Friday, May 27, 2022

PAINT BY NUMBERS

Robot artist paints a portrait of the Queen for the Platinum Jubilee

Queen's portrait
Ai-Da uses cameras in her eyes and computer memory to paint and draw images. (Picture: http://www.ai-darobot.com)

A portrait of the Queen painted by a humanoid robot artist has been unveiled just in time for the Platinum Jubilee.

The artwork, titled Algorithm Queen, was painted by Ai-Da — an artificial intelligence robot built in 2019 that creates drawings, painting, and sculptures.

To start with, Ai-Da uses cameras in her eyes and computer memory, before using a variety of unique algorithms to paint and draw images.

A robotic arm lets Ai-Da turn digital formations into physical drawings and paintings.

Algorithm Queen was layered and scaled to produce the final multi-dimensional portrait of the monarch.

The machine’s artistic process was designed to reflect the different aspects of technological change that have taken place during the Queen’s 70-year reign.

Queen portrait by AI humanoid robot
Video Player is loading.
Loaded8.52%
Current Time 0:00
Duration 0:46

Ai-Da — named after 18th century mathematician and scientist Ada Lovelace — is also able to converse using a specially designed language model.

‘I’d like to thank Her Majesty the Queen for her dedication, and for the service she gives to so many people,’ said the humanoid robot.

‘She is an outstanding, courageous woman who is utterly committed to public service,’

‘I think she’s an amazing human being, and I wish The Queen a very happy Platinum Jubilee,’

Queen's portrait
The artwork, titled Algorithm Queen, was painted by Ai-Da (Picture: http://www.ai-darobot.com)

Ai-Da was devised in Oxford by Aidan Meller, a specialist in modern and contemporary art, before being built in Cornwall by Engineered Arts and programmed internationally.

The robot’s capabilities were developed by PhD students and professors at the Universities of Oxford and Birmingham.

‘We are excited Ai-Da Robot has made history just in time for the Queen’s Jubilee,’ said Meller.

‘We are in unprecedented technological times, and so we are pleased we can take a moment to think about all that has changed during the Queen’s life,’

Algorithm Queen will be exhibited publicly in London later this year and revealed on the Ai-Da Robot artist website at 10 am on May 27.

 

Will AI text-to-image generators put illustrators out of a job?

Imagen, from Google, is the latest example of an AI seemingly able to produce high-quality images from a text prompt – but they aren't quite ready to replace human illustrators

TECHNOLOGY | ANALYSIS 26 May 2022

New Scientist Default Image

Examples of images created by Google’s Imagen AI

Imagen/Google

Tech firms are racing to create artificial intelligence algorithms that can produce high-quality images from text prompts, with the technology seeming to advance so quickly that some predict that human illustrators and stock photographers will soon be out of a job. In reality, limitations with these AI systems mean it will probably be a while before they can be used by the general public.

Text-to-image generators that use neural networks have made remarkable progress in recent years. The latest, Imagen from Google, comes hot on the heels of DALL-E 2, which was announced by OpenAI in April.

Both models use a neural network that is trained on a large number of examples to categorise how images relate to text descriptions. When given a new text description, the neural network repeatedly generates images, altering them until they most closely match the text based on what it has learned.

While the images presented by both firms are impressive, researchers have questioned whether the results are being cherry-picked to show the systems in the best light. “You need to present your best results,” says Hossein Malekmohamadi at De Montfort University in the UK.

One problem in judging these AI creations is that both firms have declined to release public demos that would allow researchers and others to put them through their paces. Part of the reason for this is a fear that the AI could be used to create misleading images, or simply that it could generate harmful results.

The models rely on data sets scraped from large, unmoderated portions of the internet, such as the LAION-400M data set, which Google says is known to contain “pornographic imagery, racist slurs, and harmful social stereotypes”. The researchers behind Imagen say that because they can’t guarantee it won’t inherit some of this problematic content, they can’t release it to the public.

