October 2020 • Issue 1

NNAISENSE delivers advanced neural network solutions for industrial clients that improve how their processes and products work.

AI in Industrial Applications

Business benefits in real deployments, customer success stories
Interview with our customer and strategic investor SCHOTT AG

… the (NNAISENSE) team has impressive skills on the level of well-known companies such as DeepMind.

Dr. Jens Schulte
CFO Schott AG

NNN – Why is SCHOTT so interested in AI solutions?
Dr. Jens Schulte –  Schott’s key competitive differentiators are materials and process know-how. The efficiency and productivity of our production processes are essential, and AI will help us to achieve the next level of productivity in this area.

What are the use cases that you foresee, and what business benefits do you expect?
Our focus is on gradual automation of our core production processes,  particularly when it comes to melting of glass. The evolution of automation in this area is similar to autonomous driving levels of cars, i.e., from assisted driving to full automation. Just as with cars, it is a mid- to long-term journey. Besides, we look into material informatics and predictive maintenance applications. Since a large part of our cost base is related to production and development, we expect a significant benefit to our business.

Why did you choose NNN as your AI solution partner?
First, NNAISENSE has a focus on industrial processes. Second, based on a core backbone, the firm develops individual solutions and not off-the-shelf software. And third, the team has impressive skills on the level of well-known companies such as DeepMind.

Why did you also become a strategic investor?
We see the financial upside and can participate in new AI technologies first hand – a twofold advantage for us.

What are the current results of the implementation, and what next steps are you expecting?
A first user assistance system (i.e., automation level 1) has been implemented and is currently being tested on one of our larger production aggregates. The next steps would be a combination with further data or/and an enlargement of the training data set with simulations.

In general, how do you see the state of the German high-tech industry, and AI adoption in particular, compared with the US and China?
I believe the answer always depends on the specific area and use case. Much has been said and written about Germany (and Europe) falling behind the US and China concerning consumer platforms, social networks, and cloud technology. On the other hand, I would see us well positioned in areas closer to “German engineering” such as digital industrial clouds (Siemens), business software (SAP), or process mining (Celonis). Today, these always include an AI element as well. All in all, if we continue to focus on what we can do best, I am sure Germany will continue to play a role in such high tech / AI areas as we advance.

Updates from the Research Team

New patent filed

NNAISENSE has filed a patent for a new approach to drug and pesticide development which could aid in treatment of COVID 19. The proposed framework uses Graph Neural Networks to predict chemical properties and then discover molecules that maximize those properties. The AUC scores show excellent discrimination (>0.8). Applications are in the Pharma and Agrochem industries, and we are open to discuss the commercial application of this work with interested parties.

Poster talk at PPSN2020

ClipUp: A simple and powerful optimizer for distribution-based policy evolution. Classical momentum-based gradient ascent enhanced with two simple techniques: gradient normalization and update clipping. We argue that the resulting optimizer, called ClipUp (short for “clipped updates”), is a better choice for distribution-based policy evolution because its working principles are simple and easy to understand and its hyperparameters can be tuned more intuitively in practice.
View Paper

AI to optimize datacenter energy consumption

NNAISENSE has analyzed DeepMind’s efforts to optimize energy efficiency at Google’s datacenters and is ready to apply this know-how.
Contact us to learn how you can potentially save up to 40% in cooling energy cost in your datacenter using AI.

Marco Gallieri’s talk on “Safe AI with control theory”

Dr. Gallieri, Senior Researcher at NNAISENSE, overviews how we are combining neural networks and control theory at the Second Symposium on “Machine Learning and Dynamical Systems” organized by The Fields Institute for Research in Mathematical Sciences
Watch Video

AI Industry News

IoT, Industry 4.0, manufacturing, industrial processes

What impact will COVID-19 have on AI projects and deployments?


TRUMPF Group becomes latest NNAISENSE partner and strategic investor

NNAISENSE welcomes TRUMPF as its newest strategic partner and investor. We very much look forward to collaborating with this innovation leader in sheet metal fabrication machinery and industrial lasers. NNAISENSE will support TRUMPF in their latest laser technologies, applying our AI expertise to create industry-leading solutions.

The TRUMPF Group has 77 subsidiaries and branch offices and more than 13,400 employees, with sales in excess of $4.3 billion.

New team member

… I’ve followed the progress of NNAISENSE over the years and now seems to be the ideal time to come on board.

