New AI is amazing at depixelating images — but it also highlights the problem of data bias

BLOG: Heidelberg Laureate Forum

Laureates of mathematics and computer science meet the next generation
Heidelberg Laureate Forum

It’s almost become a cliche for AI stories, but this new algorithm does a brilliant job of highlighting it: the technology is exciting and remarkable, but far from mature and robust — and data bias can be an issue.

Image credits: Duke University.

The algorithm, called PULSE, attempts to produce high-resolution images from pixelated portraits. It’s not the first time algorithms have been developed for the purpose of de-pixelating images, but PULSE does it a bit differently: it’s not ‘zoom and enhance’, it’s entirely another approach.

Instead of trying to see what colors would fit in each pixel, the algorithm actually produces new, original faces. PULSE is essentially a GAN — a Generative Adversarial Network, in which two different algorithms work in an adversarial fashion. In this context, one algorithm would produce an original face not belonging to any human, and the other algorithm tries to find differences between the pixelated image and the generated face.

In recent times, advances in this field have surged. You need only visit This Person Does Not Exist and click refresh several times — you’ll see just how eerily realistic these AI-generated faces can be. PULSE uses a somewhat similar approach, except it also compares the generated images with the pixelated image to produce a plausible, realistic portrait.

Even when facial features are barely recognizable, “our algorithm still manages to do something with it, which is something that traditional approaches can’t do,” said co-author Alex Damian ’20, a Duke math major.

Promising, but not perfect: Portraits of the authors of the paper, pixelated and un-pixelated. Image credits: Duke University.

However, there’s a catch: none of these faces are real. In other words, what the algorithm is producing isn’t a high-resolution version of the pixelated image, it’s a brand-new face that fits with the input data.

This means that it can’t, and shouldn’t be used for something like facial recognition or crime scene investigation. Instead, researchers say it could be used to complement satellite imagery and medical imaging, as well as microscopy and astronomy.

You can see the results for yourself here or by checking the published paper at arXiv. You can also check out the AI and use it yourself via GitHub, and this is where things get even more interesting.

As expected, when you expose an AI to the online world, people start testing and experimenting with it.

Image credits: Twitter / tg_bomze.

First, one of the study authors themselves played with the AI, generating a human image of the famous game character B.J. Blazkowicz, who starred in the famous Wolfenstein games. For millions of old-school gamers, it was a chance to see how the protagonist of several cult classics may have looked like in reality. It’s hard to imagine that a man like B.J. Blazkowicz could exist in reality, but if he did, that’s probably what he would have looked like.

But then, other people came across something weird. In some cases, the algorithm behaved weirdly. Here’s an example of the type of unexpected result:

Image credits: Twitter / Chicken3gg

Most people would look at the left picture and they’d see a pixelated photo of former US President Barack Obama. Yet, the image on the right looks nothing like Obama.

As mentioned, the faces don’t belong to an actual person so it was bound to generate some differences, but this is more than an isolated example: it’s a trend, Twitter users found.

Here’s another example, depicting a pixelated an unpixelated image of US Congresswoman Alexandria Ocasio-Cortez.

Image credits: Twitter / Robert Osazuwa Ness.

Twitter user Robert Osazuwa Ness posted another example, of his wife:

Image credits: Twitter / Robert Osazuwa Ness.

… and another one of actress Lucy Liu:

Image credits: Twitter / Robert Osazuwa Ness.

By now, you probably see the trend: the AI appears to have a bias in how it depixelates faces. Obviously the AI has no opinion or biases, so what’s going on?

The problem with any AI, or any data project really, is that it’s only as good as its input data. If you want to make good wine, you need good grapes, and finding the right grapes is challenging.

This is probably what’s happening here: the algorithm was trained with a specific dataset, and if most people in that dataset tend to look in a specific way, the AI is more inclined to generate images that fit these trends.

This data bias is an artifact of the dataset, and the authors themselves did not appear fully aware of it until it was pointed out.

“It does appear that PULSE is producing white faces much more frequently than faces of people of color,” wrote the algorithm’s creators on Github. “This bias is likely inherited from the dataset […] though there could be other factors that we are unaware of.”

At this point, it’s still too soon whether this is a shortcoming of the algorithm itself or an artifact of the training data, but the problem is surprisingly widespread in machine learning.

It’s actually a well-known problem in facial recognition: algorithms often work worse on non-white people. It’s not entirely clear why this happens, but an often-quoted NIST report found that even the best algorithms still struggle with this.

Black women tend to have the highest false match rate for facial recognition algorithms. Image credits: NIST.

Data bias can be a major pitfall of machine learning, and it’s an issue that researchers need to be aware of, both for socially and technical reasons. These are just a few (potentially isolated) examples of bias, but there can be many more problem

Of course, the algorithm itself isn’t perfect, and could also be improved upon.

An example of a clearly imperfect transformation, starting from a pixelated image of President Donald Trump. Image credits: Twitter / Moritz Klack.

But in the grand scheme of things, it’s more important than ever to be aware of this bias and address it. This is an interesting and important study, highlighting both the potential and the shortcomings of current machine learning.

We’re experiencing an advent of AI, and while still in its infancy, the technology is taking major strides, so the earlier we can set it on a healthy course, the better.

Avatar photo

Posted by

Andrei is a science communicator and a PhD candidate in geophysics. He is the co-founder of ZME Science, where he published over 2,000 articles. Andrei tries to blend two things he loves (science and good stories) to make the world a better place -- one article at a time.


  1. PULSE has a bias toward faces which are used in the training set.
    But humans have the same bias. I see no difference. If Obamas and Trumps face would have been part of the training set, the pixelated Obama PULSE‘d would be the full resolution Obama, the same for Trump.
    One solution would be that PULSE uses a very huge training set and delivers not only one but several possible solutions.

  2. Trump is a little child -I don’t see the problem! 🙂

    Joking aside: Maybe one could experiment with a switch for the general appearence. “Which appearence would you like as output? A) Asian B) White C) Colored D)….
    For that, I guess one has to seperate the training sets. But this could potentially hard-code racism.
    We can put it in a positive way: Color is unimportant for “PULSE”. It doesn’t know that it is so important for *us*…

    • A general bias for white people is offen falsely attributed to AI-algorithms, but is in fact a wellknown problem of photography: dark surfaces absorb more light and make it more difficult to render shading nuances. This can be „corrected“ either by proper illumination (using reflected light) or algorithmically in a similar way as in high dynamic range Images. The New York Times article The Racial Bias Built Into Photography has this problem as its subject.

  3. I read about that strange phenomenon at „The Verge“. The creators of this AI kinda apologized for that mishap.

    As you mentioned in your text:

    “The problem with any AI, or any data project really, is that it’s only as good as its input data. If you want to make good wine, you need good grapes, and finding the right grapes is challenging.”

    The dataset used to train the underlying algorithm was composed of a couple of some hundred thousand average images of British Subjects living in the U.K.

    As Great Britain’s population is predominantly white, the system is prone to generate white faces.

    Very unfortunate, Lucy Lui looks much cuter than the white woman’s face generated by the AI.

Leave a Reply

E-Mail-Benachrichtigung bei weiteren Kommentaren.
-- Auch möglich: Abo ohne Kommentar. +