Ultrafast Neural Network Training for Robot Learning from Uncertain Data

  • J. Barhen
  • V. Protopopescu


A capability for learning from uncertain data has been a major and perennial requirement for many real-life robotic applications. In that context, a new methodology for ultrafast learning using neural networks is presented. It requires only a single iteration to train a feed-forward network with near-optimal results. Uncertainty reduction algorithms are also incorporated in a natural and optimal fashion. As such, this methodology is intended to become an essential building block for future architectures of intelligent systems. Its application to multi-robot observation of multiple moving targets is illustrated.


Neural Network Probability Density Function Uncertain Data Autonomous Robot Global Optimization Algorithm 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    S.-I. Amari and M. Arbib, Competition and Cooperation in Neural Networks ,Springer (1982).CrossRefGoogle Scholar
  2. 2.
    S. Grossberg, The Adaptive Brain, I: Cognition, Learning, Reinforcement, and Rhythm ,Elsevier (1987).Google Scholar
  3. 3.
    S. Grossberg, The Adaptive Brain, II: Vision, Speech, Language, and Motor Control ,Elsevier (1987).Google Scholar
  4. 4.
    R. Arkin, Behavior-Based Robotics ,MIT Press (1998).Google Scholar
  5. 5.
    D. Rumelhart and J. McLelland, Parallel Distributed Processing ,MIT Press (1986).Google Scholar
  6. 6.
    D. Waterman and F. Hayes-Roth, Pattern-Directed Inference Systems ,Academic Press (1978).MATHGoogle Scholar
  7. 7.
    J. Barhen, W. Dress, and C. Jorgensen, “Application of Concurrent Neuromorphic Algorithms to Autonomous Robots”, in Neural Computers ,NATO ASI ,F41, pp. 321–334, Springer (1988)Google Scholar
  8. 7.a
    J. Barhen, M. Zak, and N. Toomarian, “Non-Lipschitzian Neural Dynamics”, in Advanced Neural Computers ,R. Eckmiller, ed., pp. 103–112, North Holland (1990)Google Scholar
  9. 7.b
    N. Toomarian, J. Barhen, and S. Gulati, “Neural Networks for Real-Time Robotic Applications”, in Parallel Computation Systems for Robotics ,A. Fijany and A. Bejczy eds., pp. 213–247, World Scientific (1992).CrossRefGoogle Scholar
  10. 8.
    G. Bekey and K. Goldberg, Neural Networks in Robotics ,Kluwer (1993).MATHCrossRefGoogle Scholar
  11. 9.
    O. Omidvar and P. van der Smagt, Neural Systems for Robotics ,Academic Press (1997).Google Scholar
  12. 10.
    R. Brooks, S. Grossberg, and L. Optican, eds, “Neural Control and Robotics”, special issue of Neural Networks ,11, no. 7-8 (1998).Google Scholar
  13. 11.
    T. Ziemke and N. Sharkey, eds, “Artificial Neural Networks for Robot Learning”, special issue of Autonomous Robots ,7, no. 7 (1999).Google Scholar
  14. 12.
    L. Parker and C. Touzet, “Multi-Robot Learning in a Cooperative Observation Task”, in Distributed Autonomous Robotic Systems 4 ,L. Parker, G. Bekey, and J. Barhen eds.. Springer (in press, 2000).Google Scholar
  15. 13.
    C. Touzet, “Distributed Lazy Q-learning for Cooperative Mobile Robots”, Autonomous Robots (in press, 2000).Google Scholar
  16. 14.
    A. Agah and G. Bekey, “Cognitive Architecture for Robust Adaptive Control of Robots in a Team”, Journal of Intelligent and Robotic Systems ,20, 251–273 (1997).CrossRefGoogle Scholar
  17. 15.
    M. Hassoun, Fundamentals of Artificial Neural Networks ,MIT Press (1995).Google Scholar
  18. 16.
    M. Bishop, Neural Networks for Pattern Recognition ,Oxford University Press (1997).Google Scholar
  19. 17.
    M. Beckerman, Adaptive Cooperative Systems ,Wiley Interscience (1997).Google Scholar
  20. 18.
    D. White and D. Sofge, Handbook of Intelligent Control: Neural Fuzzy and Adaptive Approaches ,Van Nostrand Reinhold (1992).Google Scholar
  21. 19.
    A. Cichocki and R. Unbenhauen, Neural Networks for Optimization and Signal Processing ,Wiley (1993).MATHGoogle Scholar
  22. 20.
    P. Mars, J. Chen, and R. Nambiar, Learning Algorithms ,CRC Press (1996).Google Scholar
  23. 21.
    N. Toomarian and J. Barhen, “Learning a Trajectory using Adjoint Functions and Teacher Forcing”, Neural Networks ,5,473–484 (1992).CrossRefGoogle Scholar
