Tế bào Nơron nhân tạo có độ chính xác và tốc độ cao
Main Article Content
Abstract
Bài báo này tập trung trình bày thiết kế tế bào nơron nhân tạo với phương pháp học giám sát có khả năng thích ứng với nhiều thuật toán đòi hỏi độ chính xác và tốc độ cao. Dựa trên thuật toán huấn luyện có giám sát và cấu tạo nơron thực, nhóm nghiên cứu xây dựng một kiến trúc nơron nhân tạo có kiến trúc tương tự đi kèm bộ xử lý số thực. Kiến trúc này dễ dàng tăng tốc độ xử lý bằng cách mở rộng số tầng thực hiện mô phỏng theo cấu trúc đường ống (pipeline). Để đảm bảo tốc độ và độ chính xác cao, nhóm nghiên cứu đã thực hiện tối ưu một số kiến trúc bộ dịch và bộ xử lý số thực song song. Chính vì vậy khi tăng thêm số tầng cho kiến trúc thì tốc độ tăng lên rất nhanh trong khi tài nguyên tăng lên không đáng kể. Kết quả tổng hợp trên chip FPGA Virtex 6 của hãng Xilinx cho thấy kiến trúc nơron của nhóm nghiên cứu đề xuất có thể hoạt động lên đến 5 tầng thực hiện theo cấu trúc pipeline và tốc độ đạt được tối đa là 108Mhz.
Keywords:
nơron nhân tạo, xử lý số thực, đường ống, bộ dịch.
References
[1] Sicheng Li, Chunpeng Wu, Helen, Boxun Li, Yu Wang, Qinru Qiu - “FPGA Acceleration of Recurrent Neural Network based Language Model”.
[2] Lei Liu, Jianlu Luo, Xiaoyan Deng, Sikun Li – “FPGA-based Acceleration of Deep Neural Networks Using High Level Method” - 2015 10th International Conference on P2P, Parallel, Grid, Cloud and Internet Computing.
[3] Eriko Nurvitadhi, Jaewoong Sim, David Sheffield, Asit Mishra, Srivatsan Krishnan, Debbie Marr - “Accelerating Recurrent Neural Networks in Analytics Servers: Comparison of FPGA, CPU, GPU, and ASIC”.
[4] Philippe Dondon, Julien Carvalho, Rémi Gardere, Paul Lahalle, Georgi Tsenov and Valeri Mladenov - “Implementation of a Feed-forward Artificial Neural Network in VHDL on FPGA”- 978-1-4799-5888-7/14/$31.00 ©2014 IEEE.
[5] Yufei Ma, Naveen Suda, Yu Cao, Jae-sun Seo, Sarma Vrudhula - “Scalable and Modularized RTL Compilation of Convolutional Neural Networks onto FPGA”.
[6] “IEEE Standard for Floating-Point Arithmetic” - September 03,2015 at 19:44:10 UTC from IEEE Xplore.
[7] Peng Li, Ming Liu, Xu Zhang and Hongda Chen - “Effcient Online Feature Extraction algorithm for Spile Sorting in A Multichannel FPGA-Based Neural Recording System” - 978-1-4799-2346-5/14/$31.00 ©2014 IEEE.
[8] SAMI EL MOUKHLIS, ABDESSAMAD ELRHARRAS, ABDELLATIF HAMDOUN - “FPGA Implementation of Artificial Neural Networks” - IJCSI International Journal of Computer Science Issues, Vol. 11, Issue 2, No 1, March 2014.
[9] Suhap Sahin, Yasar Becerikli, and Suleyman Yazici - “Neural Network Implementation in Hardware Using FPGAs” - ICONIP 2006, Part III, LNCS 4234, pp. 1105 – 1112, 2006. © Springer-Verlag Berlin Heidelberg 2006.
[10] E.Al Zuraiqi, M.Joler, C.G.Christodoulou - “Neural Networks FPGA Controller for Reconfigurable Antennas” - 978-1-4244-4968-2/10/$25.00 ©2010 IEEE.
[11] Zbigniew Szadkowski, Krzysztof Pytel, Pierre Auger Collaboration - “Artificial Neural Network as a FPGA Trigger for a Detection of Very Inclined Air Showers”- IEEE TRANSACTIONS ON NUCLEAR SCIENCE - 0018-9499 © 2015 IEEE.
[12] Alin Tisan , Jeannette Chin - “An End User Platform for FPGA-based design and Rapid Prototyping of FeedForward Artificial Neural Networks with on-chip Back Propagation learning”- 10.1109/TII.2016.2555936, IEEE.
[13] Gabriele-Maria LOZITO, Antonino LAUDANI, Francesco RIGANTI-FULGINEI, Alessandro SALVINI – “FPGA Implementations of Feed Forward Neural Network by using Floating Point Hardware Accelerators”- c 2014 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING.
[14] Asim J. Al-Khalili of Concordia University – Distinguished Emeritus Professor, P. Eng - “FLOATING POINT ADDERS AND MULTIPILERS”.
[15] Cheng-Jian Lina, Hung-Ming Tsai - “FPGA implementation of a wavelet neural network with particle swarm optimization learning” - Mathematical and Computer Modelling 47 (2008) 982–996.
[16] Charles Geoffroy, Jean-Baptiste Michaud, Marc-André Tétrault, Julien Clerk-Lamalice, Charles-Antoine Brunet, Roger Lecomte, Réjean Fontaine - “Real Time Artificial Neural Network FPGA Implementation for Triple Coincidences Recovery in PET”- 0018-9499 © 2015 IEEE - IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 62, NO. 3, JUNE 2015.
