Unnamed: 0 int64 0 7.24k | id int64 1 7.28k | raw_text stringlengths 9 124k | vw_text stringlengths 12 15k |
|---|---|---|---|
400 | 1,365 | 2D Observers for Human 3D Object Recognition?
Zili Liu
NEC Research Institute
Daniel Kersten
University of Minnesota
. Abstract
Converging evidence has shown that human object recognition
depends on familiarity with the images of an object. Further,
the greater the similarity between objects, the stronger is the
dep... | 1365 |@word neurophysiology:1 trial:2 wiesel:3 stronger:1 d2:2 simulation:6 decomposition:1 tr:5 shading:2 vigorously:1 initial:1 liu:6 disparity:26 exclusively:2 daniel:2 tuned:3 past:1 current:1 surprising:1 hpp:2 activation:2 dx:1 must:3 finest:1 physiol:2 confirming:1 plot:2 v:3 discrimination:3 cue:1 selected:2 yr... |
401 | 1,366 | Blind Separation of Radio Signals
Fading Channels
?
In
Kari Torkkola
Motorola, Phoenix Corporate Research Labs,
2100 E. Elliot Rd, MD EL508, Tempe, AZ 85284, USA
email: A540AA(Qemail.mot.com
Abstract
We apply information maximization / maximum likelihood blind
source separation [2, 6) to complex valued signals mixe... | 1366 |@word bounced:1 kong:1 version:2 seems:1 nd:1 simulation:3 carry:1 reduction:1 fragment:1 imaginary:1 current:3 com:1 yet:1 must:2 visible:1 wx:1 predetermined:1 shape:1 discernible:1 update:2 half:1 plane:6 beginning:2 short:2 awex:2 coarse:1 qam:8 location:2 constructed:2 direct:1 differential:1 sii:1 hermitian... |
402 | 1,367 | An Analog VLSI Model of the Fly
Elementary Motion Detector
Reid R. Harrison and Christof Koch
Computation and Neural Systems Program, 139-74
California Institute of Technology
Pasadena, CA 91125
[harrison,koch]@klab.caltech.edu
Abstract
Flies are capable of rapidly detecting and integrating visual motion information... | 1367 |@word h:1 version:5 r:1 series:1 past:1 bradley:1 current:18 follower:1 must:2 exposing:1 physiol:1 visible:1 remove:2 designed:1 device:1 compo:1 filtered:1 detecting:1 node:3 along:2 direct:1 supply:1 ect:2 behavior:5 aliasing:3 borst:9 actual:1 begin:1 retinotopic:1 estimating:1 matched:1 circuit:50 mass:1 dev... |
403 | 1,368 | Computing with Action Potentials
John J. Hopfield*
Carlos D. Brody t
Sam Roweis t
Abstract
Most computational engineering based loosely on biology uses continuous variables to represent neural activity. Yet most neurons communicate with action potentials. The engineering view is equivalent to using
a rate-code for ... | 1368 |@word eliminating:1 interleave:1 seems:1 pulse:6 excited:1 mammal:1 initial:1 contains:1 interestingly:1 odour:1 current:1 nt:1 yet:1 readily:1 john:1 periodically:1 speakerindependent:1 shape:2 beginning:1 short:6 along:1 burst:1 ik:1 pathway:1 olfactory:7 indeed:1 behavior:1 roughly:2 brain:1 integrator:1 inspi... |
404 | 1,369 | Learning Continuous Attractors in
Recurrent Networks
H. Sebastian Seung
Bell Labs, Lucent Technologies
Murray Hill, NJ 07974
seung~bell-labs.com
Abstract
One approach to invariant object recognition employs a recurrent neural network as an associative memory. In the standard depiction of the
network's state space, mem... | 1369 |@word neurophysiology:1 version:2 compression:1 linearized:1 contrastive:1 pressure:1 initial:3 suppressing:1 com:1 surprising:1 must:1 cottrell:2 visible:10 extensional:1 shape:1 motor:1 remove:1 update:3 short:6 location:4 simpler:1 zhang:1 along:5 rnl:1 retrieving:1 consists:1 unlearning:1 inside:1 baldi:1 ind... |
405 | 137 | 720
AN ELECTRONIC PHOTORECEPTOR
SENSITIVE TO SMALL CHANGES IN INTENSITY
T. Delbriick and C. A. Mead
256-80 Computer Science
California Institute of Technology
Pasadena, CA 91125
ABSTRACT
We describe an electronic photoreceptor circuit that is sensitive to
small changes in incident light intensity. The sensitivity to ... | 137 |@word trial:5 version:2 cm2:1 pressed:1 initial:2 series:1 past:2 subjective:1 current:17 comparing:1 reminiscent:1 drop:1 v:2 half:1 device:1 short:1 filtered:2 height:1 direct:1 incorrect:1 consists:2 sustained:1 manner:1 lov:1 expected:1 roughly:1 abscissa:1 td:1 little:3 becomes:2 brightly:1 unrelated:2 circui... |
406 | 1,370 | MELONET I: Neural Nets for Inventing
Baroque-Style Chorale Variations
Dominik Hornel
dominik@ira.uka.de
Institut fur Logik, Komplexitat und Deduktionssysteme
Universitat Fridericiana Karlsruhe (TH)
Am Fasanengarten 5
D-76128 Karlsruhe, Germany
Abstract
MELONET I is a multi-scale neural network system producing
baroqu... | 1370 |@word kong:1 illustrating:1 middle:1 nd:1 simulation:2 decomposition:1 recursively:1 initial:2 contains:1 feulner:3 current:2 nt:2 activation:1 written:1 must:2 realize:1 shape:1 update:2 beginning:1 wolfram:1 denis:1 successive:1 harmonize:1 height:1 harmonically:1 mgt:2 predecessor:1 behavior:2 multi:5 automati... |
407 | 1,371 | Detection of first and second order motion
Alexander Grunewald
Division of Biology
California Institute of Technology
Mail Code 216-76
Pasadena, CA 91125
alex@vis.caltech.edu
Heiko Neumann
Abteilung Neuroinformatik
Vniversitat VIm
89069 VIm
Germany
hneumann@neuro.informatik.uni-ulm.de
Abstract
A model of motion detec... | 1371 |@word middle:1 wiesel:2 grey:2 simulation:2 simplifying:1 excited:2 contains:1 tuned:1 neurophys:2 activation:2 physiol:2 plasticity:1 plot:4 half:2 postnatal:1 filtered:1 provides:1 detecting:1 location:5 preference:3 along:2 direct:3 differential:1 become:1 grunewald:4 inside:2 introduce:1 manner:1 freeman:1 de... |
408 | 1,372 | From Regularization Operators
to Support Vector Kernels
Alexander J. Smola
Bernhard Scholkopf
GMDFIRST
Rudower Chaussee 5
12489 Berlin, Germany
smola@first.gmd.de
Max-Planck-Institut fur biologische Kybernetik
Spemannstra.Be 38
72076 Ttibingen, Germany
bs-@mpik-tueb.mpg.de
Abstract
We derive the correspondence bet... | 1372 |@word rreg:2 briefly:1 inversion:2 polynomial:3 norm:1 tedious:1 decomposition:2 pg:1 thereby:2 solid:1 contains:1 denoting:3 comparing:2 com:1 dx:2 written:2 additive:1 kdd:1 girosi:10 offunctions:1 wll:1 leaf:1 vanishing:1 provides:2 math:1 gx:4 c2:2 differential:1 scholkopf:10 prove:1 eiw:1 mpg:1 curse:1 confu... |
409 | 1,373 | Globally Optimal On-line Learning Rules
Magnus Rattray*and David Saad t
Department of Computer Science & Applied Mathematics,
Aston University, Birmingham B4 7ET, UK.
