition system is easily fooled, confidently classifying meaningless patterns as objects with more than 99 per cent confidence. And at the 2017 UEFA Champions League final in Cardiff, UK, a face-recognition system matched the faces of 2470 football fans at the stadium and the city’s railway station to those of known criminals. This would have been useful but for 92 per cent of the matches turning out to be false alarms, despite the system being designed to be both more efficient and more reliable than humans.
There are good reasons why even the smartest systems fail, says Gigerenzer. Unlike chess, which has rules that are rigid and unchanging, the world of humans is squishy and inconsistent. In the face of real-world uncertainty, algorithms fall apart.
Here, we get to the crux of Gigerenzer’s main argument: technology, at least as we know it today, could never replace humans because there is no algorithm for common sense. Knowing, but not truly understanding, leaves AI in the dark about what is really important.
Obviously, technology can be, and often is, useful. The voice and face-recognition software on smartphones are largely convenient and the fact that YouTube seems to know what I want to watch saves the hassle of working it out for myself. Yet even if smart technology is mostly helpful, and is showing few signs of replacing us, Gigerenzer argues that we should still be aware of the dangers it can pose to our society.
“Knowing, but not truly understanding, leaves artificial intelligence in the dark about what is really important”
Digital technology has created an economy that trades on the exchange of personal data, which can be used against our best interests. Companies and political parties can purchase targeted adverts that subtly influence our online shopping choices and, even more nefariously, how we vote. “One might call this turn to an ad-based business model the ‘original sin’ of the internet,” writes Gigerenzer.
So, what can be done? Gigerenzer says that more transparency from tech firms and advertisers is vital. But technology users also need to change our relationship with it. Rather than treating technology with unflinching awe or suspicion, we must cultivate a healthy dose of scepticism, he says. In an age where we seem to accept the rise of social media addiction, regular privacy breaches and the spread of misinformation as unavoidable downsides of internet use – even when they cause significant harm to society – it is perhaps time we took stock and reconsidered.
Using personal anecdotes, cutting-edge research and cautionary real-world tales, Gigerenzer deftly explains the limits and dangers of technology and AI. Occasionally, he uses extreme examples for the sake of making a point, and in places he blurs the lines between digital technology, smart technology, algorithms and AI, which muddies the waters. Nevertheless, the overall message of Gigerenzer’s book still stands: in a world that increasingly relies on technology to make it function, human discernment is vital “to make the digital world a world we want to live in”.

More on these topics: More