Omnipresence, Human-machine Symbiosis, & Open-source AI

The latest trends in Data & AI developed by deep-tech startups from our Global Startup Challenge

Artificial Intelligence is the New Electricity” — Andrew Ng.
Founder of education platform Coursera and former director of the Baidu Lab pinpointed the disruptive power of AI technology which is revolutionizing the world.

Artificial Intelligence has already shown us that it can be omnipresent: Uber is famously developing a self-driving car, Amazon has recently unveiled the concept of an autonomously-run shop, and Google is using AI to cut data centre consumption. All these applications are needed to advance towards a better, more sustainable future, made possible by creating and fostering relevant ecosystems.

Startups from our Global Startup Challenge are embracing this: the new techniques among speech recognition, natural and neural language processing, computer vision, developed in labs or in research centres, are paving the way for a future where digitalization and AI are part of the daily norm. Lyrebird, spinoff from the University of Montreal, is developing an API for speech synthesis and Mycroft.AI complements this ecosystem by creating the world’s first open source voice assistant. Let’s not forget that collaboration and transparency are vital in the development of AI, as these startups are not only fighting the monopoly created by GAFAs, but are also corroborated by the recent headlines made by Elon Musk and Mark Zuckerberg regarding AI’s impact on humanity, raising the inevitable question of responsibility. Both Lyredbird and Mycroft. AI stressed the importance of ethical codes of practice and fully embraced the danger of a world where “the sky’s the limit” and unsupervised learning could be a reality within a time frame of 10 to 15 years.

Lyrebird creates a digital copy of Barack Obama’s voice

A second trend occurring in Data & AI is inevitably the democratization of Augmented Reality and Virtual Reality, and more importantly the increase of their capabilities. Providing a new experience of your surroundings has always been tempting, and its applications are wide ranging from industrial monitoring of factories to gaming. Lampix is transforming any surface to become a smart surface, enabling possibilities for everyday use such as office collaboration or smart homes furnitures. On the other hand, Wrnch, a startup from Canada, is embracing deep learning techniques and computer vision to extract the 3D humans in standard 2D videos & pictures. Both need cutting edge talents, research and large data sets upon which they can build and train their algorithms. This has been made possible by the recent democratization and the multiplication of actors in the field of AR and VR.

Mapix smart surfaces
Wrnch’s technology BODYSLAM, a Human Machine Interaction Engine

The third trend identified among startups in this field is recent improvement in human-machine symbiosis. Ziva Dynamics is changing how virtual humans are created and rendered in interactive experiences. By having ultra-realistic interactive characters, they are enabling a better understanding and more interaction with our human characters in a virtual environment. More generally, it brings the biological-machine frontier to our attention, where our 2016 alumni Koniku Inc, made initial advancements. They are developing what is likely to be “the first and only company on the planet building chips with biological neurons.”, enabling a deeper understanding to one of the greatest challenges of our time: “to understand the brain and the mind to ultimately allow humanity to enter a more enlightened state.” as MIT Professor of Biological Engineering Ed Boyden previously put it.

Ziva Dynamics

Lastly, while algorithms are playing a major role in recent advancements in the Data & AI field, let’s not forget that digitalization and automatization of the world is happening given that software integrates itself perfectly with hardware. More specifically, significant improvements in computational power have led to the emergence of several startups who are shaking-up long established paradigms: processing in memory solutions for datacenters as well as cloud data centre storage solutions and quantum computing. UPMEM, a French startup founded in 2015, is addressing the Memory Wall issue by developing the first scalable, programmable and efficient processing in memory solution for datacenters. Their solution is to position hundreds of processing units as close to the data as possible to make the computation highly efficient both in terms of speed and energy. Then there’s Ugloo, who are tackling the cloud data centres storage solution by developing a local and datacenter-less solution to ensure the safety and governance of our data. And finally, Sparrow Quantum are working on quantum computing, developing and commercializing photonic quantum technology components based on research and patents developed in the Niels Bohr Institute Quantum Photonics Lab.

All in all, this year’s major trends in Data & AI are emphasizing the instantaneity of their solutions at a crucial time and bringing us to the edge of the next leap in AI and autonomous solutions. Deepmind’s AI technology paved the way when they beat the human champion of the game ‘Go’, but will there be another major improvement soon? Only one way to find out: 
Come to the Summit!

Discover all of these startups at the Hello Tomorrow Global Summit on October 26–27 in Paris!

Medium readers: Benefit from a 20% discount to attend the event using the code MEDIUM here:

Show your support

Clapping shows how much you appreciated Hello Tomorrow’s story.