Abstract
We present a method for determining the globally optimal on-line
learning rule for a soft committee machine under a statistical mechanics fram... | 1373 |@word seems:1 instrumental:1 km:2 bn:3 covariance:2 solid:1 carry:1 reduction:3 configuration:1 exclusively:1 outperforms:1 activation:13 written:1 must:1 plot:1 provides:2 node:13 ron:1 sigmoidal:1 differential:1 supply:1 ik:4 theoretically:1 expected:6 frequently:1 mechanic:5 examine:1 globally:17 window:1 more... |
410 | 1,374 | Just One View:
Invariances in Inferotemporal Cell Thning
Maximilian Riesenhuber
Tomaso Poggio
Center for Biological and Computational Learning and
Department of Brain and Cognitive Sciences
Massachusetts Institute of Techno)ogy, E25-201
Cambridge, MA 02139
{max,tp }@ai.mit.edu
Abstract
In macaque inferotemporal corte... | 1374 |@word neurophysiology:1 version:2 middle:2 stronger:1 simulation:3 tried:2 excited:1 pick:1 fonn:1 thereby:1 phy:1 tuned:11 interestingly:1 perret:1 reynolds:1 current:2 comparing:2 anterior:2 neurophys:3 ka:1 must:1 subsequent:1 realistic:1 shape:4 plot:5 progressively:1 stationary:1 selected:1 beginning:1 locat... |
411 | 1,375 | Classification by Pairwise Coupling
*
Stanford University
and
TREVOR HASTIE
ROBERT TIBSHIRANI t
University of Toronto
Abstract
We discuss a strategy for polychotomous classification that involves
estimating class probabilities for each pair of classes, and then coupling the estimates together. The coupling model is... | 1375 |@word version:2 briefly:1 proportion:1 seems:1 logit:2 neigbours:1 simulation:1 accounting:1 covariance:3 carry:1 contains:1 score:2 bradley:5 ida:3 must:1 plot:1 drop:1 update:1 discrimination:1 guess:1 item:1 argm:1 recompute:1 provides:1 math:1 toronto:3 preference:2 simpler:1 constructed:3 qualitative:1 prove... |
412 | 1,376 | Agnostic Classification of Markovian
Sequences
Ran EI-Yaniv
Shai Fine
Naftali Tishby*
Institute of Computer Science and Center for Neural Computation
The Hebrew University
Jerusalem 91904, Israel
E-Dlail: {ranni,fshai,tishby}Ocs.huji.ac.il
Category: Algorithms.
Abstract
Classification of finite sequences without exp... | 1376 |@word uee:1 compression:10 norm:1 boundedness:1 inefficiency:1 substitution:1 series:1 existing:1 portuguese:1 john:2 analytic:1 remove:1 greedy:3 selected:1 vanishing:1 short:3 smith:4 unbounded:1 guard:1 direct:1 beta:1 introduce:2 pairwise:6 x60:1 provided:1 ocs:1 estimating:5 underlying:1 agnostic:11 israel:1... |
413 | 1,377 | Bidirectional Retrieval from Associative
Memory
Friedrich T. Sommer and Gunther Palm
Department of Neural Information Processing
University of Ulm, 89069 Ulm, Germany
{sommer,palm}~informatik.uni-ulm.de
Abstract
Similarity based fault tolerant retrieval in neural associative memories (N AM) has not lead to wiedesprea... | 1377 |@word trial:1 briefly:1 nd:1 simulation:3 heteroassociative:2 accounting:1 thereby:1 reduction:2 initial:8 contains:1 fragment:1 denoting:1 existing:1 activation:1 cruz:1 analytic:1 drop:1 intelligence:1 yr:1 reciprocal:1 provides:1 contribute:1 accessed:1 acti:1 pathway:1 abscissa:1 growing:1 decomposed:1 little... |
414 | 1,378 | Using Expectation to Guide Processing:
A Study of Three Real-World Applications
Shumeet 8aluja
Justsystem Pittsburgh Research Center &
School of Computer Science, Carnegie Mellon University
baluja@cs.cmu.edu
Abstract
In many real world tasks, only a small fraction of the available inputs are important
at any particula... | 1378 |@word middle:2 eliminating:1 compression:1 proportion:1 simulation:1 accounting:1 fonn:1 deems:1 initial:1 current:9 ferrier:2 nowlan:2 activation:13 must:6 takeo:1 cottrell:2 visible:1 discernible:1 remove:3 half:1 filtered:3 provides:1 detecting:1 location:9 simpler:1 constructed:1 driver:1 inside:1 manner:2 in... |
415 | 1,379 | Shared Context Probabilistic Transducers
Yoshua Bengio*
Dept. IRO ,
Universite de Montreal,
Montreal (QC) , Canada, H3C 3J7
bengioyOiro.umontreal.ca
Samy Bengio t
Microcell Labs,
1250 , Rene Levesque Ouest,
Montreal (QC) , Canada, H3B 4W8
samy.bengioOmicrocell.ca
Jean-Fran~ois Isabelle t
Microcell Labs,
1250, Rene ... | 1379 |@word kong:1 version:1 tried:1 recursively:1 series:1 att:1 contains:2 ala:2 past:2 existing:1 current:1 com:1 comparing:1 written:1 must:1 update:3 leaf:1 guess:1 weighing:1 node:51 become:1 transducer:37 descendant:1 upenn:1 decomposed:1 becomes:1 baker:2 interpreted:1 string:3 nj:2 w8:2 every:5 growth:1 yn:9 b... |
416 | 138 | 402
MODELING THE OLFACTORY BULB
- COUPLED NONLINEAR OSCILLATORS
Zhaoping Lit
J. J. Hopfield?
t Division of Physics, Mathematics and Astronomy
?Division of Biology, and Division of Chemistry and Chemical Engineering
t? California Institute of Technology, Pasadena, CA 91125, USA
? AT&T Bell Laboratories
ABSTRACT
The ol... | 138 |@word private:1 version:1 eliminating:1 hippocampus:2 nd:2 ayy:3 simulation:9 pulse:3 wog:3 mammal:1 carry:1 initial:1 odour:1 must:1 physiol:1 j1:1 wx:2 discrimination:3 nervous:1 indicative:1 xk:1 ith:2 short:1 provides:1 location:1 simpler:1 mathematical:2 burst:1 pathway:2 behavioral:1 olfactory:35 mask:1 expe... |
417 | 1,380 | Estimating Dependency Structure as a Hidden
Variable
Marina Meill and Michael I. Jordan
{mmp, jordan}@ai.mit.edu
Center for Biological & Computational Learning
Massachusetts Institute of Technology
45 Carleton St. E25-201
Cambridge, MA 02142
Abstract
This paper introduces a probability model, the mixture of trees tha... | 1380 |@word trial:3 msr:1 version:1 compression:4 simulation:1 tried:2 thereby:1 recursively:1 initial:2 liu:4 selecting:1 t7:1 bitmap:1 current:1 recovered:1 nt:1 must:1 mst:5 visible:2 realistic:1 informative:1 mstep:1 meilii:1 designed:1 intelligence:3 node:2 height:1 direct:1 become:1 beta:1 tirri:1 retrieving:1 pr... |
418 | 1,381 | Hippocampal Model of Rat Spatial Abilities
Using Temporal Difference Learning
David J Foster*
Centre for Neuroscience
Edinburgh University
Richard GM Morris
Centre for Neuroscience
Edinburgh University
Peter Dayan
E25-210, MIT
Cambridge, MA 02139
Abstract
We provide a model of the standard watermaze task, and of a m... | 1381 |@word h:1 trial:19 version:1 hippocampus:16 extinction:1 open:3 gradual:2 simulation:4 r:4 shot:3 united:1 tuned:2 interestingly:1 past:1 current:3 si:1 refines:1 plasticity:3 shape:1 designed:1 update:1 instantiate:1 short:3 rgm:2 location:8 along:1 direct:1 become:1 replication:1 behavioral:1 manner:3 acquired:... |
419 | 1,382 | Hippocampal Model of Rat Spatial Abilities
Using Temporal Difference Learning
David J Foster*
Centre for Neuroscience
Edinburgh University
Richard GM Morris
Centre for Neuroscience
Edinburgh University
Peter Dayan
E25-210, MIT
Cambridge, MA 02139
Abstract
We provide a model of the standard watermaze task, and of a m... | 1382 |@word h:1 trial:19 version:1 hippocampus:16 extinction:1 open:5 termination:2 gradual:2 simulation:4 r:4 decomposition:7 tried:1 pick:1 tr:1 solid:1 shot:3 carry:1 reduction:1 configuration:3 contains:1 exclusively:1 united:1 tuned:2 interestingly:1 past:1 outperforms:1 current:7 si:1 jkl:2 john:1 refines:1 belmo... |
420 | 1,383 | The Asymptotic Convergence-Rate of
Q-Iearning
es. Szepesvari*
Research Group on Artificial Intelligence, "Jozsef Attila" University,
Szeged, Aradi vrt. tere 1, Hungary, H-6720
szepes@math.u-szeged.hu
Abstract
In this paper we show that for discounted MDPs with discount
factor, > 1/2 the asymptotic rate of convergence... | 1383 |@word mild:1 version:1 seems:1 stronger:1 norm:1 nd:1 hu:2 simulation:1 tr:2 boundedness:1 comparing:1 yet:1 must:2 john:1 belmont:1 update:1 stationary:6 intelligence:1 device:1 ith:1 lr:2 provides:1 math:1 firstly:1 simpler:1 along:1 direct:1 become:2 eleventh:2 indeed:1 multi:1 ol:1 discounted:3 decreasing:2 r... |
421 | 1,384 | Reinforcement Learning with
Hierarchies of Machines *
Ronald Parr and Stuart Russell
Computer Science Division, UC Berkeley, CA 94720
{parr,russell}@cs.berkeley.edu
Abstract
We present a new approach to reinforcement learning in which the policies considered by the learning process are constrained by hierarchies of
p... | 1384 |@word version:1 twelfth:1 open:1 closure:1 carolina:1 decomposition:6 fonn:1 dramatic:1 thereby:1 reduction:1 initial:6 contains:2 efficacy:1 selecting:2 unintended:1 current:3 yet:1 reminiscent:1 must:1 ronald:1 concatenate:1 partition:1 numerical:1 update:3 intelligence:4 hallway:6 short:1 provides:3 traverse:1... |
422 | 1,385 | The Observer-Observation Dilemma
in Neuro-Forecasting
Hans Georg Zimmermann
Ralph Neuneier
SiemensAG
Corporate Technology
D-81730 Munchen, Germany
Georg.Zimmermann@mchp.siemens.de
Siemens AG
Corporate Technology
D-81730 Munchen, Germany
Ralph.Neuneier@mchp.siemens.de
Abstract
We explain how the training data can b... | 1385 |@word trial:1 version:3 hsieh:1 profit:1 series:3 suppressing:1 neuneier:8 written:1 enables:1 sponsored:1 update:3 resampling:1 tenn:2 mathematical:1 consists:1 combine:1 manner:1 huber:1 market:3 expected:2 behavior:1 brain:3 td:1 actual:1 considering:2 project:1 underlying:1 easiest:1 interpreted:1 unified:5 a... |
423 | 1,386 | Automated Aircraft Recovery
via Reinforcement Learning:
Initial Experiments
Jeffrey F. Monaco
Barron Associates, Inc.
