Nicolas Corsi, CEO of ViDi Systems, is a visionary who has built a business around Deep Learning and Artificial Intelligence. Nicolas takes world class people to get a world class company selling world class products. The venture started four years ago when he met with Reto Wyss, a Ph.D. in computational neuroscience. It turned out that Reto had cracked the well-hidden tricks nature uses to process and reason on visual stimuli. Very quickly the idea to start a business to bring Reto’s discoveries to the world emerged. Now, their company with its products enables computers, machines, and cars to understand images so they can interact with the reality. ViDi’s Deep Learning Software autonomously learns from images, leading to decisions with superhuman performances, without any programming. Nicolas is fully entitled to brag, and he says: “Our technology has the potential to make a dent in the world.” He believes that failure never becomes an obstacle; instead, they accelerate growth and learning.
Nicolas has 20 years’ of professional experience in High-Tech Startups, Business Development, M&A, Product Management, B2B Sales, and Marketing. Nicolas has an-active role in the establishment of 8 Sales Subsidiaries in Europe, North America, and Asia. He Co-founded four new companies, investment or acquisition into four companies; one is fully exit. Since 2010, he is a CBDO CPA Group and board member of 7 high-tech companies. In the year of 2003-2010, Nicolas played a significant role in Sales and Marketing Director ETEL SA, Switzerland (Heidenhain GmbH Group). In 2000-2003, he was a Managing Director of ETEL USA, Chicago. Nicolas earned his B.Sc in Electronics and Embedded Microsystems from the University of Applied Sciences, Western Switzerland and later went on to pursue Executive MBA in Management of Technology from EPFL, Switzerland.
“One of our customer, a researcher active in the field of oncology, told us that thanks to our software he can now accomplish in a single day a painstaking task that would have taken him 6-8 weeks, hence helping him to greatly speed up his research, ultimately making the world a better place,” reveals Nicolas.
ViDi Systems: Presenting Infinite Possibilities with Artificial Intelligence
ViDi Systems develops and commercializes the most advanced Artificial Intelligence software for Image Analysis. Building on 10-year Machine Learning and Computational Neuroscience Research, they invented a truly unique Active Deep Learning Architecture able to process and reason on images without any programming; instead of manually defining a set of logical rules and then tediously adjusting many parameters for each vision problem, they let the computer learn from images all by itself.
Their software is already revolutionizing the Factory Automation market where it routinely performs very difficult visual inspection tasks that were up to now impossible to automate such as finding the faintest optical defects on highly variant surfaces such as non-woven textiles, decorated watch parts or machined automotive components. Hundreds of successful feasibilities show that technology scales very well also outside of factories. That is why ViDi tackles now large-potential markets such as Medical Imaging, Driving Assistance, Aerial Image Analysis, Infrastructure Monitoring and Surveillance.
ViDi’s Technology: “Magnum Opus” of Artificial Intelligence Technology
The notion behind ViDi’s line of work is that to interact with the world in a meaningful way; machines must first understand images. ViDi provides the missing technology for digital creations to make sense of visual stimuli so that the clients can cope with real-life situations. Traditional computer vision solutions are very limited in performance and can hardly manage changing or unpredictable environments. Mainstream, Machine Learning solutions rely mostly on Deep Convolutional Neural Networks (ConvNets) which require highly supervised learning, huge training datasets and are very demanding in term of computing power. This has so far limited the commercial success of Machine Learning based products to a few applications, mainly speech and object recognition, search engines, and content filtering. The gap between what can be done with AI in the lab and what is actually done in the real-world applications is huge. ViDi bridges that gap, allowing technologists across multiple domains to create groundbreaking products for real people.
Nicolas elaborated this technology in detail which includes:
Active Deep Learning
Human vision is an active process that sequentially samples the scenery in an intelligent, task-specific way using a small, high-resolution fovea with a large, low-resolution surround. Based on the latest understanding of the well-hidden tricks nature uses to process and reason on visual stimuli, they developed a novel way to analyze images actively with focus and attention. This radical advancement in Machine Learning has been successfully implemented into ViDi Core, a powerful yet nimble AI engine.
Supervised and Unsupervised Learning
At ViDi, they do not oppose supervised and unsupervised learning. Human learning is largely unsupervised. However, certain things are faster and better learned with a few examples and good direction from an expert. Unlike other dogmatic solutions, ViDi offers a full range of possibilities, from highly supervised to completely unsupervised learning. Because one-fits-all approaches are just unrealistic.
A Practical Set of Tools
Analyzing hundreds of real vision cases, they found out that they all could be solved by decomposing them into only three base tasks. And by combining these, very complex challenges can be easily splintered. They encapsulated these base tasks into three software tools: ViDi blue to detect and localize features; ViDi red to find anomalies and segment images; and ViDi green to classify the images.
Bright Future Upstream
Initial customers now validate ViDi’s technology, their product is used by an increasing number of happy customers, and its value is recognized. They see an excellent growth of their turnover. Also, large, long-term contracts are currently being negotiated. In the next two years, the challenge will be to transform the company so that it can sustain the growth long-term. Expanding their sales, marketing, and support capability worldwide will be one of the challenges, as well as keeping the technological edge to enter new market verticals. In the same time, they should never lose focus while taking the good opportunities that are constantly coming along.