OpenAI claims to be improving DALL-E 2’s “safety system” by “refining the text filters and tuning the automated detection & response system for content policy violations”, while Google is seeking to address the challenges by developing a “vocabulary of potential harms”. Neither firm was able to speak to New Scientist before publication of this article.

Unless these problems can be solved, it seems unlikely that big research teams like Google or OpenAI will offer their text-to-image systems for general use. It is possible that smaller teams could choose to release similar technology, but the sheer amount of computing power required to train these models on huge data sets tends to limit work on them to big players.

Despite this, the friendly competition between the big firms is likely to mean the technology continues to advance rapidly, as tools developed by one group can be incorporated into another’s future model. For example, diffusion models, where neural networks learn how to reverse the process of adding random pixels to an image in an effort to improve them, have shown promise in machine-learning models in the past year. Both DALL-E 2 and Imagen rely on diffusion models, after the technique proved effective in less-powerful models, such as OpenAI’s Glide image generator.

“For these types of algorithms, when you have a very strong competitor, it means that it helps you build your model better than those other ones,” says Malekmohamadi. “For example, Google has multiple teams working on the same type of [AI] platform.”



Read more: https://zephr.newscientist.com/article/2322056-will-ai-text-to-image-generators-put-illustrators-out-of-a-job/#ixzz7UTh2uhH8

America's gun culture - in seven charts


A school shooting in Uvalde, Texas, involving young children has reignited the national US debate about access to firearms. What does the data tell us about gun culture and its impact?

Firearms deaths are a fixture in American life.

There were 1.5 million of them between 1968 and 2017 - that's higher than the number of soldiers killed in every US conflict since the American War for Independence in 1775.

In 2020 alone, more than 45,000 Americans died at the end of a barrel of a gun, whether by homicide or suicide, more than any other year on record. The figure represents a 25% increase from five years prior, and a 43% increase from 2010.

But the issue is a highly political one, pitting gun control advocates against sectors of the population fiercely protective of their constitutionally-enshrined right to bear arms.

How many guns are there in the US?

While calculating the number of guns in private hands around the world is difficult, figures from the Small Arms Survey - a Swiss-based leading research project - estimate that there were 390 million guns in circulation in 2018.

The US ratio of 120.5 firearms per 100 residents, up from 88 per 100 in 2011, far surpasses that of other countries around the world.

More recent data also suggests that gun ownership grew significantly over the last several years. One study, published by the Annals of Internal Medicine in February, found that 7.5 million US adults - just under 3% of the population - became first new gun owners between January 2019 and April 2021.

This, in turn, exposed 11 million people to firearms in their homes, including 5 million children. About half of new gun owners in that time period were women, while 40% were either black or Hispanic.

A separate study, published by the American Academy of Pediatrics in 2021, linked a rise in gun ownership during the pandemic to higher rates of gun injuries among - and inflicted by - children.

How do US gun deaths break down?

According to the US Centers for Disease Control and Prevention (CDC), a total of 45,222 people died from gun-related injuries of all causes during 2020, the last year for which complete data is available.

And while mass shooting and gun murders generally garner more media attention, of the total, 54% - about 24,300 deaths - were suicides.

A 2016 study published in the American Journal of Public Health found there was a strong relationship between higher levels of gun ownership in a state and higher firearm suicide rates for both men and women.

Advocates for stricter gun laws in the United States often cite this statistic when pushing lawmakers to devote more resources to mental health and fewer to easing gun restrictions.

How do US gun killings compare with other countries?

In 2020, 43% of the deaths - amounting to 19,384 people - were homicides, according to data from the CDC. The figure represents a 34% increase from 2019, and a 75% increase over the course of the previous decade.

Nearly 53 people are killed each day by a firearm in the US, according to the data.

The data also shows that the vast majority of murders, 79%, were carried out with guns.