Jordan Building
1160 Pepsi Place, Suite 300
Charlottesville VA 22901
monaco@bainet.com
David G. Ward
Barron Associates, Inc.
Jordan Building
1160 Pepsi Place, Suite 300
Charlottesville VA 22901
ward@b... | 1386 |@word aircraft:30 trial:1 version:1 oae:1 twelfth:1 simulation:7 initial:28 uma:1 longitudinal:7 recovered:1 com:2 wd:4 buckingham:1 must:2 readily:1 subsequent:3 designed:3 atlas:1 update:4 selected:3 short:1 indefinitely:2 authority:3 attack:3 five:2 direct:2 rudder:1 falsely:1 multi:2 manager:1 notation:1 boun... |
424 | 1,387 | Local Dimensionality Reduction
Stefan Schaal
1,2,4
sschaal@usc.edu
http://www-slab.usc.edulsschaal
Sethu Vijayakumar 3, I
Christopher G. Atkeson 4
sethu@cs.titech.ac.jp
http://ogawawww.cs.titech.ac.jp/-sethu
cga@cc.gatech.edu
http://www.cc.gatech.edul
fac/Chris.Atkeson
IERATO Kawato Dynamic Brain Project (IST), ... | 1387 |@word trial:1 collinearity:1 briefly:1 version:3 inversion:2 advantageous:1 seems:1 simulation:1 jacob:2 covariance:5 recursively:1 reduction:15 interestingly:1 atlantic:1 comparing:1 surprising:1 must:1 additive:4 shape:1 plot:1 intelligence:3 guess:2 nervous:1 device:1 yr:1 provides:1 sigmoidal:1 along:1 wpls:2... |
425 | 1,388 | On Efficient Heuristic Ranking of
Hypotheses
Steve Chien, Andre Stechert, and Darren Mutz
Jet Propulsion Laboratory, California Institute of Technology
4800 Oak Grove Drive, MIS 525-3660, Pasadena, CA 91109-8099
steve.chien@jpl.nasa.gov, Voice: (818) 306-6144 FAX: (818) 306-6912
Content Areas: Applications (Stochastic... | 1388 |@word trial:4 nd:1 tadepalli:1 dekker:1 additively:3 u11:1 initial:1 configuration:1 selecting:2 genetic:1 existing:1 current:1 si:1 yet:1 must:3 additive:1 intelligence:1 selected:2 fewer:2 compo:1 provides:3 math:1 successive:1 oak:1 mathematical:1 differential:2 incorrect:3 consists:1 combine:4 pairwise:12 spa... |
426 | 1,389 | Local Dimensionality Reduction
Stefan Schaal
1,2,4
sschaal@usc.edu
http://www-slab.usc.edulsschaal
Sethu Vijayakumar 3, I
Christopher G. Atkeson 4
sethu@cs.titech.ac.jp
http://ogawawww.cs.titech.ac.jp/-sethu
cga@cc.gatech.edu
http://www.cc.gatech.edul
fac/Chris.Atkeson
IERATO Kawato Dynamic Brain Project (IST), ... | 1389 |@word collinearity:1 trial:1 version:3 middle:1 inversion:2 seems:3 proportion:1 advantageous:1 briefly:1 simulation:4 jacob:2 covariance:5 recursively:1 reduction:15 initial:1 substitution:1 interestingly:1 atlantic:1 reaction:1 comparing:1 surprising:1 activation:6 grapheme:2 must:3 written:1 cottrell:4 additiv... |
427 | 139 | 11
AN OPTIMALITY PRINCIPLE FOR
UNSUPERVISED LEARNING
Terence D. Sanger
MIT AI Laboratory, NE43-743
Cambridge, MA 02139
(tds@wheaties.ai.mit.edu)
ABSTRACT
We propose an optimality principle for training an unsupervised feedforward neural network based upon maximal
ability to reconstruct the input data from the network... | 139 |@word aircraft:1 simulation:1 seek:1 decomposition:3 contains:1 disparity:16 elliptical:1 nowlan:1 si:1 yet:1 john:1 physiol:1 shape:2 remove:1 sponsored:1 update:1 half:4 intelligence:3 plane:1 affair:1 ith:2 record:1 filtered:1 ire:1 quantized:3 node:1 location:1 successive:1 x128:1 qualitative:1 mask:14 tomaso:... |
428 | 1,390 | Online learning from finite training sets
in nonlinear networks
David Barber t
Peter Sollich*
Department of Physics
University of Edinburgh
Edinburgh ERg 3JZ, U.K.
Department of Applied Mathematics
Aston University
Birmingham B4 7ET, U.K.
P.Sollich~ed.ac.uk
D.Barber~aston . ac.uk
Abstract
Online learning is one o... | 1390 |@word eor:1 km:6 seek:1 simulation:11 covariance:3 carry:1 kappen:1 initial:4 comparing:1 nt:1 activation:18 yet:1 additive:1 realistic:1 happen:1 remove:1 update:10 v:3 alone:1 selected:1 yr:3 provides:2 successive:2 kinh:1 become:2 differential:1 qualitative:1 incorrect:1 specialize:1 manner:1 expected:1 mechan... |
429 | 1,391 | Effects of Spike Timing Underlying
Binocular Integration and Rivalry in a
Neural Model of Early Visual Cortex
Erik D. Lumer
Wellcome department of Cognitive Neurology
Institute of Neurology, University College of London
12 Queen Square, London, WC1N 3BG, UK
Abstract
In normal vision, the inputs from the two eyes are ... | 1391 |@word determinant:1 version:1 middle:1 nd:1 confirms:1 simulation:5 moment:1 extrastriate:3 configuration:1 series:1 selecting:1 com:1 activation:2 physiol:1 tilted:6 visible:1 shape:1 motor:1 plot:4 v:20 exl:1 patterning:5 accordingly:1 preference:1 relayed:1 along:2 differential:5 consists:3 pathway:8 combine:1... |
430 | 1,392 | Approximating Posterior Distributions
in Belief Networks using Mixtures
Christopher M. Bishop
Neil Lawrence
Neural Computing Research Group
Dept. Computer Science & Applied Mathematics
Aston University
Binningham, B4 7ET, U.K.
Tommi Jaakkola
Michael I. Jordan
Center for Biological and Computational Learning
Massa... | 1392 |@word briefly:1 polynomial:1 seek:1 configuration:6 contains:1 si:3 yet:1 written:1 visible:3 plot:2 tenn:1 intelligence:2 inam:1 lr:1 node:1 location:1 along:1 prove:1 introduce:3 indeed:1 multi:1 considering:1 bounded:1 factorized:3 exactly:1 unit:6 omit:1 appear:1 frey:1 treat:1 initialization:1 factorization:... |
431 | 1,393 | Comparison of Human and Machine Word
Recognition
M. Schenkel
Dept of Electrical Eng.
