ارائه اپراتور جدید جایگزین پخش قطره جوهر در روش یادگیری فعال

نوع مقاله : علمی-پژوهشی

نویسندگان

1 دانشکده مهندسی برق - دانشگاه صنعتی شریف

2 دانشکده فنی و مهندسی - دانشگاه مازندران

چکیده

روش یادگیری فعال یکی از روش‌های یادگیری فازی است که  الگو­گرفته از پردازش در مغز انسان است. اپراتور پخش قطره جوهر، موتور اصلی پردازشی در این روش است که به دور از فرمول­های پیچیده، به­دنبال یافتن رابطه بین خروجی و هر یک از ورودی‌ها است. زیاد­بودن حافظه مورد نیاز برای پیاده‌سازی صفحات پخش قطره جوهر و همچنین حجم محاسبات زیاد لازم برای استخراج ویژگی‌ها از جمله مشکلات پیش روی اپراتور پخش قطره جوهر است. در این مقاله یک روش جایگزین برای اپراتور پخش قطره جوهر ارائه شده است که سبب کاهش چشم‌گیر پیچیدگی محاسباتی می‌شود. الگوریتم ارائه­شده با استفاده از دو بردار حافظه به توصیف صفحات پخش جوهر می‌پردازد که مشکل اتلاف زیاد حافظه را حل می‌کند. الگوریتم پیشنهادی، ویژگی‌های مسیر باریک و پراکندگی داده در صفحات پخش قطره جوهر را که مهم‌ترین مفاهیم برای استفاده در مرحله استنتاج الگوریتم یادگیری فعال هستند را با صرف کمترین هزینه و زمان محاسباتی می‌یابد. برای بررسی صحت عملکرد الگوریتم، شبیه‌سازی‌هایی بر روی مجموعه داده‌های استاندارد در حوزه مدل‌سازی و طبقه‌بندی ارائه شده است. زمان و دقت  الگوریتم پیشنهادی با روش یادگیری فعال، شبکه‌های عصبی چندلایه پرسپترون و سیستم استنتاج عصبی-فازی تطبیق‌پذیر مقایسه شده است.

کلیدواژه‌ها


عنوان مقاله [English]

A Novel Alternative Operator for Ink Drop Spread (IDS) in Active Learning Method (ALM)

نویسندگان [English]

  • S. Haghzad Klidbary 1
  • S. Bagheri Shouraki 1
  • I. Esmaili Paeen Afrakoti 2
1 Faculty of Electrical Engineering, Sharif University of Technology, Tehran, Iran
2 Departments of Engineering and Technology, University of Mazandaran, Babolsar, Iran
چکیده [English]

Active Learning method is one of the fuzzy learning methods inspired from the computation of human brain. Ink drop spread operator is the main computational engine in ALM which without using any complex formula, seeks for the relationship between the output and the inputs of the system. One of the challenges of IDS operator is that not only a large memory is required for implementing IDS planes, but also extracting features imposes high computational costs. In this paper, one learning method, as a replacement for IDS operator, is represented that considerable reduces the computational complexity. The represented algorithm defines the IDS planes with only two memory vectors and solves the problem of huge wastage of memory in these planes. This algorithm starts to learn the available patterns in learning data to find two features of Narrow p < /em>ath and Spread in planes which are the most important concepts in the level of active inference learning with minimum computational and time cost. To investigate the accuracy of algorithm’s performance, some simulations in modelling and classification have been done on standard data sets. Time and accuracy of proposed algorithm is compared with traditional ALM, MLP and ANFIS algorithms.