That's a significantly larger proportion of homicides than is the case in Canada, Australia, England and Wales, and many other countries.

Are mass shootings becoming deadlier?

Deaths from the "mass shootings" that attract international attention, however, are harder to track.

While the country does not have a single definition for "mass shootings", the FBI has for over a decade tracked "active shooter incidents" in which "an individual actively engaged in killing or attempting to kill people in a populated area".

According to the FBI, there were 345 "active shooter incidents" in the United States between 2000-2020, resulting in more than 1,024 deaths and 1,828 injuries.

The deadliest such attack, in Las Vegas in 2017, killed more than 50 people and left 500 wounded. The vast majority of mass shootings, however, leave fewer than 30 people dead.

Who supports gun control?

Despite widespread and vocal public outrage - often in the wake of gun violence - American support for stricter gun laws in 2020 fell to the lowest level since 2014, according to polling by Gallup.

Only 52% of Americans surveyed said they wanted stricter gun laws, while 35% said they should remain the same.

Eleven percent surveyed said laws should be "made less strict".

The issue is also one that is hyper-partisan and extremely divisive, falling largely along party lines.

"Democrats are nearly unanimous in their support for stricter gun laws," the same Gallup study noted, with nearly 91% in favour of stricter gun laws.

Only 24% Republicans, on the other hand, agreed with the same statement, along with 45% of Independent voters.

Some states have taken steps to ban or strictly regulate ownership of assault weapons. Laws vary by state but California, for example, has banned ownership of assault weapons with limited exceptions.

Presentational white space

Some controls are widely supported by people across the political divide - such as restrictions governing the sale of guns to people who are mentally ill or on "watch" lists.

Who opposes gun control?

Despite years of financial woes and internal strife, the National Rifle Association (NRA) remains the most powerful gun lobby in the United States, with a substantial budget to influence members of Congress on gun policy.

In January, the NRA filed for bankruptcy as part of a fraud case against some of its own senior staff. Even after the move, it vowed to continue "confronting anti-Second Amendment activities, promoting firearm safety and training, and advancing public programs across the United States".

Over the last several election cycles, it, and other organisations, have consistently spent more on pro-gun rights messaging than their rivals in the gun control lobby.

A number of states have also gone as far as to largely eliminate restrictions on who can carry a gun. In June 2021, for example, Texas Governor Greg Abbott signed into law a "permitless carry bill" that allows the state's residents to carry handguns without a license or training.

Similarly, on 12 April Georgia became the 25th in the nation to eliminate the need for a permit to conceal or openly carry a firearm. The law means any citizen of that state has the right to carry a firearm without a licence or a permit.

The law was backed by the NRA, and leaders within the organisation called the move "a monumental moment for the Second Amendment".

‘Horrifying’ conspiracy theories swirl around Texas shooting 

BY 

MAY 26, 2022



PROVIDENCE, R.I. (AP) — By now it’s as predictable as the calls for thoughts and prayers: A mass shooting leaves many dead, and wild conspiracy theories and misinformation about the carnage soon follow.


It happened after Sandy Hook, after Parkland, after the Orlando nightclub shooting and after the deadly rampage earlier this month at a Buffalo grocery store. Within hours of Tuesday’s school shooting in Uvalde, Texas, another rash began as internet users spread baseless claims about the man named as the gunman and his possible motives.


Unfounded claims that the gunman was an immigrant living in the U.S. illegally, or transgender, quickly emerged on Twitter, Reddit and other social media platforms. They were accompanied by familiar conspiracy theories suggesting the entire shooting was somehow staged.


The claims reflect broader problems with racism and intolerance toward transgender people, and are an effort to blame the shooting on minority groups who already endure higher rates of online harassment and hate crimes, according to disinformation expert Jaime Longoria.

“It’s a tactic that serves two purposes: It avoids real conversations about the issue (of gun violence), and it gives people who don’t want to face reality a patsy, it gives them someone to blame,” said Longoria, director of research at the Disinfo Defense League, a non-profit that works to fight racist misinformation.


