The company published a 10-page research paper (PDF) this week detailing how their system works and likened it to existing photo retouching tools such as red-eye correction.

Facebook’s idea is not entirely original, a similar feature exists in Adobe Photoshop Elements. The engineers acknowledge this in their paper but believe their ExGAN (Exemplar Generative Adversarial Networks) technique offers superior results.

In the examples below, the first column represents an exemplar photo of the subject with their eyes open. The second column shows the original closed eye photo to be fixed. In the third column, a closed eye correction using Adobe Photoshop Elements can be seen. Finally, the fourth column shows the results of Facebook’s algorithm.

As can be seen in the examples above, the engineers’ AI cleverly takes into account variables such as lighting and can adjust the resulting correction accordingly.

The idea is still in its research phase but offers an impressive demonstration of what’s possible when AI is applied to photography. It’s unclear when or if Facebook will roll this out for photos on their social network.

Photos hold our memories and will remain dear to us throughout our lives and perhaps even after down through generations. Any correctional feature which could save an otherwise nice photo from being deleted is sure to be welcomed by users.

What are your thoughts on Facebook’s closed eye correction?

The post Facebook can open closed eyes in photos using AI correction appeared first on AI News.

“In the report, you’ll see that in the first three months of this year we took down 837 million pieces of spam and disabled 583 million fake accounts,” wrote Facebook CEO Mark Zuckerberg in a post. “Thanks to AI tools we’ve built, almost all of the spam was removed before anyone reported it, and most of the fake accounts were removed within minutes of being registered.”

The internet can be a fantastic tool, it’s connected us across borders like never before and provides a near-infinite resource of knowledge. However, it’s also provided space for hate, dangerous ideology, and criminal activity.

Here are some of Facebook’s key achievements:

• We took down 21 million pieces of adult nudity and sexual activity in Q1 2018 — 96% of which was found and flagged by our technology before it was reported. Overall, we estimate that out of every 10,000 pieces of content viewed on Facebook, 7 to 9 views were of content that violated our adult nudity and pornography standards.
• For graphic violence, we took down or applied warning labels to about 3.5 million pieces of violent content in Q1 2018 — 86% of which was identified by our technology before it was reported to Facebook.
• For hate speech, our technology still doesn’t work that well and so it needs to be checked by our review teams. We removed 2.5 million pieces of hate speech in Q1 2018 — 38% of which was flagged by our technology.

Content moderators have some of the most distressing jobs imaginable and often have to receive counselling for what they’re subjected to on a daily basis — from child abuse, to beheadings, and more.

AI is stepping in to reduce the amount of content which has to be moderated by a human and help to get it taken down quickly before it reaches many people. Yet it’s still not perfect.

Guy Rosen, VP of Product Management, wrote:

“Artificial intelligence, while promising, is still years away from being effective for most bad content because context is so important.

For example, artificial intelligence isn’t good enough yet to determine whether someone is pushing hate or describing something that happened to them so they can raise awareness of the issue.”

Facebook’s candidness that it still needs to do more to combat hate speech will be welcomed by many international governments who have long called for social networks to take more responsibility for the content on their platforms.

You can find the full report here.

What are your thoughts on Facebook’s enforcement report?

The post Facebook’s AI continues to fight the worst of the web appeared first on AI News.

The social media giant acquired Instagram for $1 billion in 2012. At the time, some criticised its value with its 30 million users. Now, however, the platform has over 800 million active monthly users.

Facebook trained a machine learning algorithm to categorise images by itself using over 3.5 billion posted on Instagram. The algorithm took into account the 17,000 hashtags used for the images.

Speaking during Facebook’s F8 developer conference, CTO Mike Schroeper said:

“The biggest limiting factor to making progress in computer vision — as in many fields of AI — is that we rely almost entirely on hand-labelled, human-curated, data sets.

If a person hasn’t spent the time to label something specific in an image, even the most advanced computer vision systems won’t be able to detect it at runtime because it hasn’t seen it in the training set.

We built some breakthrough technology that takes publicly available hashtagged images at an unprecedented scale. We have trained on 3.5 billion training images using a public set of images without any human curated images in that data set.”

That image set is 10x the size of the 300 million Google uses to train their algorithms, which the company disclosed in a paper (PDF) released in August 2017.