Ralf Haller
EVP Sales & Marketing at NNAISENSE

NNN : Hi Ralf, welcome to NNAISENSE. Tell us how you came to us after many years as an independent business development and marketing consultant?
Ralf Haller: I know IDSIA from a very successful event I did about five years ago in Lugano, “Silicon Valley meets Switzerland”, where we had 600 attendees and also all the local (incl. TV) media present.  Many customer projects were presented there and we had a 1:1 talk with Prof. Jürgen Schmidhuber as well. I was one of the first to organize machine learning tech business events in Switzerland. So, I’ve followed the progress of NNAISENSE for over the years and now seems to be the ideal time to come on board as you have a proven platform that is used by industrial clients like Schott, Repsol, and others and scaling is possible.

What qualifies you to take on this role in an AI engineering team?  Not afraid of all the deep AI expertise?
Haha, no not afraid at all. I worked in Silicon Valley for five years and I’m used to engaging with very technical expert teams. Also, I’m an electrical engineer myself and have worked for 25 years all over the word, always in high-tech.  For nine years now, I’ve been involved around 200 high-tech trend events where AI startups and solutions have played a role. Math and physics are a hobby of mine, not kidding, which helps me to focus and relax, and I truly enjoy it.  I even took machine learning classes and obtained a range of AI/ML/DL/RL certificates from Stanford University and other think-tanks.  In short, I think I have a good enough understanding of what’s behind all this, and, most importantly, what business applications, opportunities and challenges there are.

What are your immediate to-dos?
Lots of things on my plate. Existing clients and prospects have to be looked at, we will start (this) newsletter to stay in touch with all of our community. And, of course new business is to be developed, and strategic partners have to be found. Plus, some strategic things and introduction of a new CRM tool among many other things.

NNAISENSE celebrated its 6-year anniversary on Sept 8, 2020

As some folks said, we’re now ready to go to school and can wear long pants.
Interview with NNAISENSE CEO, Dr. Faustino “Tino” Gomez

I want NNAISENSE to become the premier provider of intelligent automation solutions that use the very latest advances in the field of machine learning.

Dr. Faustino Gomez
Co-Founder and CEO at NNAISENSE

NNN: How were the first 5 years up to end of 2019?
Dr. Faustino Gomez:  The first 5 years were an exciting journey starting with the founders transitioning from their former lives at IDSIA to the unfamiliar world of private enterprise.  It hasn’t always been easy, but over this period we’ve been able to assemble an elite team, and successfully raise two rounds of funding that have given us the bandwidth to both serve our great customers in building industry-first solutions and continue our long track record of applied research.   Highlights from this period include our public demos with Audi (NeurIPS 2016), with FESTO (Hanover Messe 2019), winning the “Learning to Run” Deep Reinforcement Learning competition (NeurIPS 2017), and our groundbreaking work with Schott and EOS. Because of these achievements, we now have a much better understanding of the market for industrial AI, and the track record to more easily open doors and encourage the adoption of this technology on a larger scale.

What happened in 2020 then?
The year started well: we made both the CBInsights AI100 and the Forbes AI DACH30 in March.  Obviously, the rest of the year has been challenging for everyone.  Despite the economic downturn, we’ve been able to weather the crisis and use it as an opportunity to sharpen our focus.  The pandemic is probably far from over, but we can already see that we will emerge stronger, and the addition of Mr. Haller (EVP Sales & Marketing) to the team is a big part of our strategy to hit the ground running as things pick up.

Has Switzerland been a good location for this startup? Would one not rather find such talent and company in Silicon Valley?
As a European company with an emphasis on the European market, Switzerland is a great strategic location to base our operations.  Apart from the favorable business environment, it’s a very enticing setting to attract and retain the scarce expert talent which is the key asset in this business. 

What are some of the key benefits of working with NNAISENSE to implement AI?
I think the main strength we bring to table is the caliber of our team which consists primarily of top researches that have not only made significant scientific contributions, but also have a predilection for how to apply the state-of-the-art to challenging problems in the real world. Additionally, because we work with a variety of companies, focusing on certain types of applications rather than on specific industrial sectors, our customers benefit from the know-how we’ve accumulated from working with diverse data sourced from diverse industries that would be not be available to them in-house.

What is your vision for the company in the medium to long term?
I want NNAISENSE to become the premier provider of intelligent automation solutions that use the very latest advances in the field of machine learning. That will require increasing our capacity by expanding the team and continuing to build on our track record of successful deployments. The more solutions we have out in the field demonstrating the value of adopting AI, the more we can advance the technology through lessons learned, and scale the business by requiring ever diminishing levels of customization per customer.

Jürgen Schmidhuber interviewed by Daniel Fagella of Emerj

Dr. Jürgen Schmidhuber, Co-founder and Chief Scientist at NNAISENSE explores the potential use-cases of AI for manufacturing applications, including robotic dexterity, multi-use robots, and more.

Listen to podcast interview

Tell us what you do

Piazza Molino Nuovo 17
6900, Switzerland
1224 East 12th St., suite 313 Texas, 78702, USA
trusted by innovators