University of Sydney
Sydney, NSW 2006, Australia
schenkel@sedal.usyd.edu.au
C. Latimer
Dept of Psychology
University of Sydney
Sydney, NSW 2006, AustTalia
M. Jabri
Dept of Electrical Eng.
University of Sydney
Sydney, NSW 2006, Austr... | 1393 |@word cnn:1 middle:1 seems:1 grey:1 eng:2 nsw:3 pick:2 paid:1 initial:3 substitution:1 series:1 score:10 selecting:1 contains:3 document:10 legality:7 interestingly:2 current:1 contextual:1 lang:1 additive:1 matured:1 blur:1 shape:2 discrimination:1 half:2 leaf:2 accordingly:2 inspection:1 short:3 height:1 direct... |
432 | 1,394 | Combining Classifiers Using
Correspondence Analysis
Christopher J. Merz
Dept. of Information and Computer Science
University of California, Irvine, CA 92697-3425 U.S.A.
cmerz@ics.uci.edu
Category: Algorithms and Architectures.
Abstract
Several effective methods for improving the performance of a single learning algo... | 1394 |@word kong:3 repository:2 version:2 decomposition:1 xtest:2 dramatic:1 reduction:1 loc:1 existing:1 must:3 fn:1 partition:6 resampling:1 alone:1 v:9 oblique:1 record:1 caveat:1 boosting:2 five:1 direct:1 become:1 consists:2 combine:6 redefine:1 alspector:1 frequently:1 little:2 actual:1 provided:1 underlying:1 ma... |
433 | 1,395 | A 1,OOO-Neuron System with One
Million 7-bit Physical Interconnections
Yuzo Hirai
Institute of Information Sciences and Electronics
University of Tsukuba
1-1-1 Ten-nodai, Tsukuba, Ibaraki 305, Japan
e-mail: hirai@is.tsukuba.ac.jp
Abstract
An asynchronous PDM (Pulse-Density-Modulating) digital neural
network system ha... | 1395 |@word version:1 loading:2 nd:1 open:1 donham:1 pulse:21 paid:1 initial:1 configuration:1 contains:1 electronics:1 analysed:1 written:1 realize:2 interrupted:1 msb:1 rc:1 along:1 differential:2 become:1 consists:3 inside:1 manner:1 alspector:1 behavior:2 integrator:1 terminal:2 ol:2 company:1 little:1 increasing:1... |
434 | 1,396 | Multiple Threshold Neural Logic
Jehoshua Bruck
Vasken Bohossian
~mail:
California Institute of Technology
Mail Code 136-93
Pasadena, CA 91125
{vincent, bruck}~paradise.caltech.edu
Abstract
We introduce a new Boolean computing element related to the Linear Threshold element, which is the Boolean version of the neur... | 1396 |@word version:3 polynomial:7 simulation:1 xiy:1 xiyi:1 written:1 chicago:1 hajnal:3 intelligence:1 ptm:2 characterization:2 math:4 five:1 unbounded:3 along:1 consists:3 prove:3 interscience:1 introduce:2 indeed:1 behavior:1 growing:1 multi:1 brain:1 ptb:1 inspired:1 goldman:3 provided:2 bounded:1 circuit:38 what:... |
435 | 1,397 | Function Approximat.ion with the
Sweeping Hinge Algorithm
Don R. Hush, Fernando Lozano
Dept. of Elec. and Compo Engg.
University of New Mexico
Albuquerque, NM 87131
Bill Horne
MakeWaves, Inc.
832 Valley Road
Watchung, NJ 07060
Abstract
We present a computationally efficient algorithm for function approximation with ... | 1397 |@word h:3 middle:2 knd:1 polynomial:2 norm:1 nd:5 suitably:1 seek:1 simplifying:1 minus:2 tr:1 initial:5 pub:1 tuned:1 current:3 xiyi:1 si:1 must:2 grahm:1 fn:8 numerical:2 partition:40 engg:1 subsequent:1 designed:2 plot:1 update:5 greedy:1 ria:1 compo:1 provides:1 node:20 location:2 revisited:1 hyperplanes:6 si... |
436 | 1,398 | EM Algorithms for PCA and SPCA
Sam Roweis?
Abstract
I present an expectation-maximization (EM) algorithm for principal
component analysis (PCA). The algorithm allows a few eigenvectors and
eigenvalues to be extracted from large collections of high dimensional
data. It is computationally very efficient in space and ti... | 1398 |@word determinant:1 version:1 inversion:3 compression:2 norm:1 covariance:37 decomposition:1 fonn:1 asks:1 pick:1 tr:1 solid:1 accommodate:1 shot:3 reduction:1 initial:1 series:1 xiy:1 denoting:1 diagonalized:1 current:3 yet:1 must:4 written:2 john:2 additive:1 informative:1 shape:1 update:2 generative:2 fewer:1 ... |
437 | 1,399 | Task and Spatial Frequency Effects on Face
Specialization
Matthew N. Dailey
Garrison W. Cottrell
Department of Computer Science and Engineering
U.C. San Diego
La Jolla, CA 92093-0114
{mdailey,gary}@cs.ucsd.edu
Abstract
There is strong evidence that face processing is localized in the brain.
The double dissociation... | 1399 |@word middle:3 briefly:1 gradual:1 jacob:5 covariance:1 tr:1 initial:2 configuration:1 contains:1 hereafter:2 empath:1 tuned:1 current:3 nt:1 nowlan:2 neurophys:1 cottrell:8 shape:1 plot:1 drop:3 discrimination:3 infant:4 cue:3 filtered:6 coarse:1 preference:1 five:2 differential:1 specialize:1 combine:1 acquired... |
438 | 14 | 301
ENCODING GEOMETRIC INVARIANCES IN
HIGHER-ORDER NEURAL NETWORKS
C.L. Giles
Air Force Office of Scientific Research, Bolling AFB, DC 20332
R.D. Griffin
Naval Research Laboratory, Washington, DC
20375-5000
T. Maxwell
Sachs-Freeman Associates, Landover, MD 20785
ABSTRACT
We describe a method of constructing higher-o... | 14 |@word seems:3 retraining:1 simulation:9 simplifying:1 moment:1 substitution:1 contains:2 loc:1 selecting:1 yet:1 must:7 readily:1 reminiscent:1 numerical:1 subsequent:1 intelligence:1 device:1 ith:1 math:1 simpler:1 mathematical:1 constructed:3 persistent:1 manner:3 behavior:1 multi:3 freeman:1 encouraging:1 increa... |
439 | 140 | 20
ASSOCIATIVE LEARNING
VIA INHIBITORY SEARCH
David H. Ackley
Bell Communications Research
Cognitive Science Research Group
ABSTRACT
ALVIS is a reinforcement-based connectionist architecture that
learns associative maps in continuous multidimensional environments. The discovered locations of positive and negative rei... | 140 |@word version:1 stronger:1 seems:2 a8i:2 open:2 simulation:1 propagate:2 uncovers:1 t_:1 initial:1 configuration:19 contains:3 current:18 activation:5 si:6 yet:1 must:1 cheap:1 motor:2 leaf:1 selected:3 beginning:1 short:1 granting:1 record:2 supplying:2 dissertation:1 provides:2 location:4 five:2 become:2 burr:2 ... |
440 | 1,400 | Synchronized Auditory and Cognitive 40 Hz
Attentional Streams, and the Impact of
Rhythmic Expectation on Auditory Scene Analysis
Bill Baird
Dept Mathematics, U.C.Berkeley, Berkeley, Ca. 94720.
