From the Machine Learning Conference

We prepared a mini-conference on “machine learning” for our customers, students and the general public together with IBM, NVIDIA and DNS, which was held on Tuesday morning, 24 October 2017.  The conference took place on the grounds of the CIIRC at the CTU.   Seventy participants had the opportunity to see presentations on machine learning and artificial intelligence, the practical benefits they offer and, above all, the future implications of their use – as predicted by scientists and specialists from IBM and NVIDIA.

In the introduction, Petr Plodík from our company explained why it is now that we are talking about machine learning, why M Computers is organizing the conference and why it is being held on the grounds of the CIIRC at the CTU.   The introduction was followed by the presentations of guests from IBM and NVIDIA.

The History and Future of Machine Learning and Artificial Intelligence (AI)

The main star of the conference was Dr. Klaus Rindtorff, a scientist from the IBM Laboratory in Bölingen, Germany.  As early as the 1990s, Dr. Klaus had been using neural networks to recognize and authenticate signatures in a project for a banking institution.  In his presentation, he described the history of neural networks with alternating periods of interest and the progress and investments in these technologies in the “ice age”, in which the development of these technologies had frozen.  It was interesting to trace back the development of the last years, which was demonstrated with the ImageNetChallenge competition, which is organized every year.  It is an image recognition competition (150,000 images), in which the year-to-year increase in image recognition reliability by computers can be seen.  This year, the computer has even surpassed the success rate of a human,  and it has done so with an image processing speed exceeding 4000 images per minute on a single server.  At the conclusion of his lecture, Dr. Klaus explained how the Watson-IBM computer, which answers questions in natural language, works.

NVIDIA is introducing new features

Another guest was Ralph Hinsche from NVIDIA. He presented his vision of how machine learning and artificial intelligence are changing the world.   He mentioned the new technologies announced by NVIDIA two weeks ago at a conference in Munich.  He showed the listeners who were present at the conference the latest NVIDIA V100 accelerator, which has over 21 billion transistors, 5120 processors and a performance of 15.7 trillion operations with decimal numbers in a palm-sized box. He also presented the “Pegasus” Robotaxi drive PX, a computer based on artificial intelligence that is capable of driving a taxi (or so-called robo-taxi) at the highest level of the ASIL-D safety standard (learn more about the ASIL standard).

We recommend viewing the main talk of the GTC Europe 2017 conference by Jensen Huang, the founder and CEO of NVIDIA:

Practical demonstrations of machine learning

Radek Špimr from IBM had prepared a presentation and presented the OpenPOWER initiative and its applications in the world of machine learning.

In others parts of the conference, Milan Le from the Czech start-up Quantasoft and Radoslav Škoviera from ČVUT CIIRC presented real-life applications of neural network technology in their projects.

More information about the conference and downloadable presentations are available at: http://aidays.cz

We would like to give a big thanks to IBM, NVIDIA and DNS for their help and cooperation in the preparation and implementation of the conference:

 

            

 

Foto from the conference