Srinivas Narayanan, Engineering Director at Facebook, added:

“We have now created the world’s best computer vision system. It achieves the highest score ever of 84.5% accuracy on ImageNet — a dataset widely used for benchmarking.”

Back in August last year, AI News reported that IBM had set a record for image recognition with a 33.8 percent accuracy rate after analysing 7.5 million pictures over seven hours. Facebook may have just set the new standard.

What are your thoughts on Facebook’s computer vision achievement?

The post Facebook trained its AI using 3.5 billion Instagram images appeared first on AI News.

Facebook’s new suicide prevention tool aims to use pattern recognition to detect posts or live videos where someone might be expressing thoughts of suicide, and to help respond to reports faster.

In a post announcing the feature, Facebook wrote: “We are starting to roll out artificial intelligence outside the US to help identify when someone might be expressing thoughts of suicide, including on Facebook Live. This will eventually be available worldwide, except the EU.”

EU data protection regulations

Facebook’s notable lack of support for its latest AI advancement in EU member countries is likely due to strict data protection regulations.

In recent months, I’ve spoken to lawyers, executives from leading companies, and even concerned members of the European Parliament itself about the EU’s stringent regulations stifling innovation across member states.

Julia Reda, an MEP, says: “When we’re trying to regulate the likes of Google, how do we ensure that we’re not also setting in stone that any European competitor that might be growing at the moment would never emerge?”

My discussions highlighted the fear that European businesses will struggle without the data their international counterparts have access to, and startups may look to non-EU countries to set up their operations.

However, the situation has taken a more serious turn with the potential for loss of life. Beyond the inability to launch potentially life-saving features like Facebook’s suicide prevention, the regulations will slow innovation in fields benefiting from AI such as healthcare.

We often cover medical developments on AI News, and most of these advancements rely on data collection to improve machine learning models. GDPR puts significant restrictions on how, when, and why firms can collect and use this data — which simply do not exist to such an extent anywhere else in the world.

“You’ve got your Silicon Valley startup that can access large amounts of money from investors, access specialist knowledge in the field, and will not be fighting with one arm tied behind its back like a competitor in Europe,” comments Peter Wright, Solicitor and Managing Director of Digital Law UK. “Very often we hear ‘Where are the European Googles and Facebooks?’ Well, it’s because of barriers like this which stop organisations like that being possible to grow and develop.”

The issue stems from austere EU data protection regulations being unsuitable for today’s world. There’s little debate against the need to safeguard data, and penalise where this has been insufficient, but even companies with a history of protecting their users are concerned about the extent of this legislation.

“We deal with a very large amount of customer data at F-Secure and I don’t go a working day without hearing a GDPR discussion around me,” comments Sean Sullivan, Security Advisor at Finnish cyber security company F-Secure. “It’s a huge effort, and many people are involved within my part of the organisation.”

“And not just at the legal level; we have ‘data people’ working with our product developers on our software architecture. We’ve always been a privacy focused company, but the last year has been a whole new level in my experience.”

The penalties for non-compliance with GDPR are severe and could devastate a company. Startups in particular, especially in areas such as AI, will struggle from being unable to collect anywhere near as much data as current leaders such as Google hold. However, that doesn’t mean established companies have it easy.

“Fortunately, we have the people we need. I imagine Facebook is still in the position of needing to find California-based GDPR experts who can work with the local developer teams,” explains Sullivan. “I’m confident it has people in Europe who are working on the high level issues, but I doubt that all of the product teams will be able to find the needed resources to be confident of GDPR compliance.”

“There will be more tech innovations that won’t be rolled out in the EU. Hopefully not for long, but at least for the near future.”

With its billions of users, there’s a good chance everyone has friends and family who use Facebook. I’m certain if anyone expresses suicidal thoughts on the platform we’d all want them to receive help as soon as possible.

For many consumers, this situation will be the first to bring awareness to the negative impacts of the EU’s strict data regulations. For businesses, this serves as yet another example.

If you’re having thoughts of suicide or self-harm, please find a list of international helplines here.

What are your thoughts on the EU’s data protection regulations?

The post Editorial: Facebook excluding the EU from AI advancement heralds a trend appeared first on AI News.