Michael Turmon: Papers


Howard, A., Turmon, M., Matthies, L., Benyang Tang, Angelova, A., & Mjolsness, E. (2006). Towards learned traversability for robot navigation: From underfoot to the far field. Journal Of Field Robotics, 23(11-12), 1005–1017.

Bajracharya, M., Benyang Tang, Howard, A., Turmon, M., & Matthies, L. (2008). Learning long-range terrain classification for autonomous navigation. ICRA 2008. IEEE International Conference on Robotics and Automation, 2008, 4018–4024.

Turmon, M., Bajracharya, M., Tang, B., & Matthies, L. (2010), Unsupervised visual terrain classification from proprioception, technical report.

Rankin, A., Bajracharya, M., Huertas, A., Howard, A., Moghaddam, B., Brennan, S., Ansar, A., Tang, B., Turmon, M., and Matthies, L. (2010), Stereo-vision based perception capabilities developed during the Robotics Collaborative Technology Alliances program, Proc. SPIE, Vol. 7692.

Solar Image Classification

These papers document work on solar image classification, which made it into production use on MDI and HMI.

Turmon, M., Pap, J., & Mukhtar, S. (2002). Statistical pattern recognition for labeling solar active regions: Application to SOHO/MDI imagery. Astrophysical Journal, 568(1), 396–407.

Jones, H. P., Chapman, G. A., Harvey, K. L., Pap, J. M., Preminger, D. G., Turmon, M. J., & Walton, S. R. (2008). A comparison of feature classification methods for modeling solar irradiance variation. Solar Physics, 248(2), 323–337.

Turmon, M., Jones, H. P., Malanushenko, O., and Pap, J. Statistical Feature Recognition for Multidimensional Solar Imagery, Solar Physics, April 2010, 277-298.

M. Turmon, J.T. Hoeksema, X. Sun, and M. Bobra. HARPs: Tracked Active Region Patch Data Product from SDO/HMI 2012 AGU Fall Meeting (poster).

M. Turmon, J.T. Hoeksema, and M. Bobra. TARPs: Tracked Active Region Patches from SOHO/MDI, 2013 AGU Fall Meeting (poster).

M. Turmon, J.T. Hoeksema, and M. Bobra. Tracked Active Region Patches for MDI and HMI, AAS Solar Physics Division Meeting, June 2014.

HMI Data Products

J. T. Hoeksema, Y. Liu, K. Hayashi, X. Sun, J. Schou, S. Couvidat, A. Norton, M. Bobra, R. Centeno , K.D. Leka, G. Barnes and M. Turmon, The HMI Vector Magnetic Field Pipeline: Overview and Performance, Solar Physics, Sept. 2014.

M.G. Bobra, X. Sun, J.T. Hoeksema, M. Turmon, Y. Liu, K. Hayashi, G. Barnes, and K.D. Leka. The Helioseismic and Magnetic Imager (HMI) Vector Magnetic Field Pipeline: SHARPs - Space-weather HMI Active Region Patches, Solar Physics, Sept. 2014.

Earlier solar work

The 2002 paper in ApJ above is definitive, but it took us a long time to write it up; this is some of our earlier work using the same techniques.

M. Turmon, J. M. Pap, and S. Mukhtar, Automatically finding solar active regions using SoHO/MDI photograms and magnetograms, in Proc. SoHO 6/GONG 1998 Workshop on the Structure and Dynamics of the Sun, 979-984, 1998.

Turmon, M. (1997), Identification of solar features via Markov random fields, Second Meeting of the International Association for Statistical Computing (IASC-2).

Turmon, M. and Pap, J. M., 1997, Segmenting chromospheric Images with Markov Random Fields, in Statistical Challenges in Modern Astronomy II , eds. G. J. Babu and E. D. Feigelson, Springer, p. 408.

Solar Image Processing: Experimental

These papers are mainly about using reversible jump MCMC to retrieve large-scale solar structures. The work did not make it into production use.

M. Turmon and S. Mukhtar, Representing solar active regions with triangulations, in Proc. Compstat-98, pp. 473-478, 1998.

