TY - CHAP AU - Bártfai, Gusztáv ED - IEEE, Power Electronics Society TI - An Adaptive Resonance Theory-based Neural Network for Autonomous Learning via Iterative Knowledge Redescription T2 - Proceedings of the International Joint Conference on Neural Networks PB - Institute of Electrical and Electronics Engineers (IEEE) CY - New York, New York SN - 9780738133669 T3 - ; 2021-July. PY - 2021 PG - 8 DO - 10.1109/IJCNN52387.2021.9533351 UR - https://m2.mtmt.hu/api/publication/32813602 ID - 32813602 N1 - Funding Agency and Grant Number: Faculty of Information Technology and Bionics at Pazmany Peter Catholic University Funding text: I wish to thank the Faculty of Information Technology and Bionics at Pazmany Peter Catholic University for their support. AB - This paper introduces a neural network architecture called ARTIR, which autonomously learns classifications of arbitrary input sequences via iterative redescription of knowledge that it builds up during learning. Knowledge redescription is achieved via changes to both the network architecture and the connection weights between neurons. Based on ARTMAP, the supervised version of the Adaptive Resonance Theory (ART) neural network, and motivated by the Representational Redescription hypothesis in cognitive science, the ARTIR network creates new, successively deeper internal features iteratively. The process is driven by the complexity of the classification task as measured by the number of input categories the ARTMAP network develops during training. The network is forced to discover a new internal feature that better discriminates the output classes every time the currently perceived difficulty of the learning task exceeds a certain limit as indicated by the number of input categories the ARTMAP network develops in response to an input-target sequence. Once a new candidate feature - found to correlate well enough with an output class - is accepted, ARTIR re-learns the mapping using the additional feature and uncommits its categories that are no longer needed as a result of re-learning. The iterative learning process then continues where the newly created feature, along with all original inputs and internal features created during previous stages of learning, becomes subject of internal knowledge redescription for future iterations. Experiments carried out on two binary decision problems demonstrate the viability of the approach, and the results point to further questions that will be the subject of future research. LA - English DB - MTMT ER - TY - JOUR AU - Rekeczky, Csaba AU - Bártfai, Gusztáv AU - Bob, Siller TI - FPGAs bringen Vorteile gegenüber DSPs JF - ELEKTRONIK INDUSTRIE J2 - ELECTRON IND VL - 2012 PY - 2012 IS - 6 SP - 18 EP - 21 SN - 0174-5522 UR - https://m2.mtmt.hu/api/publication/3003486 ID - 3003486 LA - German DB - MTMT ER - TY - JOUR AU - Argyros, A AU - Bártfai, Gusztáv AU - Eitzinger, C AU - Kemény, Zsolt AU - Csáji, Balázs Csanád AU - Kék, László AU - Lourakis, M AU - Reisner, W AU - Sandrisser, W AU - Sarmis, T AU - Umgeher, G AU - Viharos, Zsolt János TI - Smart sensor based vision system for automated processes. JF - INTERNATIONAL JOURNAL OF FACTORY AUTOMATION ROBOTICS AND SOFT COMPUTING J2 - INT J FACT AUTOM ROBOT SOFT COMPUT VL - 3 PY - 2007 SP - 118 EP - 123 PG - 6 SN - 1828-6984 UR - https://m2.mtmt.hu/api/publication/165614 ID - 165614 LA - English DB - MTMT ER - TY - CHAP AU - Argyros, A AU - Bártfai, Gusztáv AU - Eitzinger, C AU - Kemény, Zsolt AU - Csáji, Balázs Csanád AU - Kék, László AU - Lourakis, M AU - Reisner, W AU - Sandrisser, W AU - Sarmis, T AU - Umgeher, G AU - Viharos, Zsolt János ED - Pennacchio, S TI - Smart sensor based vision system for automated processes. T2 - Emerging technologies, robotics and control systems. Vol. 2 PB - Internationalsar CY - Palermo PY - 2007 SP - 24 EP - 29 PG - 6 UR - https://m2.mtmt.hu/api/publication/165481 ID - 165481 LA - English DB - MTMT ER - TY - JOUR AU - Földesy, Péter AU - Kék, László AU - Zarándy, Ákos AU - Roska, Tamás AU - Bártfai, Gusztáv TI - Fault-tolerant design of analogic CNN templates and algorithms - Part I: The binary output case JF - IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS I - FUNDAMENTAL THEORY AND APPLICATIONS J2 - IEEE T CIRC SYST FUND VL - 46 PY - 1999 IS - 2 SP - 312 EP - 322 PG - 11 SN - 1057-7122 DO - 10.1109/81.747209 UR - https://m2.mtmt.hu/api/publication/1185448 ID - 1185448 AB - This paper addresses the issue of designing a class of fault-tolerant cellular neural network (CNN) templates that, combined with CNN analogic algorithms, work correctly and reliably on given CNN universal machine (CNN-UM) chips. In particular, a generic method for finding nonpropagating binary-output CNN templates is proposed, This method is based on measurements of actual CNN-UM chips and combines adaptive optimization and decomposition of theoretically ideal CNN templates in order to correct the erroneous behavior of actual CNN-UM chips, which is mainly caused by imperfections introduced during fabrication, More specifically, the entire array of cells in a CNN-UM chip is modeled by a single feed-forward virtual cell whose optimal parameters are found by a simple and effective gradient-based method. In the case of binary input-output uncoupled templates (or Boolean operators), a systematic template decomposition method is applied whenever optimization fails to find a correctly working CNN template for all possible combinations of local 3 x 3 binary input patterns. The resulting templates are finally combined, yielding a simple CNN analogic algorithm. Examples are presented for both binary- and analog-input operators, using two concrete stored-program CNN-UM chips to demonstrate the effectiveness of the proposed method, whose advantages and limitations are also discussed. LA - English DB - MTMT ER - TY - THES AU - Bártfai, Gusztáv TI - Towards implementation of biologically inspired autonomous and adaptive information processing systems PY - 1999 SP - 106 UR - https://m2.mtmt.hu/api/publication/161879 ID - 161879 LA - English DB - MTMT ER - TY - GEN AU - Bártfai, Gusztáv TI - Combination of adaptive resonance theory (ART) and cellular neural network (CNN) chips into a high-speed pattern recognition system. PY - 1999 SP - 1 EP - 16 PG - 16 UR - https://m2.mtmt.hu/api/publication/161870 ID - 161870 LA - English DB - MTMT ER - TY - CHAP AU - Földesy, Péter AU - Kék, László AU - Roska, Tamás AU - Zarándy, Ákos AU - Bártfai, Gusztáv ED - Tavsanoglu, V TI - Fault tolerant CNN template design and optimatization based on chip measurements T2 - 1998 fifth IEEE International Workshop on Cellular Neural Networks and Their Applications PB - IEEE CY - Piscataway (NJ) SN - 9780780348684 PY - 1998 SP - 404 EP - 409 PG - 6 DO - 10.1109/CNNA.1998.685415 UR - https://m2.mtmt.hu/api/publication/161590 ID - 161590 LA - English DB - MTMT ER - TY - GEN AU - Földesy, Péter AU - Kék, László AU - Zarándy, Ákos AU - Roska, Tamás AU - Bártfai, Gusztáv TI - Fault tolerant design of analogic CNN templates and algorithms. Part I: The binary output case. PY - 1998 SP - 1 EP - 18 PG - 18 UR - https://m2.mtmt.hu/api/publication/161429 ID - 161429 LA - English DB - MTMT ER - TY - JOUR AU - Roska, Tamás AU - Bártfai, Gusztáv AU - Szolgay, Péter AU - Szirányi, Tamás AU - Radványi, András AU - Kozek, T AU - Ugray, Zs AU - Zarándy, Ákos TI - A digital multiprocessor hardware accelerator board for Cellular Neural Networks: CNNHAC JF - INTERNATIONAL JOURNAL OF CIRCUIT THEORY AND APPLICATIONS J2 - INT J CIRC THEOR APP VL - 20 PY - 1992 IS - 5 SP - 589 EP - 601 PG - 13 SN - 0098-9886 DO - 10.1002/cta.4490200512 UR - https://m2.mtmt.hu/api/publication/1001439 ID - 1001439 LA - English DB - MTMT ER -