--- abstract: |- The problem of representing large sets of complex state sequences (CSSs)---i.e., sequences in which states can recur multiple times---has thus far resisted solution. This paper describes a novel neural network model, TEMECOR, which has very large capacity for storing CSSs. Furthermore, in contrast to the various back-propagation-based attempts at solving the CSS problem, TEMECOR requires only a single presentation of each sequence. TEMECOR's power derives from a) its use of a combinatorial, distributed representation scheme, and b) its method of choosing internal representations of states at random. Simulation results are presented which show that the number of spatio-temporal binary feature patterns which can be stored to some criterion accuracy (e.g., 97%) increases faster-than-linearly in the size of the network. This is true for both uncorrelated and correlated pattern sets, although the rate is slightly slower for correlated patterns. altloc: - http://home.comcast.net/~rinkus/wcnn95.pdf chapter: ~ commentary: ~ commref: ~ confdates: 'July 17-21, 1995' conference: The 1995 World Congress on Neural Networks confloc: 'Washington, D.C.' contact_email: ~ creators_id: [] creators_name: - family: Rinkus given: Gerard J. honourific: '' lineage: '' date: 1995 date_type: published datestamp: 2004-03-06 department: ~ dir: disk0/00/00/34/72 edit_lock_since: ~ edit_lock_until: ~ edit_lock_user: ~ editors_id: [] editors_name: [] eprint_status: archive eprintid: 3472 fileinfo: /style/images/fileicons/application_pdf.png;/3472/1/wcnn95.pdf full_text_status: public importid: ~ institution: ~ isbn: ~ ispublished: pub issn: ~ item_issues_comment: [] item_issues_count: 0 item_issues_description: [] item_issues_id: [] item_issues_reported_by: [] item_issues_resolved_by: [] item_issues_status: [] item_issues_timestamp: [] item_issues_type: [] keywords: spatiotemporal pattern memory sequence associative lastmod: 2011-03-11 08:55:29 latitude: ~ longitude: ~ metadata_visibility: show note: ~ number: ~ pagerange: 442-448 pubdom: FALSE publication: ~ publisher: 'Lawrence Erlbaum Associates, Inc. and INNS Press' refereed: TRUE referencetext: | Cleeremans, A. (1993) Mechanisms of Implicit Learning: Connectionist Models of Sequence Processing. A Bradford Book, The MIT Press, Cambridge, MA. Elman, J. L. (1990) “Finding Structure in Time” Cognitive Science, 14, 179-212. Guyon, I., Personnaz, L., & Dreyfus, G. (1988) “Of points and loops” In Eckmiller, R. & Malsburg, C.v.d. (Eds.) Neural Computers, NATO ASI Series, Vol. F41, 261-269. Springer-Verlag, Berlin, Germany. Jordan, M. I. (1986) “Serial Order” Tech. Rep. 8604, Institute for Cognitive Science, University of California, San Diego, CA. McCloskey, M. & Cohen, N. J. (1989) “Catastrophic Interference in Connectionist Networks: The Sequential Learning Problem”, In The Psychology of Learning and Memory Vol. 24. Bower, G. H. (Ed.) Academic Press. 109-165. Rinkus, G. (1993) “Context-sensitive Spatio-temporal Memory” Tech. Rep. CAS/CNS-TR-93-031, Dept. of Cognitive and Neural Systems, Boston University, Boston, MA Rinkus, G. (1995) A Combinatorial Neural Network Exhibiting both Episodic Memory and Generalization for Spatio-Temporal Patterns. Ph.D. Thesis, Graduate School of Arts and Sciences, Boston University. In Progress. Smith, A. W. & Zipser, D. (1989) “Learning Sequential Structure with the Real-Time Recurrent Learning Algorithm” International Journal of Neural Systems, 1, 125-131. Williams, R.J. & Zipser, D. (1989) “A learning algorithm for continually running fully recurrent neural networks” Neural Computation, 1, 270-280. relation_type: [] relation_uri: [] reportno: ~ rev_number: 12 series: ~ source: ~ status_changed: 2007-09-12 16:51:09 subjects: - comp-neuro-sci - comp-sci-neural-nets succeeds: ~ suggestions: ~ sword_depositor: ~ sword_slug: ~ thesistype: ~ title: 'TEMECOR: An Associative, Spatio-temporal Pattern Memory for Complex State Sequences' type: confpaper userid: 4757 volume: ~