PANN™ technology
PANN™ (Progressive Artificial Neural Network) is a patented technology based on a fundamentally new neural network architecture. This technology is based on the latest knowledge in the field of biology and is closer in its principles to biological networks than any AI technology existing today. Based on PANN™ technology, we are creating a new generation neural network and AI infrastructure with fundamentally different qualities and characteristics. In the scientific articles attached below, we describe the fundamental functioning principles of this new type of neuron:
Main Principles of the General Theory of Neural Network with Internal Feedback
Analog and Digital Modeling of a Scalable Neural Network
Fundamentally new quality
PANN™ technology can radically improve the neural network performance. With data volume increase, the need for computing power and training time grows linearly, not exponentially, as in conventional neural networks. This means that the time required to train the neural network, energy consumption, costs and human resource needs are reduced. Thus, intelligence in the right proportion can be added everywhere: to any product, software, gadget or infrastructure element.
Problems solved by PANN™
PANN™ enables:
- High speed training
- High accuracy
- Simple training algorithm
- Easy realization in analog or digital form
- Scalability – the possibility of building ANNs of any size and complexity
- Introduction of additional images into trained network without the necessity of retraining the entire ANN (prevents loss of previously accumulated information)
- Reliable, without network paralysis
Consequently, PANN™ realizes the advantages of ANNs predicted by science.
Quantum leap in AI:
PANN™ technology eliminates the ‘inherent defect’ and limitations of classical neural networks and opens up new horizons in the development of AI applications.
- For everyone: to use
- Computing requires less sophisticated hardware.
- Calculations can be performed at user’s own facilities and guarantee the safety and confidentiality of data.
- Affordable for startups and small businesses.
- Compatible with various hardware: from microchips, video cards, and smartphones to supercomputers; boosts hardware computing capabilities by a multiple.
- Low costs:
- Drastic reduction in the cost of infrastructure and energy consumption.
- Less human resource investment due to a sharp acceleration of network training and ease of maintenance.
- Innate high security and reliability.
- Unparalleled performance:
- Hundreds and thousands of times faster training and network speed: the advantage increases with an increase in task complexity and data volume.
- Up-training ability.
- Almost zero probability of ‘freezing’ with an increase in task complexity and network size.
- Opportunities to scale the network both in the training process and in the process of operation.
22nd ETRIA World Conference TRIZ Future 2022 “Systematic Innovation Partnerships with AI and IT”
Vladimir Proseanic (CEO of Progress, Inc. and Omega Server, Inc.) and Boris Zlotin (Chief Scientist of Progress, Inc. and Omega Server, Inc.) were among the key speakers of the 22nd ETRIA World Conference TRIZ Future 2022 “Systematic Innovation Partnerships with AI and IT” that took place at the Warsaw University of Technology (Warsaw, Poland). They presented their work in “TRIZ and Progressive Artificial Neural Networks (PANN™)”.
The conference is organized by The European TRIZ Association (ETRIA) together with the Faculty of Electronics and Information Technology (FEIT), Warsaw University of Technology. The TRIZ Future conference invites original papers, both scientific and practitioners’, that combine systematic invention generation, creative design, and digital technologies to solve problems in any field of human activity. This year’s edition focuses on Artificial Intelligence methods and Information Technology, which are becoming more ubiquitous and more important every year.
The conference aims at linking organizations and individuals from the industry, research, and education worlds to share their experiences on systematic innovation and promote TRIZ-based tools worldwide. It will provide an international forum to exchange new ideas on knowledge-based innovation, present recent achievements in this area, and strengthen collaboration between academic and industrial players.