Turmon, M. and Mukhtar, S. (1997). Recognizing chromospheric objects via Markov chain Monte Carlo. IEEE ICIP (International Conference on Image Processing), 1997, vol. 3, 320–323.

M. Turmon, S. Mukhtar and J. Pap. Bayesian inference for identifying solar active regions. In H. Mannila, D. Heckerman and D. Pregibon, eds., Proc. Third Conf. on Knowledge Discovery and Data Mining. MIT Press, 1997.

M. Turmon, L. Ramsey, E. Mjolsness and V. Gluzman. A language for probabilistic modeling of scientific data, in Proc. Second Conf. Highly Structured Stochastic Systems, 1999.

Astronomical Transients

Mahabal, A. A., Djorgovski, S. G., Turmon, M., Jewell, J., Williams, R. R., Drake, A. J., Graham, M. G., et al. (2008). Automated Probabilistic Classification of Transients and Variables. Astron. Nachr. AN 329, No. 3, pp. 288–291.

Djorgovski, S. G., Donalek, C., Mahabal, A., Moghaddam, B., Turmon, M., Graham, M., Drake, A., et al. (2011). Towards an Automated Classification of Transient Events in Synoptic Sky Surveys, in Proc. Conf. Intell. Data Understanding (CIDU) 2011, ed. A. Srivasatva et al., p. 174.

Nonlinear Dynamics

Chin, T. M., Turmon, M. J., Jewell, J. B., & Ghil, M. (2007). An ensemble-based smoother with retrospectively updated weights for highly nonlinear systems. Monthly Weather Review, 135(1), 186–202.

Fault Tolerant Computing

Turmon, M., Granat, R., Katz, D., & Lou, J. (2003). Tests and tolerances for high-performance software-implemented fault detection. IEEE Transactions on Computers, 52(5), 579–591.

Turmon, M., Granat, R., & Katz, D. (2000). Software-Implemented Fault Detection for High-Performance Space Applications. IEEE International Conf. Dependable Systems and Networks, June 2000.

Turmon, M. & Granat, R. (2000). Algorithm-based fault tolerance for spaceborne computing: basis and implementations. IEEE Aerospace Conference, pp. 411-420, vol. 4.

R. Granat, K. L. Wagstaff, B. Bornstein, B. Tang, and M. Turmon (2009). Simulating and Detecting Radiation-Induced Errors for Onboard Machine Learning, Proc. Third IEEE International Conference on Space Mission Challenges for Information Technology (SMC-IT), 125-131.


Work on learning rates in neural networks, aimed at explaining why VC theory is a bad predictor of accuracy-vs-complexity relationships.

M. Turmon, Assessing generalization of feedforward neural networks, Cornell University, 1995. Trying to give dissertations a good name.

M. Turmon and T.L. Fine. Sample size requirements for feedforward neural networks, in Advances in Neural Information Processing Systems (NIPS) 7, MIT Press, 1995.

T. L. Fine and M. Turmon (1993). Sample Size Requirements of Feedforward Neural Network Pattern Classifiers. Proc. IEEE International Symposium on Information Theory, 1993, 432.

M. Turmon and T. L. Fine. Empirically estimating generalization ability of feedforward neural networks. In Proc. World Conf. Neural Networks, INNS Press, 1995, pp. 600-605.

M. Turmon and T. L. Fine (1995). Assessing generalization of feedforward neural networks, in Proc. IEEE Intl. Symp. on Information Theory, 1995, 168. The full-length paper is here.

Expectation-Maximization Algorithm

M. J. Turmon and M. I. Miller. Maximum-likelihood estimation of constrained means and Toeplitz covariances with application to direction finding, IEEE Trans. Sig. Proc., 42(5):1074-1086, May 1994. There is also a short tech report on the Cramer-Rao bound for the constrained covariance estimation problem.

Turmon, M. Symmetric Gaussian mixtures, in Compstat 2004, Proceedings in Computational Statistics, pp. 1909-16, Physica-Verlag, 2004. The tech report contains more on the calculations.