Intelligent Parameter Tuning Using Segmented Recursive Reinforcement Learning Algorithm

  IJETT-book-cover  International Journal of Engineering Trends and Technology (IJETT)          
  
© 2020 by IJETT Journal
Volume-68 Issue-6
Year of Publication : 2020
Authors : Modalavalasa Hari Krishna, Dr.Makkena Madhavi Latha
  10.14445/22315381/IJETT-V68I6P201S

MLA 

MLA Style: Modalavalasa Hari Krishna, Dr.Makkena Madhavi Latha  "Intelligent Parameter Tuning Using Segmented Recursive Reinforcement Learning Algorithm" International Journal of Engineering Trends and Technology 68.6(2020):1-8. 

APA Style: Modalavalasa Hari Krishna, Dr.Makkena Madhavi Latha. Intelligent Parameter Tuning Using Segmented Recursive Reinforcement Learning Algorithm  International Journal of Engineering Trends and Technology, 68(6),1-8.

Abstract
Now-a-days, Machine Learning plays a vital role in enhancing the capabilities of traditional algorithms and processing techniques to handle profusion of data due to advances in digital technologies. Many processing problems can be solved using optimization-based solutions. In general, these solutions are normalized for different applications. Most of these solutions have control parameters to maintain relative importance to a specific application and also to optimize the performance of the solution. Parameter tuning is a straightforward task by manually determining the direction for adjustment of the parameter. But manual adjustment for these control parameters is tedious and consumes too much time and effort. Manual control becomes impractical if the solution has more parameters and requires precise tuning. With control being complex, machine learning has emerged as a key component in setting up correct and precise parameters. Reinforcement Learning (RL) can solve this problem but existing RL algorithms requires huge amount of learning time and resources. This paper aims at solving this problem and proposes novel Segmented and Recursive Reinforcement Learning (SRRL) algorithm to train the system that can automatically adjust the parameters accurately and precisely with minimal learning time. Performance of the proposing algorithm is validated in wavelet-based noise reduction technique by employing SRRL algorithm to adjust 3 control parameters of Noise based Hybrid Threshold method. After integrating with the proposing SRRL algorithm, the performance of the considered noise reduction technique is improved and provide better PSNR values with minimum learning time than with existing RL algorithms.

Reference

[1] Chenyang Shen, Yesenia Gonzalez, Liyuan Chen, Steve B. Jiang, Xun Jia, “Intelligent Parameter Tuning in Optimization-based Iterative CT Reconstruction via Deep Reinforcement Learning”, IEEE Transactions on Medical Imaging, doi: 10.1109/TMI.2018.2823679, 2018.
[2] Juan Cruz Barsce, Jorge A. Palombarini, Ernesto C. Mart´ınez , “Towards Autonomous Reinforcement Learning: Automatic Setting of Hyper-parameters using Bayesian Optimization”, doi : 978-1-5386-3057-0/17,IEEE,2017.
[3] M. Hari Krishna, G. Sateesh Kumar, “A Noise Based Hybrid Thresholding For Enhanced Noise Reduction Using Double Density Dual Tree Complex Wavelet Transform”, 13th International Conference on Electromagnetic Interference and Compatibility (INCEMIC), doi : 978-1-5090-5350-6/15, IEEE,2015.
[4] R. Saravanan, Pothula Sujatha, “A State of Art Techniques on Machine Learning Algorithms: A Perspective of Supervised Learning Approaches in Data Classification”, Second International Conference on Intelligent Computing and Control Systems (ICICCS-2018), IEEE, doi: 10.1109/ICCONS.2018.8663155,2018.
[5] Ahmad Maroof Karimi, Justin S. Fada, JiQi Liu, Jennifer L. Braid, Mehmet Koyutürk, Roger H. French, “Feature Extraction, Supervised and Unsupervised Machine Learning Classification of PV Cell Electroluminescence Images”, IEEE 7th World Conference on Photovoltaic Energy Conversion (WCPEC)-2018, IEEE Conferences, doi: 10.1109/PVSC. 2018.8547739, PP: 0418 – 04ADAPTIVE,2018.
[6] Snehasis Mukhopadhyay, Omkar Tilak, Subir Chakrabarti, “Reinforcement Learning Algorithms for Uncertain, Dynamic, Zero-Sum Games”, 17th IEEE International Conference on Machine Learning and Applications (ICMLA)-2018, IEEE Conferences, doi: 10.1109/ICMLA.2018.00015, PP: 48 – 54,2018.
[7] L. I. Rudin, S. Osher, and E. Fatemi, “Nonlinear total variation based noise removal algorithms”, Phys. D, Nonlinear Phenomena, vol. 60, nos. 1–4, pp. 259–268, Nov. 1992.
[8] Michael G. Pecht, Myeongsu Kang, “Prognostics and Health Management of Electronics: Fundamentals, Machine Learning, and the Internet of Things”, IEEE Books, Wiley-IEEE Press, doi: 10.1002/9781119515326, e-ISBN: 9781119515326.
[9] Parag Kulkarni, “Reinforcement and Systemic Machine Learning for Decision Making”, IEEE Books, Wiley-IEEE Press, doi: 10.1002/9781118266502, e-ISBN: 9781118266502, 2012.
[10] Oge Marques, “Practical Image and Video Processing Using MATLAB”, IEEE Books, Wiley-IEEE Press, doi: 10.1002/9781118093467, e- ISBN: 9781118093467, 2011.
[11] M.Waltz, K.Fu, “A heuristic approach to reinforcement learning control systems” IEEE Transactions on Automatic Control, DOI: 10.1109/TAC.1965.1098193, volume-10, Issue-4 , 1965.
[12] [12] V. Solo, "Selection of tuning parameters for support vector machines," Proceedings. (ICASSP '05). IEEE International Conference on Acoustics, Speech, and Signal Processing, 2005., Philadelphia, PA, 2005, pp. v/237-v/240 Vol. 5.
[13] E. B. Kosmatopoulos and A. Kouvelas, "Large Scale Nonlinear Control System Fine-Tuning Through Learning," in IEEE Transactions on Neural Networks, vol. 20, no. 6, pp. 1009-1023, June 2009.
[14] Jun Yu; Dacheng Tao, "Modern Machine Learning Techniques," in Modern Machine Learning Techniques and Their Applications in Cartoon Animation Research , , IEEE, 2013, pp.63-104.
[15] John D. Kelleher; Brendan Tierney, "4 MACHINE LEARNING 101," in Data Science, MITP, 2018, pp.97-150

Keywords
Intelligent Parameter Tuning, Noise based Hybrid Threshold, Machine Learning, Reinforcement Learning, Segmented and Recursive Reinforcement Learning.