baird@math.berkeley.edu
Abstract
We have developed a neural network architecture that implements a theory of attention, learn... | 1400 |@word trial:2 hippocampus:1 simulation:4 attended:5 pick:1 thereby:1 reduction:2 moment:2 series:2 contains:1 current:1 activation:3 must:4 interrupted:1 distant:2 arrayed:1 motor:13 hypothesize:3 discrimination:1 infant:1 tenn:1 selected:1 pacemaker:1 nervous:1 tone:50 alone:1 plane:1 short:2 core:1 pointer:1 ma... |
441 | 1,401 | Coding of Naturalistic Stimuli by
Auditory Midbrain Neurons
H. Attias* and C.E. Schreiner t
Sloan Center for Theoretical Neurobiology and
W.M. Keck Foundation Center for Integrative Neuroscience
University of California at San Francisco
San Francisco, CA 94143-0444
Abstract
It is known that humans can make finer disc... | 1401 |@word h:4 nd:1 bf:3 integrative:1 solid:7 carry:1 phy:2 imaginary:1 comparing:1 si:1 slb:1 enables:1 hypothesize:1 plot:1 discrimination:3 v:2 nervous:1 tone:2 short:1 lr:1 detecting:1 provides:2 location:2 simpler:1 along:1 constructed:1 become:1 consists:2 manner:1 theoretically:1 indeed:1 examine:1 window:1 be... |
442 | 1,402 | The Rectified Gaussian Distribution
N. D. Socci, D. D. Lee and H. S. Seung
Bell Laboratories, Lucent Technologies
Murray Hill, NJ 07974
{ndslddleelseung}~bell-labs.com
Abstract
A simple but powerful modification of the standard Gaussian distribution is studied. The variables of the rectified Gaussian are
constrained t... | 1402 |@word neurophysiology:1 version:3 polynomial:2 covariance:6 initial:1 cyclic:1 configuration:3 com:1 yet:1 must:6 belmont:1 wx:1 analytic:1 motor:2 designed:1 depict:1 update:5 stationary:5 generative:1 discovering:1 slowing:1 short:1 location:2 zhang:1 along:3 become:1 introduce:1 roughly:1 behavior:3 multi:1 br... |
443 | 1,403 | Modeling Complex Cells in an A wake
Macaque During Natural Image Viewing
William E. Vinje
vinjeCsocrates.berkeley.edu
Department of Molecular and
Cellular Biology, Neurobiology Division
University of California, Berkeley
Berkeley, CA, 94720
Jack L. Gallant
gallantCsocrates.berkeley.edu
Department of Psychology
Univers... | 1403 |@word neurophysiology:1 approved:1 simplifying:1 accounting:1 pg:1 pick:1 dramatic:1 extrastriate:1 valois:4 series:1 score:1 tuned:1 yet:2 dx:1 realistic:1 discrimination:1 alone:2 half:3 cue:1 beginning:1 characterization:1 location:1 psth:10 five:1 alert:1 become:1 consists:2 fixation:5 sustained:1 fitting:5 r... |
444 | 1,404 | Reinforcement Learning for Continuous
Stochastic Control Problems
Remi Munos
CEMAGREF, LISC, Pare de Tourvoie,
BP 121, 92185 Antony Cedex, FRANCE.
Rerni.Munos@cemagref.fr
Paul Bourgine
Ecole Polyteclmique, CREA,
91128 Palaiseau Cedex, FRANCE.
Bourgine@poly.polytechnique.fr
Abstract
This paper is concerned with the pr... | 1404 |@word open:1 kus90j:2 covariance:1 contraction:1 bourgine:5 initial:3 series:1 ecole:1 kry80j:2 current:2 discretization:2 ixj:1 ka:2 dx:1 fn:3 numerical:2 update:2 intelligence:2 ficial:1 xk:20 lr:4 finitedifference:1 successive:3 c6:1 zii:1 c2:14 differential:3 zkj:1 ik:1 prove:3 hjb:2 discretized:1 bellman:3 d... |
445 | 1,405 | Hybrid NNIHMM-Based Speech Recognition
with a Discriminant Neural Feature Extraction
Daniel Willett, Gerhard RigoU
Department of Comfuter Science
Faculty of Electrica Engineering
Gerhard-Mercator-University Duisburg, Germany
{willett,rigoll}@tb9-ti.uni-duisburg.de
Abstract
In this paper, we present a novel hybrid arc... | 1405 |@word faculty:1 oae:1 stronger:1 lwk:4 covariance:1 profit:1 xlw:5 ld:1 feb91:1 reduction:2 daniel:1 denoting:1 past:6 outperforms:3 wd:1 activation:1 discrimination:1 rrt:2 cook:1 provides:1 quantizer:1 codebook:2 node:4 five:1 consists:2 prove:2 combine:4 umbach:1 manner:1 behavior:1 multi:4 considering:3 incre... |
446 | 1,406 | Bayesian Robustification for Audio Visual
Fusion
Javier Movellan *
movellanOcogsci.ucsd.edu
Department of Cognitive Science
University of California, San Diego
La Jolla, CA 92092-0515
Paul Mineiro
pmineiroOcogsci.ucsd.edu
Department of Cognitive Science
University of California, San Diego
La Jolla, CA 92092-0515
Abst... | 1406 |@word trial:1 norm:1 loading:1 nd:1 open:1 grey:1 prasad:2 tried:1 covariance:1 decomposition:1 speechreading:3 hereafter:1 current:1 contextual:2 cottrell:1 drop:1 resampling:1 stationary:6 alone:1 lexicon:2 along:1 differential:1 symposium:1 consists:2 combine:2 redefine:1 expected:2 alspector:1 blackbox:1 audi... |
447 | 1,407 | Reinforcement Learning for Continuous
Stochastic Control Problems
Remi Munos
CEMAGREF, LISC, Pare de Tourvoie,
BP 121, 92185 Antony Cedex, FRANCE.
Rerni.Munos@cemagref.fr
Paul Bourgine
Ecole Polyteclmique, CREA,
91128 Palaiseau Cedex, FRANCE.
Bourgine@poly.polytechnique.fr
Abstract
This paper is concerned with the pr... | 1407 |@word trial:3 proportion:8 nd:1 open:1 verrelst:3 kus90j:2 simulation:1 contraction:1 covariance:1 jacob:1 bourgine:5 solid:3 initial:3 series:1 score:1 ecole:1 kry80j:2 subjective:1 current:2 discretization:2 ixj:1 ka:2 marquardt:1 activation:2 comparing:2 dx:1 john:3 fn:5 numerical:4 benign:18 enables:1 plot:4 ... |
448 | 1,408 | New Approximations of Differential
Entropy for Independent Component
Analysis and Projection Pursuit
Aapo Hyvarinen
Helsinki University of Technology
Laboratory of Computer and Information Science
P.O. Box 2200, FIN-02015 HUT, Finland
Email: aapo.hyvarinen<Ohut.fi
Abstract
We derive a first-order approximation of the... | 1408 |@word cu:1 version:1 polynomial:6 calculus:1 simulation:1 decomposition:1 solid:2 negentropy:5 dx:5 must:3 john:1 numerical:1 shape:1 enables:2 pertinent:1 plot:1 cook:2 xk:1 tcp:1 provides:1 firstly:1 simpler:1 c2:1 differential:16 introduce:1 theoretically:1 huber:1 ica:5 indeed:3 growing:1 company:1 mass:1 wha... |
449 | 1,409 | Generalized Prioritized Sweeping
David Andre
Nir Friedman
Ronald Parr
Computer Science Division, 387 Soda Hall
University of California, Berkeley, CA 94720
{dandre,nir,parr}@cs.berkeley.edu
Abstract
Prioritized sweeping is a model-based reinforcement learning method
that attempts to focus an agent's limited comput... | 1409 |@word msr:1 briefly:1 version:1 middle:1 tadepalli:1 minus:1 tr:1 carry:1 contains:2 fragment:1 unintended:1 past:1 existing:1 current:2 nt:1 refines:1 ronald:1 happen:1 designed:1 update:26 pursued:1 selected:1 leaf:1 fewer:1 intelligence:1 greedy:1 record:1 provides:1 recompute:1 consists:1 combine:2 manner:2 i... |
450 | 141 | 232
SPEECH PRODUCTION USING A NEURAL
NETWORK WITH A COOPERATIVE
LEARNING MECHANISM
Mitsuo Komura
Akio Tanaka
International Institute for Advanced Study of Social Information Science,
Fujitsu Limited
140 Miyamoto, Numazu-shi Shizuoka, 410-03 Japan
ABSTRACT
We propose a new neural network model and its learning
algorit... | 141 |@word uj:2 normalized:1 assigned:5 strategy:1 ll:1 solid:1 noted:2 separate:1 thank:1 m:4 generalized:1 selecting:1 rara:1 kitagawa:1 length:2 hin:2 s2x:1 activation:2 synthesizer:1 si:9 must:2 normal:1 exp:1 great:1 ratio:3 difficult:1 lg:1 yil:1 early:1 khz:1 purpose:1 analog:4 synthesized:3 unknown:1 item:1 upp... |
451 | 1,410 | A mathematical model of axon guidance by
diffusible factors
Geoffrey J. Goodhill
Georgetown Institute for Cognitive and Computational Sciences
Georgetown University Medical Center
3970 Reservoir Road
Washington DC 20007
geoff@giccs.georgetown.edu
Abstract
In the developing nervous system, gradients of target-derived d... | 1410 |@word seems:1 cm2:20 solid:1 surprising:1 must:3 readily:1 numerical:1 shape:1 asymptote:3 plot:1 cue:1 nervous:5 meakin:2 characterization:1 provides:1 math:1 contribute:1 firstly:2 ipi:3 mathematical:6 direct:1 epithelium:1 fitting:1 dan:1 expected:2 roughly:3 growing:2 brain:1 considering:1 increasing:1 become... |
452 | 1,411 | Instabilities in Eye Movement Control: A Model
of Periodic Alternating Nystagmus
ErnstR. Dow
Center for Biophysics and
Computational Biology,
Beckman Institute
University of Illinois at UrbanaChampaign,Urbana, IL 61801.
