‘Deepfake’ technology can now create real-looking human faces
A new study from Nvidia researchers show just how far artificial image-generation technology has come in recent years.
- In 2014, researchers introduced a novel approach to generating artificial images through something called a generative adversarial network.
- Nvidia researchers combined that approach with something called style transfer to create AI-generated images of human faces.
- This year, the Department of Defense said it had been developing tools designed to detect so-called 'deepfake' videos.
A new paper from researchers at Nvidia shows just how far AI image generation technology has come in the past few years. The results are pretty startling.
Take the image below. Can you tell which faces are real?
Karros et al.
Actually, all of the above images are fake, and they were produced by what the researchers call a style-based generator, which is a modified version of the conventional technology that's used to automatically generate images. To sum up quickly:
In 2014, a researcher named Ian Goodfellow and his colleagues wrote a paper outlining a new machine learning concept called generative adversarial networks. The idea, in simplified terms, involves pitting two neural networks against each other. One acts as a generator that looks at, say, pictures of dogs and then does its best to create an image of what it thinks a dog looks like. The other network acts as a discriminator that tries to tell fake images from real ones.
At first, the generator might produce some images that don't look like dogs, so the discriminator shoots them down. But the generator now knows a bit about where it went wrong, so the next image it creates is slightly better. This process continues until, in theory, the generator creates a good image of a dog.
What the Nvidia researchers did was add to their generative adversarial network some principles of style transfer, a technique that involves recomposing one image in the style of another. In style transfer, neural networks look at multiple levels of an image in order to discriminate between the content of the picture and its style, e.g. the smoothness of lines, thickness of brush stroke, etc.
Here are a couple examples of style transfer.
In the Nvidia study, the researchers were able to combine two real images of human faces to generate a composite of the two. This artificially generated composite had the pose, hair style, and general face shape of the source image (top row), while it had the hair and eye colors, and finer facial features, of the destination image (left-hand column).
The results are surprisingly realistic, for the most part.
Karros et al.
Concerns over 'deepfake' technology
The ability to generate realistic artificial images, often called deepfakes when images are meant to look like recognizable people, has raised concern in recent years. After all, it's not hard to imagine how this technology could allow someone to create a fake video of, say, a politician saying something abhorrent about a certain group. This could lead to a massive erosion of the public's willingness to believe anything that's reported in the media. (As if concerns about 'fake news' weren't enough.)
To keep up with deepfake technology, the Department of Defense has been developing tools designed to detect deepfake videos.
"This is an effort to try to get ahead of something," said Florida senator Marco Rubio in July. "The capability to do all of this is real. It exists now. The willingness exists now. All that is missing is the execution. And we are not ready for it, not as a people, not as a political branch, not as a media, not as a country."
However, there might be a paradoxical problem with the government's effort."Theoretically, if you gave a [generative adversarial network] all the techniques we know to detect it, it could pass all of those techniques," David Gunning, the DARPA program manager in charge of the project, told MIT Technology Review. "We don't know if there's a limit. It's unclear."
Researchers have just discovered the remains of a hybrid human.
90,000 years ago, a young girl lived in a cave in the Altai mountains in southern Siberia. Her life was short; she died in her early teens, but she stands at a unique point in human evolution. She is the first known hybrid of two different kinds of ancient humans: the Neanderthals and the Denisovans.
These thought leaders, founders, and entrepreneurs are propelling the kind of future we want to be a part of.
- The tech industry may be dominated by men in terms of numbers, but there are lots of brilliant women in leadership positions that are changing the landscape.
- The women on this list are founders of companies dedicated to teaching girls to code, innovators in the fields of AI, VR, and machine learning, leading tech writers and podcasters, and CEOs of companies like YouTube and Project Include.
- This list is by no means all-encompassing. There are many more influential women in tech that you should seek out and follow.
Most said they want to act on their desire someday. But do open relationships actually work?
- The study involved 822 Americans who were in monogamous relationships at the time.
- Participants answered questions about their personalities, sexual fantasies, and intentions to act on those fantasies.
- Research suggests practicing consent, comfort, and communication makes open relationships more likely to succeed.
Consensual non-monogamy fantasies<p>For the new study, published in <a href="https://link.springer.com/article/10.1007/s10508-020-01788-7" target="_blank">Archives of Sexual Behavior</a>, researchers asked 822 people in monogamous relationships to:</p><ul><li>Describe their favorite sexual fantasy, defined as "mental images you have while you are awake that you find to be sexually arousing or erotic."</li><li>Select which themes apply to that fantasy, such as having sex with multiple people at the same time, experimenting with taboos, or engaging in a sexually open relationship.</li><li>Answer whether they intended to carry out these fantasies, and discuss them with their partner.</li><li>Complete assessments on relationship satisfaction, erotophilia and personality, as measured by the Big Five Personality inventory.</li></ul><p>The results showed that 32.6 percent of participants said being part of a sexually open relationship was "part of their favorite sexual fantasy of all time." More surprising is that, of that one-third, 80 percent said they want to act on this fantasy in the future.</p>
Pretzelpaws via Wikipedia Commons<p style="margin-left: 20px;">"The present research confirms the important distinction between sexual fantasy and sexual desire in that not everyone wanted to act on their favorite sexual fantasy of all time," study author Justin J. Lehmiller told <a href="https://www.psypost.org/2020/09/one-third-of-people-in-monogamous-relationships-fantasize-about-being-in-some-type-of-open-relationship-study-suggests-58102" target="_blank">PsyPost</a>. "This suggests that fantasies may serve different functions for different people."</p><p>Even though most participants said they want to act out their fantasy in the future, far fewer reported acting out sexual fantasies in the past. Other findings included:</p><ul><li>Men were more likely to fantasize about CNMRs.</li><li>So were people who scored high in <a href="https://en.wikipedia.org/wiki/Erotophilia#:~:text=Erotophilia%20is%20a%20personality%20trait,ranging%20from%20erotophobia%20to%20erotophilia." target="_blank">erotophilia</a> and sociosexual orientation.</li><li>The psychological predictors of fantasizing about CNMRs differed from predictors about infidelity fantasies.</li></ul>
Do open relationships work?<p>A <a href="https://www.tandfonline.com/doi/full/10.1080/00224499.2019.1669133" target="_blank">2019 study</a> from psychologists at the University of Rochester suggests it <em>is </em>possible<em>, </em>but especially when both partners practice a trio of behaviors: consent, communication, and comfort — or, the Triple-C Model.<br></p>But the study also suggests not all forms of open relationships are equally viable. For example, people in one-sided CNMRs — where one partner stays monogamous, the other seeks outside sexual relationships — were nearly three times more dissatisfied in their relationships than the monogamous group <em>and </em>the consensual non-monogamous group.
The results of this study showed depressive symptoms being highest in adolescence, declining in early adulthood and then climbing back up again into one's early 30s.