London as world’s AI capital. Amazon Lex. FB drones provide wifi to disaster areas. Baidu + self-driving cars. IBM Watson as art museum guide.
A couple of months ago, Yoshua Bengio said that Canada is at the forefront of the AI revolution. When I asked the community about this statement, I got quite a varied response, with some agreeing with him, others voting for the US or the UK, and a handful for Israel for which country is at the forefront of the AI revolution.
Last week, NetGuru argued that London is on its way to becoming the ‘world’s AI capital’, due to its hot AI startups, the amount of investment in these startups (as well as M&As), the high level of AI academic research, and more. Do you think that London is set to become the world’s AI capital? Looking forward to hearing your thoughts.
AWS said it is opening up its Amazon Lex AI service to all customers so they can build applications.
Amazon Lex uses the same machine learning technology as Amazon Alexa. Amazon Lex features algorithms that enable applications that can have conversations and process voice and text. For AWS, opening up Lex to its broad customer base could give it scale. Another key item to note about Amazon Lex is that AWS is offering the AI as a managed service. As a result, more enterprises are likely to be able to try Lex and build applications.
The company announced on Wednesday what it’s calling “Tether-tenna technology,” essentially a small, unmanned helicopter that will provide Wi-Fi access to crisis zones when existing Wi-Fi towers are down or damaged.
Yael Maguire, head of Facebook’s connectivity lab, estimates that one helicopter could connect “in the neighborhood of thousands to tens of thousands of people.” The helicopter-drone, which is roughly the size of a Volkswagen Beetle, is literally tethered to a fiber line and a power source on the ground, which enables the chopper to stay airborne for days at a time. (Facebook says its goal is to keep it up for weeks or months.)
Baidu has announced a new project named “Apollo” which will provide an open, complete and reliable software platform for its partners in the automotive and autonomous driving industry to develop their own autonomous driving systems with reference vehicles and hardware platform.
“AI has great potential to drive social development, and one of AI’s biggest opportunities is intelligent vehicles,” said Qi Lu, Group President and Chief Operating Officer at Baidu. The company will also initiate a partnership alliance, working with partners who will provide the best and most compatible vehicles, sensors, and other components.
Jun-Yan Zhu and Taesung Park, the researchers behind the AI model powering the doodle-to-cat-monster tool, are back with another impressive image manipulation implementation that lets you turn horses into zebras, apples into oranges, winters into summers and so much more.
In their paper, they lay out a new model that essentially allows you to transform images in a ‘cycle consistent’ way, meaning that any changes to the original image are expected to ultimately remain fully
reversible. Check out the code here.
Almost three-quarters of Brazilians have never been inside a museum, according to a 2010 study from the Brazilian Institute of Economic Research. There are probably many reasons for this, but among them is the feeling that art can seem inaccessible unless you’ve studied it.
So, how do you get art to speak to you specifically? By getting it to speak, period.
For the launch of IBM Watson in Brazil, Ogilvy Brazil created an interactive guide that lets people have conversations with work housed at the Pinacoteca de São Paulo Museum. “The Voice of Art” replaces pre-recorded audio guides with a Watson-powered program that gleans data from books, old
newspapers, recent articles, biographies, interviews and the internet.
Researchers are training artificial intelligence to identify tuberculosis on chest X-rays, an initiative that could help screening and evaluation efforts in TB-prevalent areas lacking access to radiologists.
For the study, Dr Paras Lakhani and his colleague, Dr Baskaran Sundaram, obtained 1,007 X-rays of patients with and without active TB. The cases were used to train two models – AlexNet and GoogLeNet – which learned from TB-positive and TB-negative X-rays. The models’ accuracy was tested on 150 cases.