edow@uiuc.edu
Thomas J. Anastasio
Department of Molecular and Integrative Physiology, Center for B... | 1411 |@word cylindrical:1 bf:1 termination:1 integrative:1 simulation:2 r:3 eng:2 reduction:1 initial:5 cad:1 vor:31 interrupted:1 realistic:1 plasticity:1 half:1 gio:1 constructed:1 become:1 consists:1 pathway:11 acquired:1 expected:1 rapid:1 behavior:1 roughly:3 nor:1 uiuc:2 multi:1 brain:5 decreasing:1 prolonged:7 j... |
453 | 1,412 | Boltzmann Machine learning using mean
field theory and linear response correction
H.J. Kappen
Department of Biophysics
University of Nijmegen, Geert Grooteplein 21
NL 6525 EZ Nijmegen, The Netherlands
F. B. Rodriguez
Instituto de Ingenieria del Conocimiento & Departamento de Ingenieria Informatica.
Universidad Aut6nom... | 1412 |@word inversion:2 grooteplein:1 simulation:1 kappen:5 configuration:2 contains:1 paramagnetic:7 si:11 guez:3 numerical:2 partition:2 plot:2 stationary:1 intelligence:1 affair:1 steepest:1 simpler:1 mathematical:1 become:3 consists:4 introduce:1 manner:1 indeed:1 weightspace:1 multi:1 increasing:1 becomes:2 spain:... |
454 | 1,413 | Selecting weighting factors in logarithmic
opinion pools
Tom Heskes
Foundation for Neural Networks, University of Nijmegen
Geert Grooteplein 21, 6525 EZ Nijmegen, The Netherlands
tom@mbfys.kun.nl
Abstract
A simple linear averaging of the outputs of several networks as
e.g. in bagging [3], seems to follow naturally fr... | 1413 |@word version:1 middle:1 seems:2 replicate:1 logit:1 grooteplein:1 simulation:1 decomposition:6 jacob:1 minus:4 kappen:1 initial:1 selecting:8 ala:1 dx:3 written:2 remove:1 affair:1 supplying:1 lx:2 simpler:2 constructed:1 mbfys:1 relying:1 decomposed:1 becomes:1 estimating:1 notation:1 what:1 kind:2 interpreted:... |
455 | 1,414 | Visual Navigation in a Robot using
Zig-Zag Behavior
M. Anthony Lewis
Beckman Institute
405 N. Mathews Avenue
University of Illinois
Urbana, lllinois 61801
Abstract
We implement a model of obstacle avoidance in flying insects on a small,
monocular robot. The result is a system that is capable of rapid navigation
throug... | 1414 |@word briefly:1 open:1 seek:1 sensed:1 eng:1 solid:1 substitution:1 score:1 tuned:1 amp:1 current:1 activation:3 must:4 readily:1 john:1 vor:1 periodically:2 motor:9 designed:1 half:1 intelligence:1 plane:2 reciprocal:1 compo:1 detecting:2 centerline:1 zhang:3 along:1 differential:1 corridor:1 pathway:8 behaviora... |
456 | 1,415 | A Hippocampal Model of Recognition Memory
Randall C. O'Reilly
Department of Psychology
University of Colorado at Boulder
Campus Box 345
Boulder, CO 80309-0345
oreilly@psych.colorado.edu
Kenneth A. Norman
Department of Psychology
Harvard University
33 Kirkland Street
Cambridge, MA 02138
nonnan@wjh.harvard.edu
James L... | 1415 |@word version:2 stronger:2 hippocampus:18 proportion:1 integrative:1 simulation:2 initial:1 contains:1 series:1 selecting:1 efficacy:1 past:1 existing:2 current:2 comparing:1 contextual:1 activation:8 yet:1 conjunctive:4 must:1 update:1 v:1 cue:2 item:31 short:2 provides:2 parkin:2 sits:1 gillund:2 mathematical:2... |
457 | 1,416 | Correlates of Attention in a Model of
Dynamic Visual Recognition*
.
Rajesh P. N. Rao
Department of Computer Science
University of Rochester
Rochester, NY 14627
rao@cs.rochester.edu
Abstract
Given a set of objects in the visual field, how does the the visual system learn
to attend to a particular object of interest wh... | 1416 |@word middle:1 simulation:3 covariance:6 eng:1 attended:3 thereby:3 extrastriate:1 initial:4 cyclic:1 contains:1 series:1 interestingly:1 current:7 activation:2 si:3 john:2 mst:1 treating:1 designed:1 generative:6 intelligence:1 ith:5 location:2 ect:1 pathway:1 huber:1 mask:2 behavior:3 themselves:1 brain:1 occlu... |
458 | 1,417 | S-Map: A network with a simple
self-organization algorithm for generative
topographic mappings
Kimmo Kiviluoto
Laboratory of Computer and
Information Science
Helsinki University of Technology
P.O. Box 2200
FIN-02015 HUT, Espoo, Finland
Kimmo.KiviluotoChut.fi
Erkki Oja
Laboratory of Computer and
Information Science
He... | 1417 |@word h:1 version:3 briefly:1 jlf:1 stronger:2 seems:2 middle:2 selforganization:1 open:2 simulation:3 tried:1 ld:4 kappen:1 initial:3 configuration:3 series:1 t7:1 offering:1 activation:10 yet:1 written:2 plot:1 update:7 generative:4 ith:2 record:1 contribute:1 location:1 organising:1 mathematical:1 qualitative:... |
459 | 1,418 | An Incremental Nearest Neighbor
Algorithm with Queries
Joel Ratsaby?
N.A.P. Inc.
Hollis, New York
Abstract
We consider the general problem of learning multi-category classification from labeled examples. We present experimental results for
a nearest neighbor algorithm which actively selects samples from
different pat... | 1418 |@word cox:1 briefly:1 proportion:1 km:2 thereby:1 harder:1 contains:2 series:1 selecting:1 chervonenkis:1 pub:2 terion:1 current:5 si:2 written:1 partition:2 atlas:2 update:1 greedy:2 selected:2 jkj:1 hyperplanes:1 ik:2 calculable:1 consists:3 manner:1 expected:1 ra:1 frequently:1 multi:9 decomposed:1 increasing:... |
460 | 1,419 | Competitive On-Line Linear Regression
V. Vovk
pepartment of Computer Science
Royal Holloway, University of London
Egham, Surrey TW20 OEX, UK
vovkGdcs.rhbnc.ac.uk
Abstract
We apply a general algorithm for merging prediction strategies (the
Aggregating Algorithm) to the problem of linear regression with the
square loss... | 1419 |@word trial:11 version:3 polynomial:1 norm:7 stronger:1 covariance:2 contains:1 ours:1 comparing:1 cruz:1 additive:1 item:4 warmuth:8 short:2 manfred:1 node:1 ron:2 desantis:3 lor:4 unbounded:1 ucsc:1 symposium:1 consists:3 prove:1 paragraph:1 expected:1 wallace:1 bounded:5 notation:2 moreover:1 kind:1 developed:... |
461 | 142 | 99
Connectionist Learning of Expert Preferences by
Comparison Training
Gerald Tesauro
IBl\f Thomas.1. '''atson Rcsearc11 Centcr
PO Box 704, Yorktown Heights, NY 10598 USA
Abstract
A new training paradigm, caned the "eomparison pa.radigm," is introduced
for tasks in which a. network must learn to choose a prdcrred patt... | 142 |@word trial:1 judgement:1 inversion:1 seems:1 llsed:1 nd:15 cha:1 r:3 fonn:1 asks:1 tr:4 liu:1 loc:1 score:15 seriously:1 current:3 com:1 nt:5 cad:1 surprising:2 si:1 must:3 import:2 numerical:3 applica:2 half:4 selected:2 une:1 painstaking:2 alit:1 record:3 lr:1 rch:1 ire:2 preference:5 simpler:3 five:1 height:1 ... |
462 | 1,420 | How to Dynamically Merge Markov
Decision Processes
Satinder Singh
Department of Computer Science
University of Colorado
Boulder, CO 80309-0430
baveja@cs.colorado.edu
David Cohn
Adaptive Systems Group
Harlequin, Inc.
