1. E.D. Sontag. Feedback stabilization using two-hidden-layer nets. IEEE Trans. Neural Networks, 3:981-990, 1992. [PDF] Keyword(s): machine learning, neural networks, feedback stabilization. Abstract:

   This paper compares the representational capabilities of one hidden layer and two hidden layer nets consisting of feedforward interconnections of linear threshold units. It is remarked that for certain problems two hidden layers are required, contrary to what might be in principle expected from the known approximation theorems. The differences are not based on numerical accuracy or number of units needed, nor on capabilities for feature extraction, but rather on a much more basic classification into "direct" and "inverse" problems. The former correspond to the approximation of continuous functions, while the latter are concerned with approximating one-sided inverses of continuous functions - and are often encountered in the context of inverse kinematics determination or in control questions. A general result is given showing that nonlinear control systems can be stabilized using two hidden layers, but not in general using just one.




2. E.D. Sontag. Feedforward nets for interpolation and classification. J. Comput. System Sci., 45(1):20-48, 1992. [PDF] [doi:http://dx.doi.org/10.1016/0022-0000(92)90039-L] Keyword(s): machine learning, neural networks, VC dimension, boolean systems. Abstract:

   This paper deals with single-hidden-layer feedforward nets, studying various aspects of classification power and interpolation capability. In particular, a worst-case analysis shows that direct input to output connections in threshold nets double the recognition but not the interpolation power, while using sigmoids rather than thresholds allows doubling both. For other measures of classification, including the Vapnik-Chervonenkis dimension, the effect of direct connections or sigmoidal activations is studied in the special case of two-dimensional inputs.




3. E.D. Sontag. Universal nonsingular controls. Systems Control Lett., 19(3):221-224, 1992. Note: Erratum appeared in SCL 20(1993), p. 77, can be found in same file.[PDF] [doi:http://dx.doi.org/10.1016/0167-6911(92)90116-A] Keyword(s): controllability, real-analytic functions. Abstract:

   For analytic systems satisfying the strong accessibility rank condition, generic inputs produce trajectories along which the linearized system is controllable. Applications to the steering of systems without drift are briefly mentioned.




4. Y. Wang and E.D. Sontag. Algebraic differential equations and rational control systems. SIAM J. Control Optim., 30(5):1126-1149, 1992. [PDF] Keyword(s): identifiability, observability, realization theory, input/output system representations. Abstract:

   It is shown that realizability of an input/output operators by a finite-dimensional continuous-time rational control system is equivalent to the existence of a high-order algebraic differential equation satisfied by the corresponding input/output pairs ("behavior"). This generalizes, to nonlinear systems, the classical equivalence between autoregressive representations and finite dimensional linear realizability.




5. Y. Wang and E.D. Sontag. Generating series and nonlinear systems: analytic aspects, local realizability, and i/o representations. Forum Math., 4(3):299-322, 1992. [PDF] Keyword(s): identifiability, observability, realization theory, input/output system representations, real-analytic functions. Abstract:

   This paper studies fundamental analytic properties of generating series for nonlinear control systems, and of the operators they define. It then applies the results obtained to the extension of facts, which relate realizability and algebraic input/output equations, to local realizability and analytic equations.

1. F. Albertini and E.D. Sontag. For neural networks, function determines form. In Proc. IEEE Conf. Decision and Control, Tucson, Dec. 1992, IEEE Publications, 1992, pages 26-31, 1992. Keyword(s): machine learning, neural networks, recurrent neural networks.



2. M.A. Dahleh, E.D. Sontag, D.N.C. Tse, and J.N. Tsitsiklis. Worst-case identification of nonlinear fading memory systems. In Proc. Amer. Automatic Control Conf., Chicago, June 1992, pages 241-245, 1992. [PDF] Keyword(s): information-based complexity, fading-memory systems, stability, system identification, structured uncertainty. Abstract:

   Preliminary version of paper published in Automatica in 1995.