[17] Nusirwan Anwar bin Abdul Rahman, Kit Chong Wei, John See - “RGB-H-CbCr Skin Colour Model for Human Face Detection”- Faculty of Information Technology, Multimedia University.
[18] Ravikant G. Biradar, Abhishek Chatterjee, Prabhakar Mishra, Koshy George - “FPGA Implementation of a Multilayer Artificial Neural Network using System-on-Chip Design Methodology” - 978-1-4799-7171-8/15/$31.00 ©2015 IEEE.
[19] Jorge C. Romero-Aragon, Edgar N. Sanchez, Alma Y Alanis - “FPGA Neural Identifer for Insulin-Glucose Dynamics”- World Automation Congress ©2014 TSI Press.
[2] Lei Liu, Jianlu Luo, Xiaoyan Deng, Sikun Li – “FPGA-based Acceleration of Deep Neural Networks Using High Level Method” - 2015 10th International Conference on P2P, Parallel, Grid, Cloud and Internet Computing.
[3] Eriko Nurvitadhi, Jaewoong Sim, David Sheffield, Asit Mishra, Srivatsan Krishnan, Debbie Marr - “Accelerating Recurrent Neural Networks in Analytics Servers: Comparison of FPGA, CPU, GPU, and ASIC”.
[4] Philippe Dondon, Julien Carvalho, Rémi Gardere, Paul Lahalle, Georgi Tsenov and Valeri Mladenov - “Implementation of a Feed-forward Artificial Neural Network in VHDL on FPGA”- 978-1-4799-5888-7/14/$31.00 ©2014 IEEE.
[5] Yufei Ma, Naveen Suda, Yu Cao, Jae-sun Seo, Sarma Vrudhula - “Scalable and Modularized RTL Compilation of Convolutional Neural Networks onto FPGA”.
[6] “IEEE Standard for Floating-Point Arithmetic” - September 03,2015 at 19:44:10 UTC from IEEE Xplore.
[7] Peng Li, Ming Liu, Xu Zhang and Hongda Chen - “Effcient Online Feature Extraction algorithm for Spile Sorting in A Multichannel FPGA-Based Neural Recording System” - 978-1-4799-2346-5/14/$31.00 ©2014 IEEE.
[8] SAMI EL MOUKHLIS, ABDESSAMAD ELRHARRAS, ABDELLATIF HAMDOUN - “FPGA Implementation of Artificial Neural Networks” - IJCSI International Journal of Computer Science Issues, Vol. 11, Issue 2, No 1, March 2014.
[9] Suhap Sahin, Yasar Becerikli, and Suleyman Yazici - “Neural Network Implementation in Hardware Using FPGAs” - ICONIP 2006, Part III, LNCS 4234, pp. 1105 – 1112, 2006. © Springer-Verlag Berlin Heidelberg 2006.
[10] E.Al Zuraiqi, M.Joler, C.G.Christodoulou - “Neural Networks FPGA Controller for Reconfigurable Antennas” - 978-1-4244-4968-2/10/$25.00 ©2010 IEEE.
[11] Zbigniew Szadkowski, Krzysztof Pytel, Pierre Auger Collaboration - “Artificial Neural Network as a FPGA Trigger for a Detection of Very Inclined Air Showers”- IEEE TRANSACTIONS ON NUCLEAR SCIENCE - 0018-9499 © 2015 IEEE.
[12] Alin Tisan , Jeannette Chin - “An End User Platform for FPGA-based design and Rapid Prototyping of FeedForward Artificial Neural Networks with on-chip Back Propagation learning”- 10.1109/TII.2016.2555936, IEEE.
[13] Gabriele-Maria LOZITO, Antonino LAUDANI, Francesco RIGANTI-FULGINEI, Alessandro SALVINI – “FPGA Implementations of Feed Forward Neural Network by using Floating Point Hardware Accelerators”- c 2014 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING.
[14] Asim J. Al-Khalili of Concordia University – Distinguished Emeritus Professor, P. Eng - “FLOATING POINT ADDERS AND MULTIPILERS”.
[15] Cheng-Jian Lina, Hung-Ming Tsai - “FPGA implementation of a wavelet neural network with particle swarm optimization learning” - Mathematical and Computer Modelling 47 (2008) 982–996.
[16] Charles Geoffroy, Jean-Baptiste Michaud, Marc-André Tétrault, Julien Clerk-Lamalice, Charles-Antoine Brunet, Roger Lecomte, Réjean Fontaine - “Real Time Artificial Neural Network FPGA Implementation for Triple Coincidences Recovery in PET”- 0018-9499 © 2015 IEEE - IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 62, NO. 3, JUNE 2015.
[17] Nusirwan Anwar bin Abdul Rahman, Kit Chong Wei, John See - “RGB-H-CbCr Skin Colour Model for Human Face Detection”- Faculty of Information Technology, Multimedia University.
[18] Ravikant G. Biradar, Abhishek Chatterjee, Prabhakar Mishra, Koshy George - “FPGA Implementation of a Multilayer Artificial Neural Network using System-on-Chip Design Methodology” - 978-1-4799-7171-8/15/$31.00 ©2015 IEEE.
[19] Jorge C. Romero-Aragon, Edgar N. Sanchez, Alma Y Alanis - “FPGA Neural Identifer for Insulin-Glucose Dynamics”- World Automation Congress ©2014 TSI Press.