کلیدواژه‌ها [English]

  • Brain learning simulation
  • active learning method
  • ink drop spread operator
  • fuzzy inference system
  • pattern classification
[1]      L. A. Zadeh, Fuzzy sets, Information and control, vol. 8, no. 3, pp. 338-353, 1965.
[2]      Tanaka, K. and H.O. Wang, Fuzzy control systems design and analysis: a linear matrix inequality approach, John Wiley & Sons, 2004.
[3]      پوریا جعفری، محمد تشنه‌لب، مهسان توکلی کاخکی، "طراحی کنترل‌کننده فازی تطبیقی مستقیم برای سیستم‌های مرتبه کسری غیرخطی به کمک جبران‌ساز"، مجله مهندسی برق دانشگاه تبریز، جلد 47، شماره 3، 1396.
[4]      Mamdani, E.H., Application of fuzzy algorithms for control of simple dynamic plant, Electrical Engineers, Proceedings of the Institution of, vol. 121, no. 12, pp. 1585-1588, 1974.
[5]      Takagi, T. and M. Sugeno, Fuzzy identification of systems and its applications to modeling and control, IEEE transactions on systems, man, and cybernetics, vol. 1,  pp. 116-132, 1985.
[6]      Terano, T., K. Asai, and M. Sugeno, Applied fuzzy systems, Academic Press, 2014.
[7]      ندا خانبانی، امیر مسعود افتخاری مقدم، "ارائه یک روش تشخیص زبان علامت مبتنی بر رویکرد MLRF فازی با استفاده از اطلاعات عمق تصویر"، مجله مهندسی برق دانشگاه تبریز، جلد 47، شماره 3، 1396.
[8]      سمیرا رفیعی، پرهام مرادی، "بهبود عملکرد الگوریتم خوشه‌بندی فازی سی- مینز با وزن‌دهی اتوماتیک و محلی ویژگی‌ها"، مجله مهندسی برق دانشگاه تبریز، جلد 46، شماره2، 1395.
[9]      Bezdek, J.C., Fuzzy mathematics in pattern classification, 1973.
[10]      Del Campo, I., et al, Efficient hardware/software implementation of an adaptive neuro-fuzzy system, IEEE Transactions on Fuzzy Systems, vol. 16, no. 3, pp. 761-778, 2008.
[11]      Himavathi, S., D. Anitha, and A. Muthuramalingam, Feedforward neural network implementation in FPGA using layer multiplexing for effective resource utilization, IEEE Transactions on Neural Networks, vol. 18, no. 3, pp. 880-888, 2007.
[12]      Soudry, D., et al., Memristor-based multilayer neural networks with online gradient descent training, IEEE transactions on neural networks and learning systems, vol. 26, no. 10, pp. 2408-2421, 2015.
[13]      Ortega-Zamorano, F., et al., Efficient implementation of the backpropagation algorithm in fpgas and microcontrollers, IEEE transactions on neural networks and learning systems, vol. 27, no. 9, pp. 1840-1850, 2016.
[14]      Shouraki, S.B., A novel fuzzy approach to modeling and control and its hardware implementation based on brain functionality and specifications, 2000.
[15]      Sugeno, M. and T. Yasukawa, A fuzzy-logic-based approach to qualitative modeling, IEEE Transactions on fuzzy systems, vol. 1, no. 1,  pp. 7-31, 1993.
[16]      Murakami, M. and N. Honda, A study on the modeling ability of the IDS method: A soft computing technique using pattern-based information processing, International journal of approximate reasoning, vol. 45, no. 3, pp. 470-487,2007.
[17]      Sakurai, Y., A study of the learning control method using PBALM-a nonlinear modeling method, PhD, The University of Electro-Communications, Tokyo, 2005.
[18]      Shouraki, S.B. and N. Honda. Fuzzy controller design by an active learning method, in 31th Symposium of Intelligent Control. Tokyo, Japan. 1998.
[19]      Ghorbani, M.J., M.A. Choudhry, and A. Feliachi. Distributed multi-agent based load shedding in power distribution systems, in Electrical and Computer Engineering (CCECE), 2014 IEEE 27th Canadian Conference on. 2014.