3. Y. Lin and E.D. Sontag. Gradient techniques for steering systems with no drift. In Proc. Conf. Inform. Sci. and Systems, Princeton University, March 1992, pages 1003-1008, 1992.



4. R. Schwarzschild, E.D. Sontag, and M.L.J. Hautus. Output-Saturated Systems. In Proc. Amer. Automatic Control Conf. , Chicago, June 1992, pages 2504-2509, 1992.



5. H.T. Siegelmann and E.D. Sontag. On the computational power of neural nets. In COLT '92: Proceedings of the fifth annual workshop on Computational learning theory, New York, NY, USA, pages 440-449, 1992. ACM Press. [doi:http://doi.acm.org/10.1145/130385.130432] Keyword(s): analog computing, neural networks, computational complexity, super-Turing computation, recurrent neural networks.



6. H.T. Siegelmann and E.D. Sontag. Some results on computing with neural nets. In Proc. IEEE Conf. Decision and Control, Tucson, Dec. 1992, IEEE Publications, 1992, pages 1476-1481, 1992. Keyword(s): analog computing, neural networks, computational complexity, super-Turing computation, recurrent neural networks.



7. H.T. Siegelmann, E.D. Sontag, and C.L. Giles. The Complexity of Language Recognition by Neural Networks. In Proceedings of the IFIP 12th World Computer Congress on Algorithms, Software, Architecture - Information Processing '92, Volume 1, pages 329-335, 1992. North-Holland. Keyword(s): machine learning, neural networks, computational complexity, machine learning, recurrent neural networks, theory of computing and complexity.



8. E.D. Sontag. Neural nets as systems models and controllers. In Proc. Seventh Yale Workshop on Adaptive and Learning Systems, Yale University, 1992, pages 73-79, 1992. [PDF] Keyword(s): machine learning, neural networks, recurrent neural networks, neural networks. Abstract:

   A conference paper. Placed here because it was requested, but contains little that is not also contained in the survey on neural nets mentioned above.




9. E.D. Sontag. Systems combining linearity and saturations, and relations to neural nets. In Nonlinear Control Systems Design 1992, IFAC Symposia Series, 1993, M. Fliess Ed., Pergamon Press, Oxford, 1993, pages 15-21, 1992. Note: (Also in Proc. Nonlinear Control Systems Design Symp., Bordeaux, June 1992, M. Fliess, Ed., IFAC Publications, pp. 242-247). Keyword(s): machine learning, neural networks, recurrent neural networks.



10. E.D. Sontag and Y. Lin. Stabilization with respect to noncompact sets: Lyapunov characterizations and effect of bounded inputs. In Nonlinear Control Systems Design 1992, IFAC Symposia Series, M. Fliess Ed., Pergamon Press, Oxford, 1993, pages 43-49, 1992. Note: Also in Proc. Nonlinear Control Systems Design Symp., Bordeaux, June 1992,(M. Fliess, Ed.), IFAC Publications, pp. 9--14. [PDF] Keyword(s): saturation, bounded inputs.



11. E.D. Sontag and Y. Wang. I/O equations in discrete and continuous time. In Proc. IEEE Conf. Decision and Control, Tucson, Dec. 1992, IEEE Publications, 1992, pages 3661-3662, 1992. Keyword(s): identifiability, observability, realization theory.



12. Y. Yang, H.J. Sussmann, and E.D. Sontag. Stabilization of linear systems with bounded controls. In Nonlinear Control Systems Design 1992, IFAC Symposia Series, 1993, M. Fliess Ed., Pergamon Press, Oxford, 1993, pages 51-56, 1992. Note: Also in Proc. Nonlinear Control Systems Design Symp., Bordeaux, June 1992,(M. Fliess, Ed.), IFAC Publications, pp. 15-20.Keyword(s): saturation, bounded inputs.

This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders.

This document was translated from BibT_EX by bibtex2html