Adjunct Assistant Professor
School of Operations Research and Industrial Engineering
Cornell Theory Center (Center for Theory and Simulation
in Science and Engineering)
School of Operations Research and Cornell Theory Center
237 Rhodes Hall
Ithaca, NY 14853
Telephone: (607) 254-4606
Fax: (607) 255-9129
Algorithms and complexity theory in particular approximate solution
of NP-hard optimization problems.
Discrete optimization current focus on combinatorial problems
related to inconsistent linear systems with applications in several fields
including image and signal processing.
Machine learning and artificial neural networks from an
Previous focus was on the hardness
of learning problem, in particular of designing near-optimal linear
classifiers. Currently we are exploring support vector machines.
In Spring 96 I taught Mathematical Programming II
(OR&IE 631) with
I'm currently teaching Mathematical Programming I
Two constructive methods for designing compact feedforward networks of
threshold units, with Bertrand Guenin,
To appear in International Journal of Neural Systems.
An efficient line detection algorithm based on a new combinatorial
optimization formulation, with M. Mattavelli and V. Noel,
Proceedings of the 1998 International Conference on Image Processing
(ICIP98), Chicago IL, October 1998.
On the approximability of minimizing nonzero variables or unsatisfied relations
in linear systems,
with Viggo Kann, Theoretical Computer Science Vol. 209 (1998) 237-260.
Preliminary version available as
ECCC Technical Report 96-15.
A perceptron-based approach to piecewise linear modeling with an application to
time series, with Marco Mattavelli and Jean-Marc Vesin,
Proceedings of International Conference on Artificial Neural Networks (ICANN'97), Lecture Notes
in Computer Science, Vol. 1327, Springer-Verlag (1997) 547-552.
A new approach to piecewise linear modeling of time series, with M. Mattavelli,
J.-M. Vesin and R. Gruter, Proceedings of the Seventh IEEE Digital Signal
Processing Workshop, Loen, Norway, IEEE Press, 1996.
Estimating piecewise linear models using combinatorial optimization
techniques, with M. Mattavelli, Proceedings of VIII European Signal
Processing Conference (EUSIPCO'96), Trieste, Italy, 1996.
The complexity and approximability of finding maximum feasible subsystems
of linear relations, with
Theoretical Computer Science,
Vol. 147 (1995) 181-210.
Using perceptron-like algorithms for the analysis and parameterization of object
motion, with Marco
Mattavelli, In F. Girosi et al. (editors), Neural Networks
for Signal Processing V, Proceedings of the 1995 IEEE workshop, IEEE Press
From finding maximum feasible subsystems of linear systems to feedforward
neural network design, Ph.D. dissertation No. 1282,
Department of Mathematics,
Swiss Federal Institute of Technology at Lausanne
( EPFL), October 1994.
- A review of combinatorial problems arising in feedforward neural networks,
with Eddy Mayoraz and D. de Werra, Discrete Applied Mathematics, Vol. 52
- On the approximability of finding maximum feasible subsystems
of linear systems, with Viggo Kann, Proceedings of STACS'94,
Lecture Notes in Computer Science, Vol. 775 (1994), 521-532.
- On the complexity of training perceptrons, in T. Kohonen et al. (editors),
Artificial Neural Networks, Vol. 1, Elsevier Science Publisher, North-Holland,
Amsterdam (1991), 55-60.
- Computing optical flow across multiple scales: a coarse-to-fine approach,
with Roberto Battiti and Christof Koch, International Journal of Computer
Vision, Vol. 6 No. 2 (1991), 133-145.
- Stability-capacity diagram of a neural network with Ising bonds,
with Stam Nicolis, Journal of Physics (France), (September 1989), 2333-2345.
Soon available on-line:
A combinatorial optimization approach to extract piecewise linear
structure in nonlinear data and an application to optical flow segmentation,
with Marco Mattavelli.
- On the probabilistic and thermal perceptron algorithms, with
- The MIN PCS problem and piecewise linear model estimation, with
Marco Mattavelli, extended abstract.
- Maximizing consistency versus minimizing disagreement and the relevance
of non-learnability results.
306 East State St Apt. 502
Ithaca, NY 14850
Telephone + Fax : (607) 273-7798