MetaCartSign in to MyCiteSeer

Include Citations | Advanced Search | Help

Include Citations | Advanced Search | Help

  Combining Uni-Modal Classifiers to Improve Learning (1994) [2 citations — 0 self]

Download:
Download as a PDF | Download as a PS
by Virginia De Sa
In Integration of Elementary Functions into Complex Behavior
ftp://keck.ucsf.edu/pub/desa/bielefeldbook.ps
Add To MetaCart

Abstract:

One of the key ideas in both robotics and neuroscience is that complex behaviour can arise from the interaction of many cooperating simple agents or modules. In this paper we suggest that this idea can be extended; just as combining simple agents may be important for complex behaviour, combining tasks is important for learning the parts themselves. In particular we show that combining classifications across different modalities can help solve the teaching signal dilemma and allow the development of task relevant classifications without external supervision. We recap some psychophysical and neurobiological data supporting the idea that information from different modalities can assist (or interfere) with classification in another modality and describe a neural network algorithm that is able to take advantage of the structure between the pattern distributions to different sensory modalities to eliminate the need for a teaching signal during training of each network. The algorithm is demonstrated on the problem of learning to recognize speech both acoustically and visually. Simultaneous presentation of moving mouth images and emanating sound waves allows the development of lip-reading and acoustic speech classifiers. The resulting classifiers approach the performance of supervised classifiers without requiring hand-labeling of the training patterns.

Citations

2138 Learning Internal Representations by Error Propagation – Rumelhart, Hinton, et al. - 1986
780 Self-organized formation of topologically correct feature maps – Kohonen - 1982
271 Adaptive pattern classification and universal recoding: parallel development and coding of neural feature detectors – Grossberg - 1976
222 Feature discovery by competitive learning – Rumelhart, Zipser - 1985
158 Hearing lips and seeing voices – MCGURK, MACDONALD - 1976
115 Self-organizing neural network that discovers surfaces in random-dot stereograms – Becker, Hinton - 1992
94 Visual contribution of speech intelligibility in noise – Sumby, Pollack - 1954
76 Soft competitive adaptation: Neural network learning algorithms based on fitting statistical mixtures. Unpublished doctoral dissertation – Nowlan - 1991
75 Maximum likelihood competitive learning – Nowlan - 1990
63 Improved versions of learning vector quantization – Kohonen
55 Mutual information maximization: Models of cortical self organization – Becker - 1996
55 Hearing lips and seeing voices. Nature – McGurk, MacDonald
29 State dependent activity in monkey visual cortex I: Single cell activity – Haenny, Schiller - 1988
27 Category learning through multimodality sensing – Sa, Ballard - 1998
26 The representation and storage of information in neuronal networks in the primate cerebral cortex and hippocampus – Rolls - 1989
25 Learning Classification with Unlabelled Data – Sa - 1994
20 The discovery of structure by multi-stream networks of local processors with contextual guidance – Phillips, Kay, et al. - 1995
19 Temporal texture and activity recognition – Polana, Nelson - 1997
18 Visual influences on speech perception processes – MacDonald, McGurk - 1978
17 Discovering predictable classifications – Schmidhuber, Prelinger - 1993
15 Effect of cooling area 18 on striate cortex cells in the squirrel monkey – Sandell, Schiller - 1982
14 Unsupervised Classification Learning from Cross-Modal Environmental Structure – Sa - 1994
11 Integration of auditory information in the cat's visual cortex – Fishman, Michael - 1973
9 Contextually guided unsupervised learning using local multivariate binary processors – Kay, Floreano, et al. - 1998
7 Visual learning in the perception of texture: Simple and contingent aftereffects of texture density – Durgin, Proffitt - 1996
7 Extraretinal representations in area V4 of macaque monkey – Maunsell, Sclar, et al. - 1991
6 Minimizing disagreement for self-supervised classification – Sa - 1994
6 Neuronal convergence of noxious, acoustic and visual stimuli in the visual cortex of the cat – Murata, Cramer, et al. - 1965
6 Auditory Specificity in Unit Recordings from Cat's Visual Cortex – Spinelli, Starr, et al. - 1968
5 Responses somesthesiques, visuel et auditives, recuellies, au niveau du cortex "associatif " infrasylvien chez le chat curarise non anesthesie – Buser, Borenstein - 1959
5 Self-teaching through correlated input – Sa, Ballard - 1992
5 Contingent aftereffects of texture density: Perceptual learning and contingency – Durgin - 1995
5 Visual System's View of Acoustic Space – Morrell - 1972
4 The experimental development of color-tone synesthesia – Howells - 1944
4 Macro and microelectrode studies of somatic responses in the lateral geniculate body – Meulders, Colle, et al. - 1965
4 Visual system's view of acoustic space. Nature – Morrell - 1972
4 Receptive field organization of ganglion cells in the cat's retina," Experimental Neurology – Spinelli - 1967
4 Centrifugal optic nerve responses evoked by auditory and somatic stimulation," Experimental Neurology – Spinelli, Pribram, et al. - 1965
4 Afferent and efferent activity in single units of the cat's optic nerve," Experimental Neurology – Spinelli, Weingarten - 1966
2 Pattern classification by the bayes machine – Diamantini, Spalvieri - 1995
1 The intermodal representation of speech in infants. Infant Behavior and Development – Kuhl, Meltzoff - 1984
1 Neural network definitions of highly predictable protein secondary structure classes – Lapedes, Steeg, et al. - 1994