Menlo Park, CA 94025
cohn@harlequin.com
Abstract
We are frequently called upon to perform multiple ta... | 1420 |@word version:2 eliminating:1 decomposition:1 pick:2 tr:1 initial:10 selecting:1 existing:1 current:1 com:1 si:14 must:3 belmont:2 additive:1 update:6 v:1 stationary:1 greedy:4 fewer:1 guess:2 intelligence:1 reappears:1 ith:1 short:1 location:1 successive:1 zhang:2 become:1 combine:1 theoretically:2 ra:8 expected... |
463 | 1,421 | Bayesian model of surface perception
William T. Freeman
MERL, Mitsubishi Electric Res. Lab .
201 Broadway
Cambridge, MA 02139
Paul A. Viola
Artificial Intelligence Lab
Massachusetts Institute of Technology
Cambridge, MA 02139
freeman~erl.com
viola~ai.mit.edu
Abstract
Image intensity variations can result from seve... | 1421 |@word neurophysiology:1 illustrating:1 version:1 middle:1 judgement:6 instruction:1 seek:4 mitsubishi:1 shading:20 configuration:1 score:5 interestingly:1 com:1 yet:3 assigning:1 distant:1 visible:1 numerical:3 confirming:1 shape:42 designed:1 treating:1 intelligence:1 half:1 nervous:1 compo:1 filtered:1 provides... |
464 | 1,422 | An application of Reversible-J ump
MCMC to multivariate spherical Gaussian
mixtures
Alan D. Marrs
Signal & Information Processing Dept.
Defence Evaluation & Research Agency
Gt. Malvern, UK WR14 3PS
marrs@signal.dra.hmg.gb
Abstract
Applications of Gaussian mixture models occur frequently in the
fields of statistics and... | 1422 |@word determinant:1 version:1 proportion:1 barney:3 initial:1 series:1 current:2 comparing:2 z2:6 yet:1 must:1 conforming:1 subsequent:1 treating:1 stationary:1 ith:1 smith:2 firstly:1 along:2 beta:2 ik:2 combine:8 introduce:1 frequently:1 spherical:9 encouraging:2 becomes:2 estimating:1 what:1 titterington:1 tra... |
465 | 1,423 | Gradients for retinotectal mapping
Geoffrey J. Goodhill
Georgetown Institute for Cognitive and Computational Sciences
Georgetown University Medical Center
3970 Reservoir Road
Washington IX: 20007
geoff@giccs.georgetown.edu
Abstract
The initial activity-independent formation of a topographic map
in the retinotectal sys... | 1423 |@word version:1 hippocampus:2 proportionality:1 initial:3 reaction:1 current:1 anterior:2 yet:1 must:1 refines:1 plasticity:1 shape:14 opin:1 medial:1 cue:2 half:1 beginning:1 lr:1 location:1 zhang:1 mathematical:2 along:7 direct:2 supply:1 rostral:5 growing:1 brain:3 actual:1 considering:1 increasing:5 becomes:1... |
466 | 1,424 | Learning nonlinear overcomplete
representations for efficient coding
Michael S. Lewicki
Terrence J. Sejnowski
lewicki~salk.edu
terry~salk.edu
Howard Hughes Medical Institute
Computational Neurobiology Lab
The Salk Institute
10010 N. Torrey Pines Rd.
La Jolla, CA 92037
Abstract
We derive a learning algorithm for in... | 1424 |@word determinant:1 illustrating:1 norm:2 decomposition:1 simplifying:2 solid:1 phy:1 contains:1 pub:1 si:1 must:1 additive:2 remove:1 plot:2 fewer:1 provides:2 org:1 become:1 fitting:3 ra:1 ica:4 rapid:1 considering:1 solver:1 becomes:1 underlying:3 notation:2 matched:1 z:2 finding:3 transformation:1 exactly:1 s... |
467 | 1,425 | ?
Experiences with Bayesian Learning In
a
Real World Application
Peter Sykacek, Georg Dorffner
Austrian Research Institute for Artificial Intelligence
Schottengasse 3, A-10ID Vienna Austria
peter, georg@ai.univie.ac.at
Peter Rappelsberger
Institute for Neurophysiology at the University Vienna
Wahringer StraBe 17, A-lO... | 1425 |@word neurophysiology:1 seems:1 nd:1 tedious:1 tried:1 abou:1 contains:1 score:1 comparing:1 activation:10 must:1 distant:1 sponsored:1 discrimination:1 intelligence:2 selected:1 guess:1 inspection:1 short:1 detecting:1 coarse:1 node:1 toronto:1 simpler:1 five:1 stager:1 ray:3 concerted:1 rapid:1 embody:1 examine... |
468 | 1,426 | RCC Cannot Compute Certain FSA,
Even with Arbitrary Transfer Functions
Mark Ring
RWCP Theoretical Foundation GMD Laboratory
GMD - German National Research Center for Information Technology
Schloss Birlinghoven
D-53 754 Sankt Augustin, Germany
email: Mark .Ring@GMD.de
Abstract
Existing proofs demonstrating the computa... | 1426 |@word tr:1 contains:1 existing:2 current:5 x81:2 activation:1 must:4 designed:1 inspection:1 accepting:1 node:3 toronto:1 sigmoidal:6 along:1 consists:1 paragraph:1 manner:2 theoretically:1 expected:1 actual:1 becomes:2 provided:1 begin:1 notation:1 kind:1 unspecified:1 sankt:1 developed:1 temporal:1 every:3 osci... |
469 | 1,427 | Enhancing Q-Learning for
Optimal Asset Allocation
Ralph Neuneier
Siemens AG, Corporate Technology
D-81730 MUnchen, Germany
Ralph.Neuneier@mchp.siemens.de
Abstract
This paper enhances the Q-Iearning algorithm for optimal asset allocation proposed in (Neuneier, 1996 [6]). The new formulation simplifies
the approach by ... | 1427 |@word kong:1 achievable:1 seems:1 stronger:1 paid:1 profit:3 tr:1 initial:1 liquid:1 interestingly:1 past:1 neuneier:10 current:1 must:2 written:1 john:1 realistic:1 sponsored:1 fund:1 update:2 stationary:2 beginning:1 institution:1 preference:1 constructed:2 become:1 consists:2 combine:1 market:16 expected:6 beh... |
470 | 1,429 | Modelling Seasonality and Trends in Daily
Rainfall Data
Peter M Williams
School of Cognitive and Computing Sciences
University of Sussex
Falmer, Brighton BN1 9QH, UK.
email: peterw@cogs.susx.ac.uk
Abstract
This paper presents a new approach to the problem of modelling daily
rainfall using neural networks. We first mo... | 1429 |@word stronger:1 nd:1 confirms:1 recursively:1 initial:3 cyclic:2 series:6 selecting:1 past:3 current:1 legleye:2 incidence:1 activation:1 written:1 numerical:1 distant:1 shape:2 stationary:1 alone:1 indicative:1 parametrization:1 short:1 ifx:1 supplying:1 provides:1 location:1 successive:1 mathematical:1 supply:... |
471 | 143 | 240
TEMPORAL REPRESENTATIONS IN A
CONNECTIONIST SPEECH SYSTEM
Erich J. Smythe
207 Greenmanville Ave, #6
Mystic, CT 06355
ABSTRACT
SYREN is a connectionist model that uses temporal information
in a speech signal for syllable recognition. It classifies the rates
and directions of formant center transitions, and uses an... | 143 |@word merrill:1 briefly:2 middle:1 rising:1 eliminating:1 pulse:3 tr:2 ne1work:1 contains:1 series:1 tuned:1 past:3 existing:1 activation:33 yet:1 lang:1 must:5 john:1 motor:1 designed:5 succeeding:1 update:3 v:1 cue:1 nervous:2 beginning:1 indefinitely:1 sudden:1 provides:1 characterization:1 node:52 location:1 f... |
472 | 1,430 | Perturbative M-Sequences for Auditory
Systems Identification
Mark Kvale and Christoph E. Schreiner?
