Posted on August 8, 2018 by staff

AI Briefing: Picasso makeover for anonymous news sources


When anonymity is needed in a news report or investigative documentary, the tried and tested method is to blur the faces of those who require protection.

Now one university is looking to deliver a high-tech solution with AI which takes a leaf out Picasso’s book and turns to cubism.

The team at Simon Fraser University says it uses ‘art abstraction’ to repaint the frames of the source video.

Professor Steve DiPaola said: “When artists paint a portrait, they try to convey the subject’s outer and inner resemblance.

“With our AI, which learns from more than 1,000 years of artistic technique, we have taught the system to lower the outer resemblance and keep as high as possible the subject’s inner resemblance – in other words, what they are conveying and how they are feeling.

“Our system uses five levels of AI processing to simulate a smart painter, using art abstraction to repaint the video as if an artist was painting every frame.

“The result is more engaging, especially since not everyone listens to stories – so the art component becomes more relevant.”

The team received a Google/Knight Foundation grant to research the tech.



The theme park hearing problems

Isn’t it funny how everyone becomes a mechanical surveyor just before a rollercoaster ride starts? ‘What’s that noise? Should this be so loose? Is that rust?’

The majority of the time those fears are unfounded – and people hop off a ride with a smile and a wild new hair-do.

There is something to it, though. You can hear faults in mechanical parts, but you’ve got to be smart. As smart as AI.

Netherlands-based Efteling Park is taking part in a specifically designed challenge organised by The Next Web and Vodafone to do just this. It’s hoping to find a company that can develop AI which can ‘hear’ when a rollercoaster needs its screws tightening.

The theme park hopes that this tech can to reduce regular maintenance times dramatically, reducing downtime, leaving more time for anxious thrill seekers to hope they don’t get stuck on the loop-the-loop.

An AI model for models

Researchers from the University of Pittsburgh are building a model of the most persuasive faces in the ad business.

It could be a huge draw for advertising agencies: knowing which types of face persuade people to buy means less guess-work, more sales and no more dealing with temperamental superstars.

By analysing existing ads, and with a little help from humans, the researchers have shown how faces change across a wide range of ad groups.

The images show the amalgamated faces appearing in ads in the travel, clothing and restaurants industry as well as ads opposing smoking, alcohol abuse and domestic violence. And they’re just getting started.

The researchers, Christopher Thomas and his colleague Adriana Kovashka, are now imagining a future in which ads are entirely personalised to the viewer.

These computer-generated models might one day be the face of the ads you see, and will be designed to look more like you in an effort to increase your responsiveness.

An AI service with a smile

Despite its best efforts, the Microsoft brand won’t ever be ‘cool’. But when it’s not powering nearly every office computer in the world, the company can often be found on some very interesting side projects.

Its latest is an AI that can tell when you’re laughing. The company has installed a Laugh Battle machine at the newly opened National Comedy Center in New York, the first museum dedicated to comedy.

Two competitors sit across from each other and try to make each other laugh. The first one to crack loses, and it’s up to the AI to call out the winner.

The system uses cameras and a deep neural network trained on more than 100,000 photos. It can recognise a range of facial expressions that are linked to eight emotions including happiness and sadness as well as anger, contempt, disgust, fear, neutral and surprise.


AI healthtech start-up pressing ahead with IPO

Serial entrepreneur and former UK science minister Lord Drayson plans to take his digital healthcare firm public later this month.

His company, Senseye Health, works in partnership with universities and NHS trusts in develop AI-led medical products and services, as well as analysing the real-world data it receives to make new discoveries.

He spoke to my colleague Mo Aldalou about his plans to build a world-leading tech company.