In the hours after the shooting, posts falsely claiming the gunman was living in the country illegally went viral, with some users adding embellishments, including that he was “on the run from Border Patrol.”


“He was an illegal alien wanted for murder from El Salvador,” read one tweet liked and retweeted hundreds of times. “This is blood on Biden’s hands and should have never happened.”


The man who authorities say carried out the shooting, 18-year-old Salvador Ramos, is a U.S. citizen, Texas Gov. Greg Abbott said in a news conference on Tuesday.


Other social media users seized on images of innocent internet users to falsely identify them as the gunman and claim he was transgender. On the online message board 4Chan, users liberally shared the photos and discussed a plan to label the gunman as transgender, without any evidence to back it up.


One post on Twitter, which has since been deleted, featured a photo of a trans woman holding a green bottle to her mouth, looking into the camera, headphones hanging from one ear.


“BREAKING NEWS: THE IDENTITY OF THE SHOOTER HAS BEEN REVEALED,” claimed the user, saying the shooter was a “FEMBOY” with a channel on YouTube.


None of that was true. The photo actually depicted a 22-year-old trans woman named Sabrina who lives in New York City. Sabrina, who requested her last name not be published due to privacy concerns, confirmed to The Associated Press that the photo was hers and also said she was not affiliated with the purported YouTube account.


Sabrina said she received harassing responses on social media, particularly messages claiming that she was the shooter. She responded to a number of posts spreading the image with the misidentification, asking for the posts to be deleted.


“This whole ordeal is just horrifying,” Sabrina told the AP.


Another photo that circulated widely showed a transgender woman with a Coca-Cola sweatshirt and a black skirt. A second photo showed the same woman wearing a black NASA shirt with a red skirt. These photos didn’t show the gunman either — they were of a Reddit user named Sam, who confirmed her identity to the AP on Wednesday. The AP is not using Sam’s last name to protect her privacy.


“It’s not me, I don’t even live in Texas,” Sam wrote in a Reddit post.


Authorities have released no information on the gunman’s sexuality or gender identification.

Arizona Congressman Paul Gosar fit both unfounded claims about Ramos in a single now-deleted tweet that also misspelled his name. “It’s a transsexual leftist illegal alien named Salvatore Ramos,” Gosar tweeted Tuesday night.


Gosar’s office did not return a message seeking comment.


In some cases, misinformation about mass shootings or other events are spread by well-intentioned social media users trying to be helpful. In other cases, it can be the work of grifters looking to start fake fundraisers or draw attention to their website or organization.

Then there are the trolls who seemingly do it for fun.


Fringe online communities, including on 4chan, often use mass shootings and other tragedies as opportunities to sow chaos, troll the public and push harmful narratives, according to Ben Decker, founder and CEO of the digital investigations consultancy Memetica.


“It is very intentional and deliberate for them in celebrating these types of incidents to also influence what the mainstream conversations actually are,” Decker said. “There’s a nihilistic desire to prove oneself in these types of communities by successfully trolling the public. So if you are able to spearhead a campaign that leads to an outcome like this, you’re gaining increased sort of in-group credibility.”


For the communities bearing the brunt of such vicious online attacks, though, the false blame stirs fears of further discrimination and violence.


Something as seemingly innocuous as a transphobic comment on social media can spark an act of violence against a transgender person, said Jaden Janak, a PhD candidate at the University of Texas and a junior fellow at the Center for Applied Transgender Studies.

“These children and adults who were murdered yesterday were just living their lives,” Janak said Wednesday. “They didn’t know that yesterday was going to be their last day. And similarly, as trans people, that’s a fear that we have all the time.”


(Copyright (c) 2022 The Associated Press. All Rights Reserved. This material may not be published, broadcast, rewritten, or redistributed.)