  24. 22.
    N. Toomarian and J. Barhen, “Fast Temporal Neural Learning Using Teacher Forcing”, U.S. Patent No. 5,428,710 (March 28,1995).Google Scholar
  25. 23.
    Y. Chauvin and D. Rumelhart, Backpropagation: Theory, Architectures, and Applications ,Lawrence Erlbaum (1995).Google Scholar
  26. 24.
    J. Barhen, S. Gulati, and M. Zak, “Neural Learning of Constrained Nonlinear Transformations”, IEEE Computer ,22(6), 67–76 (1989).CrossRefGoogle Scholar
  27. 25.
    J. Barhen, N. Toomarian, and S. Gulati, “Applications of Adjoint Operators to Neural Networks”, Appl. Math. Lett. ,3(3), 13–18 (1990).MathSciNetMATHCrossRefGoogle Scholar
  28. 26.
    N. Toomarian and J. Barhen, “Neural Network Training by Integration of Adjoint Systems of Equations Forward in Time”, U.S. Patent No. 5,930,781 (July 27,1999).Google Scholar
  29. 27.
    J. Barhen, V. Protopopescu, and D. Reister, “TRUST: A Deterministic Algorithm for Global Optimization”, Science ,276,1094–1097 (1997).MathSciNetMATHCrossRefGoogle Scholar
  30. 28.
    A. Shepherd, Second-Order Methods for Neural Networks ,Springer (1997).CrossRefGoogle Scholar
  31. 29.
    F. Biegler-Konig and F. Barman, “A Learning Algorithm for Multilayered Neural Networks based on Linear Least Squares Problems”, Neural Networks ,6,127–131 (1993).CrossRefGoogle Scholar
  32. 30.
    J. Barhen, R. Cogswell, and V. Protopopescu, “Single Iteration Training Algorithm for Multi-layer Feed-forward Neural Networks,” Neural Proc. Lett ,11(2) (in press, 2000).Google Scholar
  33. 31.
    J. Barhen, D.Reister, and V. Protopopescu, “DeepNet An Ultrafast Neural Learning Code for Seismic Imaging”, Procs. IJCNN’99 ,CD-ROM, IEEE Press (1999).Google Scholar
  34. 32.
    G. Golub and C. Van Loan, Matrix Computations ,Johns Hopkins University Press (1983).Google Scholar
  35. 33.
    C. Weisbin et al ,“Sensitivity and Uncertainty of Reactor Performance Parameters”, in Advances in Nuclear Science and Technology ,14, Plenum Press, New York (1981).Google Scholar
  36. 34.
    D. Cacuci et al ,“Sensitivity Theory for General Systems of Non-linear Equations”, Nucl. Sci. Eng. ,75, 88–110 (1980).Google Scholar
  37. 35.
    R. Alsmiller, J. Barhen, and J. Horwedel, “The Application of Adjoint Sensitivity Theory to a Liquid Fuels Supply Model”, Energy ,9, 239–253 (1984).CrossRefGoogle Scholar
  38. 36.
    D. Reister and J. Barhen, “Application of Sensitivity and Uncertainty Analysis to the Estimation of Methane Hydrate Concentration”, Jour, of Petroleum Science and Engineering (in preparation, 2000).Google Scholar
  39. 37.
    D. Cox and P. Baybutt, “Methods for Uncertainty Analysis: A Comparative Survey”, Risk Analysis ,1(4), 251–258 (1981).CrossRefGoogle Scholar
  40. 38.
    R. Myers, Response Surface Methodology ,Edwards Bros. (1976).Google Scholar
  41. 39.
    B. Kosko, Fuzzy Engineering ,Prentice Hall (1996).Google Scholar
  42. 40.
    A. Dubrawski, “Stochastic Validation for Automated Tuning of Neural Network’s Hyper Parameters”, Robotics and Autonomous Systems ,21, 83–93 (1997).CrossRefGoogle Scholar
  43. 41.
    F. Aminzadeh, J. Barhen, C. Glover, and N. Toomarian, “Estimation of Reservoir Parameters using a Hybrid Neural Network”, Jour, of Petroleum Science & Engineering ,24,49–56 (1999).CrossRefGoogle Scholar
  44. 42.
    J. Barhen, D. Cacuci, J. Wagschal, and M. Bjerke, “Uncertainty Analysis of Time-Dependent Nonlinear Systems”, Nuclear Science &Engineering ,81(1), 23–44 (1982).Google Scholar
  45. 43.
    J. Tolma and P. Barton, “On Computational Differentiation”, Comp. Chem. Engin. ,22(4/5), 475–490 (1998).CrossRefGoogle Scholar
  46. 44.
    J. Barhen, N. Toomarian and V. Protopopescu, “Optimization of the Computational Load of a Hypercube Supercomputer Onboard a Mobile Robot”, Applied Optics ,26 (23), 5007–5014 (1987).CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Tokyo 2000

Authors and Affiliations

  • J. Barhen
    • 1
  • V. Protopopescu
    • 1
  1. 1.Center for Engineering Science Advanced Research, Computer Science and Mathematics DivisionOak Ridge National LaboratoryOak RidgeUSA

Personalised recommendations