Search results

  1. 1.
    0405189 - UIVT-O 20030200 CZ eng V - Research Report
    Kůrková, Věra - Sanguineti, M.
    Learning with Generalization Capability by Kernel Methods of Bounded Complexity.
    Prague: ICS AS CR, 2003. 16 s. Technical Report, V-901.
    R&D Projects: GA ČR GA201/02/0428
    Grant - others:IT-CZ Area MC6(XX) Project 22
    Institutional research plan: AV0Z1030915
    Keywords : supervised learning * generalization * model complexity * kernel methods * minimization of regularized empirical errors * upper bounds on rates of approximate optimization
    Subject RIV: BA - General Mathematics
    Permanent Link: http://hdl.handle.net/11104/0125385
    FileDownloadSizeCommentaryVersionAccess
    v901-03.pdf15276.9 KBOtheropen-access
     
     
  2. 2.
    0405002 - UIVT-O 20030185 CZ eng V - Research Report
    Kůrková, Věra - Sanguineti, M.
    Error Estimates for Approximate Optimization by the Extended Ritz Method.
    Prague: ICS AS CR, 2002. 17 s. Technical Report, V-882.
    R&D Projects: GA ČR GA201/02/0428
    Grant - others:IT-CZ Area MC6(XX) Project 22
    Institutional research plan: AV0Z1030915
    Keywords : approximate optimization * extended Ritz method * rates of convergence of approximate infima and argminima * high-dimensional optimization problems * curse of dimensionality * convex best approximation * learning from data
    Subject RIV: BA - General Mathematics
    Permanent Link: http://hdl.handle.net/11104/0125222
    FileDownloadSizeCommentaryVersionAccess
    v882-02.pdf18220.6 KBOtheropen-access
     
     
  3. 3.
    0404713 - UIVT-O 20020064 CZ eng V - Research Report
    Kůrková, Věra - Sanguineti, M.
    Tight Bounds on Rates of Variable-Basis Approximation via Estimates of Covering Numbers.
    Prague: ICS AS CR, 2001. 12 s. Technical Report, V-865.
    R&D Projects: GA ČR GA201/99/0092
    Grant - others:NATO(XE) PST.CLG.976870
    Institutional research plan: AV0Z1030915
    Keywords : high-dimensional optimization * curse of dimensionality * rates of approximation * covering numbers * distributed computing * neural networks
    Subject RIV: BA - General Mathematics
    Permanent Link: http://hdl.handle.net/11104/0124952
    FileDownloadSizeCommentaryVersionAccess
    v865-02.pdf19327 KBOtheropen-access
     
     
  4. 4.
    0404712 - UIVT-O 20020063 CZ eng V - Research Report
    Kainen, P.C. - Kůrková, Věra - Sanguineti, M.
    Minimization of Error Functionals over Variable-Basis Functions.
    Prague: ICS AS CR, 2002. 17 s. Technical Report, V-864.
    R&D Projects: GA ČR GA201/00/1482
    Institutional research plan: AV0Z1030915
    Keywords : error functionals * approximate optimalization * generalized Tychonov well-posedness * rate of decrease of intima * complexity of admissible functions * curse of dimensionality * neural networks
    Subject RIV: BA - General Mathematics
    Permanent Link: http://hdl.handle.net/11104/0124951
    FileDownloadSizeCommentaryVersionAccess
    v864-02.pdf18350.1 KBOtheropen-access
     
     
  5. 5.
    0404634 - UIVT-O 20020043 CZ eng V - Research Report
    Kůrková, Věra - Sanguineti, M.
    Covering Numbers and Rates of Neural-Network Approximation.
    Prague: ICS AS CR, 2001. 7 s. Technical Report, V-830.
    R&D Projects: GA ČR GA201/00/1489
    Grant - others:NATO(XE) PST.CLG.976870
    Institutional research plan: AV0Z1030915
    Keywords : nonlinear approximation * rates of approximation * variable-basis approximation * feedforward neural networks * covering numbers
    Subject RIV: BA - General Mathematics
    Permanent Link: http://hdl.handle.net/11104/0124875
    FileDownloadSizeCommentaryVersionAccess
    v830-01.pdf18312.4 KBOtheropen-access
     
