Your Tech Story

machine learning

Deepfake Salvador Dali is Back to Life as Fakedeep; Takes Selfies with Museum Visitors

Artificial Intelligence has always amazed people with its surprising experiments. One of its wing Fakedeep is also one really amazing technique that even can bring back the deads. Yes! This has happened in the past. In the month of February, Burger King featured producer-director Andy Warhol in one of its ad eating a Whopper. There has been a question on the appearance of Warhol as he has been dead for nearly 32 years.

Video Source: The Dali Museum (YouTube.com)

And now, with the help of same technology, Dali Lives, has brought the surrealist painter Salvador Dali back to life, who had died 30 years ago, in 1989. Dali Lives is an exhibition unit that hosts exhibitions in the Dalí Museum in Florida for Dali’s paintings. The museum has partnered with ad agency Goodby, Silverstein & Partners (GS&P) to create a look-alike virtual being of Dali, who can interact with the museum visitors and even can take selfies with them.

The company has used the machine learning powered video editing technique to re-create the virtual Dali. They made the use of various pictures of Dali, and the video interviews, to mimic his voice and gestures. The company re-created his face by pulling over 6,000 frames through his images and interviews. It took 1000 hours for the AI algorithm to learn all of Dali’s facial expressions. An actor with the same general physical characteristics of Dali’s body did his body double, and the facial expression of Dali was imposed on the actor’s face. A voice-over artist learned his speech and mimicked his voice following Dali’s past interviews.

The video of virtual Dali is displayed in a kiosk. To watch the video, the visitors have to press a button on it. Dali will appear on the screen, and then have a conversation with the visitors. For this, the engineers made sure to include some of his quotes and the famous statements in the conversation. Every visitor will have a different experience with Dali, as every time he will be talking about different things. And after a 45-minute conversation, he will ask the visitors to take a selfie with them. Here too, the visitors won’t be ones to click the selfie, in fact, Dali will click the picture, and even, will text it to those visitors.

Though people have been concerned about the usage of Fakedeep technique, as people can also misuse it to defame other people, Dali Museum has no such intentions, but it created the virtual Dali to acknowledge his 30th death anniversary.

Microsoft Announces New Extended AI and Machine Learning Services to its Azure MI

Machine Learning is not the next big thing, but the current hot topic, which is the main focus of every tech giant today. Even though this technology is so much in demand, learning and experimenting with this technology needs a lot of expertise. So, to simplify things, Microsoft has announced that it is going to launch three new services to help people try their hands on machine learning and learn it better.

AzureML
Image Source: microsoft.com

“There is a category of AI practitioners, who are learning machine learning concepts, they want to make their own models, but they are not coders. This could be IT professionals or folks with a background in statistics or mathematics. For those customers, we’re offering experiences to make models visually.” said Microsoft’s director of artificial intelligence Bharat Sandhu.

Those new services will be the enhancements to Azure Machine Learning. The Azure Machine Learning lets the users create AI architectures for cloud-hosted and on-premises apps. With the new services, Azure MI will now make the process of model creation more simpler, providing the users with a simple drag and drop visual MI dashboard, from which they can customize their MI program.

The new services that the company will be launching include an interface, which will be responsible to automate the process of model creation. This way the developers do not have to write even a single line of code. The company has also added new algorithms and optimizations to the platform, that will help in creating more accurate models. These algorithms are transparent to the users, and they can even manually write the codes for the same.

Along with the ML services, the company will also be adding new capabilities for AI, IoT, blockchain and edge computing to the Azure cloud platform. The AI Azure Cognitive Services will enable the application to make decisions, see, hear, translate, and reason. In addition to that, the company will also be including the AI features to the Azure Search.

Microsoft has also announced to add a new SQL engine, i.e. Azure SQL Database Edge, having built-in AI features, to handle the computing and database management tasks, intelligently. On the other hand, the new IoT Plug and Play feature will allow the developers to connect IoT devices directly to the cloud, eliminating the need to write the code for the peripherals.

Along with all these services, Microsoft also announced an Azure Blockchain Service, Unreal Engine, and a HoloLens 2 Developer Bundle. The company has not commented on when those services will be available for use. But since it has made all these announcements ahead of its Build developer conference in Seattle, Washington next week, the company might be planning of announcing the launch date at the event.