Michael Herzog

Curriculum Vitae

I earned my ``Diplom'' in mathematics at the University of Tübingen (Germany) with a thesis on the structure of the ``Automorphism groups of Hamming graphs'' in 1992 under the supervision of Prof. Dr. Hering. In 1993 I received a Master in philosophy. I was interested in the philosophy of mind and wrote a thesis about current approaches to intentionality and representation (supervision Prof. Dr. Keuth). My main interest, however, is the wide field of neurobiology and behavior. Under the supervision of Prof. M. Fahle in the Section of Visual Science at the Eye Hospital of Tübingen (Germany) I finished a Ph.D. thesis on ``mathematical models and experiments in perceptual learning'' earning a Ph.D. degree in biology.

I'm right now a post-doc in Prof. Fahles laboratory investigating the influence of higher cortical information on perceptual learning.

General Research Interests:

My research interest is the investigation of perceptual learning. What are good models for perceptual learning and how can these models be empirically justified? The main result of my Ph.D. thesis is that models belonging to the class of neural networks- both supervised and unsupervised- are inadequate to describe biological learning phenomena. These models fail to explain the outcomes of my experiments and cannot account for some problems which have to be met in perceptual learning. A list of these difficulties is published in Herzog & Fahle (1997b).
In the current experimental project- supported by Anne Holland-Moritz and Astrid Broos - we investigate how higher cortical influences like attention and task related information control the learning process which is believed to take place on the very early stages of visual processing. The results of these experiments are used to develop new models which can solve the problems mentioned above. In computer simulations these models will be tested.

Address: Section of Visual Science, University Eye Hospital Waldhörnlestr. 22, D72072 Tübingen, Germany
Phone: (49) - 7071 - 29 81207
Fax: (49) - 7071 - 29 5568
Email: michael.herzog@uni-tuebingen.de(michael.herzog@uni-tuebingen.de)

Publications:

  • Herzog, M.H. (1992). Automorphism of Hamming Graphs (in German), Final Thesis (Diplomarbeit) in Mathematics, University of Tübingen.

  • Herzog, M.H. (1993). New Approches to Intentionality and Representation (in German), Master Thesis in Philosophy 1993, University of Tübingen.

  • Herzog, M.H. (1996). Perceptual learning-mathematical models and experiments (in German), PH.D. Thesis in Biology, University of Tübingen.

  • Herzog, M.H., Fahle, M. (1997a). The role of feedback in learning a vernier discrimination task. Vision Research, in the press.

  • Herzog, M.H., Fahle, M. (1997b). Modeling perceptual learning: difficulties and how they can be overcome. Biological Cybernetics, submitted.

  • Herzog, M.H., Fahle, M. (1997c). Effects of biased feedback on learning and deciding in a vernier discrimination task. Vision Research, submitted.

Abstracts:

  • Herzog, M.H., Fahle, M. (1994). Learning without attention? Proceedings of the 22nd Goettingen Neurobiology Conference 1994, Volume II No. 817.

  • Herzog M.H., Fahle M. (1995a). The role of feedback in learning a hyperacuity task. Investigative Ophthalmology & Visual Science No. 35/1775.

  • Herzog, M.H. Fahle, M. (1995b). A recurrent model for learning a hyperacuity task. Perception, 24, 84a.

  • Herzog, M.H., Fahle, M. (1996a). Perceptual Learning is Not Caused By Fine Tuning of Receptive Fields Alone. Investigative Ophthalmology & Visual Science No. 37/3181.

  • Herzog, M.H., Fahle, M. (1996b). Biased Decision Criteria in Vernier Discrimination. Perception, 25,139b.

  • Herzog, M.H., Fahle, M. (1996c). Networks and learning. Invited talk at the 2nd Meeting of European Neuroscience (ENA), 119.04.

  • Herzog, M.H., Fahle, M. (1997a). Effects of biased feedback on learning and deciding in a vernier discrimination task. Investigative Ophthalmology & Visual Science No. 37/3181.

  • Herzog, M.H., Fahle, M. (1997b). Neural Networks cannot explain perceptual learning. Proceedings of the Conference in Neurobiology, Goettingen 1997.

back to