[20]      MURAKAMI, M., Practicality of modeling systems using the IDS method: Performance investigation and hardware implementation, The University of Electro-Communications, 2008.
[21]      Firouzi, M., S.B. Shouraki, and J. Conradt. Sensorimotor Control Learning Using a New Adaptive Spiking Neuro-Fuzzy Machine, Spike-IDS and STDP, in International Conference on Artificial Neural Networks, Springer, 2014.
[22]      Cranganu, C. and F. Bahrpeyma, Use of active learning method to determine the presence and estimate the magnitude of abnormally pressured fluid zones, a case study from the Anadarko Basin, Oklahoma, in Artificial Intelligent Approaches in Petroleum Geosciences, Springer. pp. 191-20, 2015.
[23]      Bahrpeyma, F., C. Cranganu, and B.Z. Dadaneh, Active learning method for estimating missing logs in hydrocarbon reservoirs, in Artificial Intelligent Approaches in Petroleum Geosciences, Springer. pp. 209-224, 2015.
[24]      Merrikh-Bayat, F., F. Merrikh-Bayat, and S.B. Shouraki, The neuro-fuzzy computing system with the capacity of implementation on a memristor crossbar and optimization-free hardware training, IEEE Transactions on Fuzzy Systems, vol. 22, no. 5, pp. 1272-1287, 2014.
[25]      Bahrpeyma, F., A. Zakerolhoseini, and H. Haghighi, Using IDS fitted Q to develop a real-time adaptive controller for dynamic resource provisioning in Cloud's virtualized environment, Applied Soft Computing 26, pp. 285-298, 2015.
[26]      Klidbary, S.H., et al. Outlier Robust Fuzzy Active Learning Method (ALM), in 7th International Conference on Computer and Knowledge Engineering (ICCKE 2017), October 26-27 2017, Ferdowsi University of Mashhad. 2017.
[27]      Shouraki, S.B., N. Honda, and G. Yuasa, Fuzzy interpretation of human intelligence, International Journal of Uncertainty, Fuzziness and Knowledge-based Systems, vol. 7, no. 4, pp. 407-414, 1999.
[28]      Firouzi, M., S.B. Shouraki, and I.E.P. Afrakoti, Pattern analysis by active learning method classifier, Journal of Intelligent & Fuzzy Systems, vol, 26, no. 1, pp. 49-62, 2014.
[29]      Javadian, M., S.B. Shouraki, and S.S. Kourabbaslou, A novel density-based fuzzy clustering algorithm for low dimensional feature space, Fuzzy Sets and Systems, 2016.
[30]      Afrakoti, I.E.P., S.B. Shouraki, and B. Haghighat, An Optimal Hardware Implementation for Active Learning Method Based on Memristor Crossbar Structures, IEEE Systems Journal, vol. 8, no. 4, pp. 1190-1199, 2014.
[31]      Shouraki, S.B. and N. Honda, Recursive Fuzzy Modeling Based on Fuzzy Interpolation, JACIII, vol. 3, no. 2, pp. 114-125, 1999.
[32]      Firouzi, M. and S.B. Shouraki, Performance Evaluation of Active Learning Method in Classification Problems, 3rd International Conference on Machine Learning and Computing, ICMLC. 2011.
[33]      Bishop, C.M., Pattern recognition, Machine Learning, 128, 2006.
[34]      Rini, D.P., S.M. Shamsuddin, and S.S. Yuhaniz, Particle swarm optimization for ANFIS interpretability and accuracy, Soft Computing, vol. 20, no. 1, pp. 251-262, 2016.
[35]      B. Gliwa, and A. Byrski, Hybrid neuro-fuzzy classifier based on NEFCLASS model, Computer Science 12, pp. 115, 2013.
[36]      I. E. P. Afrakoti, et al., Using a memristor crossbar structure to implement a novel adaptive real-time fuzzy modeling algorithm, Fuzzy Sets and Systems, 307, pp.115-128, 2016.