SpaceX founder Elon Musk watches a post-launch press conference after the SpaceX Falcon 9 rocket, which carries the spaceship Crew Dragon, is lifted off on an untouched test flight to the International Space Station from the Kennedy Space Center in Cape Canaveral, Florida, March 2, 2019.
Mike Blake | Reuters
Armchairs in the form of avocados and daikon radishes wearing tutu are among the quirky images created by a new software from OpenAI, an Elon Musk-supported laboratory for artificial intelligence in San Francisco.
OpenAI trained the software, known as Dall-E, to generate images from short text texts. It specifically used a data set of 1
The lab said Dall-E – a portmanteau by Spanish surrealist artist Salvador Dali and Wall-E, a small animated robot from the Pixar movie of the same name – had learned how to create images for a wide range of concepts.
OpenAI showed off some of the results in a blog post published on Tuesday. “We have found that [Dall-E] “has a diverse set of capabilities, including creating anthropomorphized versions of animals and objects, combining unrelated concepts in plausible ways, rendering text and applying transformations to existing images,” the company wrote.
Dall-E is built on a neural network, which is a computer system vaguely inspired by the human brain that can detect patterns and recognize relationships between large amounts of data.
While neural networks have generated images and videos before, Dall-E is unusual because it relies on text input, while the others do not.
Synthetic videos and images have become more sophisticated in recent years, to the extent that it has become difficult for people to distinguish between what is real and what is computer generated. General Controversial Networks (GANs), which use two neural networks, have been used to make fake videos of politicians, for example.
OpenAI acknowledged that Dall-E has “the potential for significant, broad societal effects”, adding that it plans to analyze how models such as Dall-E “relate to societal issues such as economic impact on certain work processes and occupations, the potential for bias in model outputs, and the long-term ethical challenges posed by this technology. “
Dall-E comes just a few months after OpenAI announced that they had built a text generator called GPT-3 (Generative Pre-training), which is also supported by a neural network.
The language generation tool is capable of producing human-like text on demand, and it became relatively famous for an AI program when people realized that it could write its own poetry, news articles and short stories.
“Dall-E is a GP2-3 Text2Image system, but trained on text plus images,” Mark Riedl, associate professor at the Georgia Tech School of Interactive Computing, told CNBC.
“Text2image is not new, but the Dall-E demo is remarkable for producing illustrations that are much more coherent than other Text2Image systems I’ve seen in recent years.”
OpenAI has competed with companies such as DeepMind and the Facebook AI Research group to build general algorithms that can perform a wide range of tasks at the human level and beyond.
Researchers have built AI that can play complex games such as chess and the Chinese board game Go, translate one human language into another and spot tumors in a mammogram. But getting an AI system to show real “creativity” is a big challenge in the industry.
Riedl said the Dall-E results show that they have learned how to mix concepts coherently, adding that “the ability to coherently mix concepts is considered a key form of creativity in humans.”
“From a creativity point of view, this is a big step forward,” Riedl added. “While there is not much agreement on what it means for an AI system to ‘understand’ something, the ability to use concepts in new ways is an important part of creativity and intelligence.”
Neil Lawrence, former director of machine learning at Amazon Cambridge, told CNBC that Dall-E looks “very impressive.”
Lawrence, now a professor of mechanical engineering at the University of Cambridge, described it as “an inspiring demonstration of the capacity of these models to store information about our world and generalize in ways that humans find very natural.”
He said: “I expect there will be all kinds of applications of this type of technology, I can not even begin to imagine. But it is also interesting when it comes to being another pretty impressive technology that solves problems we do not even know we actually had it. “
‘Does not promote the state of AI’
However, not everyone is impressed with Dall-E.
Gary Marcus, an entrepreneur who sold a machine learning startup to Uber in 2016 for an unknown sum, told CNBC that it’s interesting, but it “does not promote the state of AI.”
He also pointed out that it has not been opened, and the company has not yet published an academic article on the research.
Marcus has previously questioned whether any of the research published by rival laboratory DeepMind in recent years should be classified as a “breakthrough.”
OpenAI was set up as a non-profit company with a $ 1 billion pledge from a group of founders that included Tesla CEO Elon Musk. In February 2018, Musk left the OpenAI board, but he continues to give and advise to the organization.
OpenAI made a profit in 2019 and raised another $ 1 billion from Microsoft to fund the research. GPT-3 is set to be OpenAI’s first commercial product, and Reddit has registered as one of the first customers.