     
  6. 6.
    0403747 - UIVT-O 990101 CZ eng V - Research Report
    Kainen, P.C. - Kůrková, Věra - Vogt, A.
    Best Approximation by Linear Combinations of Characteristic Functions of Half-Spaces.
    Prague: ICS AS CR, 1999. 6 s. Technical Report, V-795.
    R&D Projects: GA ČR GA201/99/0092
    Keywords : best approximation * near best approximation * continuous selection * Chebyshev set * strictly convex space * uniformly convex space * conctraible set * tangent hypercone * modulus of convexity * Chebyshev radius
    Permanent Link: http://hdl.handle.net/11104/0124040
    FileDownloadSizeCommentaryVersionAccess
    v795-99.pdf13196.7 KBOtheropen-access
     
     
  7. 7.
    0403728 - UIVT-O 990089 CZ eng V - Research Report
    Kůrková, Věra - Sanguineti, M.
    Dimension-Independent Approximation by Neural Networks and Its Comparison with Linear Approximation.
    Prague: ICS AS CR, 1999. 16 s. Technical Report, V-789.
    R&D Projects: GA ČR GA201/99/0092
    Keywords : linear and nonlinear approximation * Kolmogorov width * dimension-independent approximation * curse of dimensionality * variation with respect to a set of functions * complexity of neural networks * one-hidden-layer perceptron networks
    Permanent Link: http://hdl.handle.net/11104/0124021
    FileDownloadSizeCommentaryVersionAccess
    v789-99.pdf15284.6 KBOtheropen-access
     
     
  8. 8.
    0403664 - UIVT-O 990015 CZ eng V - Research Report
    Kainen, P.C. - Kůrková, Věra - Vogt, A.
    Geometry and Topology of Continuous Best and Near Best Approximations.
    Prague: ICS AS CR, 1998. 9 s. Technical Report, V-769.
    R&D Projects: GA ČR GA201/96/0917
    Keywords : best approximation * near best approximation * continuous selection * Chebyshev set * strictly convex space * uniformly convex space * conctraible set * tangent hypercone * modulus of convexity * Chebyshev radius
    Permanent Link: http://hdl.handle.net/11104/0123962
    FileDownloadSizeCommentaryVersionAccess
    v769-98.pdf13157.2 KBOtheropen-access
     
     
  9. 9.
    0403635 - UIVT-O 980152 CZ cze V - Research Report
    Kůrková, Věra
    Aproximace funkcí neuronovými sítěmi.
    Prague: ICS AS CR, 1998. 18 s. Technical Report, V-770.
    R&D Projects: GA AV ČR IAA2030602
    Permanent Link: http://hdl.handle.net/11104/0123934
    FileDownloadSizeCommentaryVersionAccess
    v770-98.pdf12259.7 KBOtheropen-access
     
     
  10. 10.
    0402896 - UIVT-O 960206 CZ eng V - Research Report
    Kůrková, Věra
    Trade-off Between the Size of Weights and the Number of Hidden Units in Feedforward Networks.
    Prague: ICS AS CR, 1996. 17 s. Technical Report, V-695.
    R&D Projects: GA AV ČR IAA2030602; GA ČR GA201/96/0917
    Keywords : approximation of functions * one-hidden-layer neural networks * sigmodial perceptrons * radial-basis-functions
    Permanent Link: http://hdl.handle.net/11104/0123270
    FileDownloadSizeCommentaryVersionAccess
    v695-96.pdf13249.8 KBOtheropen-access
     
     

  This site uses cookies to make them easier to browse. Learn more about how we use cookies.