Sloan Center for Theoretical Neurobiology, Box 0444
University of California, San Francisco
513 Parnassus Ave, San Francisco, CA 94143
Abstract
In this paper we present a new method for studying auditory systems based ... | 1430 |@word mild:1 neurophysiology:2 trial:3 middle:1 polynomial:2 adrian:2 simulation:1 phy:2 series:3 tuned:2 ours:1 seriously:1 recovered:1 z2:1 perturbative:17 written:1 must:1 drop:1 plot:2 v:1 selected:1 iso:1 anaesthetised:1 sutter:1 detecting:1 contribute:1 along:1 c2:1 autocorrelation:3 manner:1 behavior:6 lit... |
473 | 1,431 | Learning to Order Things
William W. Cohen Robert E. Schapire Yoram Singer
AT&T Labs, 180 Park Ave., Florham Park, NJ 07932
{wcohen,schapire,singer} @research.att.com
Abstract
There are many applications in which it is desirable to order rather than classify
instances. Here we consider the problem of learning how to ord... | 1431 |@word repository:1 briefly:1 seems:1 pick:1 minus:1 reduction:1 initial:2 cyclic:2 series:1 att:1 score:2 document:20 ours:1 current:1 com:1 comparing:1 assigning:2 must:2 subsequent:1 numerical:3 progressively:1 update:1 greedy:8 warmuth:1 compo:1 institution:1 boosting:1 node:4 location:2 preference:48 wir:1 co... |
474 | 1,432 | The Storage Capacity
of a Fully-Connected Committee Machine
Yuansheng Xiong
Department of Physics, Pohang Institute of Science and Technology,
Hyoja San 31, Pohang , Kyongbuk, Korea
xiongOgalaxy.postech.ac.kr
Chulan Kwon
Department of Physics, Myong Ji University,
Yongin, Kyonggi, Korea
ckwonOwh.myongji.ac.kr
Jong-Hoo... | 1432 |@word version:1 briefly:1 seems:2 nd:5 r:8 lnh:1 orf:1 com:1 written:2 numerical:1 partition:1 j1:1 fund:1 guess:1 ith:1 vanishing:1 node:3 five:1 mathematical:2 along:1 introduce:1 angel:1 behavior:3 frequently:1 mechanic:4 multi:2 spherical:1 decomposed:1 increasing:1 mountain:1 interpreted:1 developed:1 zecchi... |
475 | 1,433 | A Generic Approach for Identification of
Event Related Brain Potentials via a
Competitive Neural Network Structure
Daniel H. Lange
Department of Electrical Engineering
Technion - liT
Haifa 32000
Israel
e-mail: lange@turbo.technion.ac.il
Hillel Pratt
Evoked Potential Laboratory
Technion - liT
Haifa 32000
Israel
e-mail:... | 1433 |@word trial:4 middle:1 eliminating:1 proportion:1 open:1 confirms:1 simulation:5 r:1 decomposition:1 eng:3 solid:1 series:2 daniel:1 com:1 john:1 fn:1 additive:5 alone:1 half:2 selected:1 dissertation:1 haykin:1 colored:1 math:1 node:5 five:2 become:1 specialize:1 consists:3 acquired:1 expected:1 frequently:1 ry:... |
476 | 1,434 | Hybrid reinforcement learning and its
application to biped robot control
Satoshi Yamada, Akira Watanabe, M:ichio Nakashima
{yamada, watanabe, naka}~bio.crl.melco.co.jp
Advanced Technology R&D Center
Mitsubishi Electric Corporation
Amagasaki, Hyogo 661-0001, Japan
Abstract
A learning system composed of linear control ... | 1434 |@word trial:16 consisted:1 rsj:1 objective:1 open:1 ankle:1 termination:2 mitsubishi:1 dependence:3 jacob:1 usual:6 during:1 self:1 eligibility:2 separate:1 exchange:4 qe:4 rhythm:1 reduction:1 initial:7 m:1 generalization:1 simulated:1 asme:2 sarsa:1 franklin:1 im:1 tn:1 motion:6 cp:3 gravitational:1 sloped:6 nt... |
477 | 1,435 | Incorporating Contextual Information in White
Blood Cell Identification
Xubo Song*
Department of Electrical Engineering
California Institute of Technology
Pasadena, CA 91125
xubosong@fire.work.caltech.edu
Yaser Abu-Mostafa
Dept. of Electrical Engineering
and Dept. of Computer Science
California Institute of Technology... | 1435 |@word proportion:2 nd:1 nicholson:1 pick:1 recursively:1 c1ass:1 contains:1 contextual:14 shape:2 pertinent:1 designed:2 amir:1 accordingly:1 mental:1 provides:1 five:2 dn:1 c2:11 direct:1 differential:4 manner:1 blast:11 rapid:1 nor:1 decomposed:1 automatically:1 company:2 pf:6 provided:1 maximizes:1 mass:1 deve... |
478 | 1,436 | Ensemble and Modular Approaches for
Face Detection: a Comparison
Raphael Feraud ?and Olivier Bernier t
France-Telecom CNET DTLjDLI
Technopole Anticipa, 2 avenue Pierre Marzin, 22307 Lannion cedex, FRANCE
Abstract
A new learning model based on autoassociative neural networks
is developped and applied to face detection... | 1436 |@word compression:1 bn:8 jacob:3 euclidian:1 solid:1 reduction:4 nowlan:1 realize:1 partition:1 remove:1 generative:2 pun:5 five:2 mahieux:1 constructed:1 ouput:1 prove:1 combine:2 detects:1 decomposed:1 window:5 considering:2 notation:1 moreover:1 bootstrapping:1 sung:3 every:1 collecting:1 fun:9 feraud:9 xd:2 c... |
479 | 1,437 | Wavelet Models for Video Time-Series
Sheng Ma and Chuanyi Ji
Department of Electrical, Computer, and Systems Engineering
Rensselaer Polytechnic Institute, Troy, NY 12180
e-mail: shengm@ecse.rpi.edu, chuanyi@ecse.rpi.edu
Abstract
In this work, we tackle the problem of time-series modeling of video
traffic. Different fr... | 1437 |@word version:1 loading:1 norm:1 simulation:2 bn:2 attainable:1 series:13 existing:2 comparing:1 com:1 rpi:2 srd:5 willinger:2 pertinent:2 plot:7 interpretable:1 alone:4 stationary:1 selected:1 vbr:3 short:8 regressive:1 provides:2 leland:1 autocorrelation:3 theoretically:2 expected:1 behavior:5 multi:3 cpu:1 med... |
480 | 1,438 | Learning Path Distributions using
Nonequilibrium Diffusion Networks
Paul Mineiro *
Javier Movellan
pmineiro~cogsci.ucsd.edu
movellan~cogsci.ucsd.edu
Department of Cognitive Science
University of California, San Diego
La Jolla, CA 92093-0515
Department of Cognitive Science
University of California, San Diego
La Jol... | 1438 |@word trial:3 open:1 simulation:1 dwh:1 solid:1 accommodate:1 initial:3 hereafter:1 denoting:2 interestingly:1 activation:2 dx:1 must:3 written:1 realistic:1 additive:1 numerical:1 shape:1 sdes:2 designed:1 isard:2 filtered:1 provides:1 math:1 node:4 sigmoidal:2 simpler:1 differential:2 prove:1 theoretically:1 ka... |
481 | 1,439 | Neural Basis of Object-Centered
Representations
Sophie Deneve and Alexandre Pouget
Georgetown Institute for Computational and Cognitive Sciences
Georgetown University
Washington, DC 20007-2197
sophie, alex@giccs.georgetown.edu
Abstract
We present a neural model that can perform eye movements to a
particular side of a... | 1439 |@word trial:2 cu:1 briefly:1 middle:1 instruction:3 simulation:2 invoking:1 contains:3 selecting:1 interestingly:1 must:1 tilted:1 shape:1 motor:5 plot:4 cue:6 intelligence:1 beginning:1 provides:1 location:6 preference:1 driver:6 consists:1 fixation:2 wallace:1 provided:3 retinotopic:2 mass:1 what:3 kind:1 monke... |
482 | 144 | 662
A PASSIVE SHARED ELEMENT ANALOG
ELECTRICAL COCHLEA
Joe Eisenberg
Bioeng. Group
U.C. Berkeley
David Feld
Dept. Elect. Eng.
207-30 Cory Hall
U.C. Berkeley
Berkeley, CA. 94720
Edwin Lewis
Dept Elect. Eng.
U.C. Berkeley
ABSTRACT
We present a simplified model of the micromechanics of the human
cochlea, realized with... | 144 |@word version:4 middle:2 rising:1 loading:1 replicate:1 simulation:2 pulse:3 eng:2 decomposition:3 pressure:2 reduction:1 series:2 ours:1 current:2 must:2 realize:3 evans:3 v:2 alone:1 selected:1 device:1 nervous:1 tone:6 beginning:1 indefinitely:1 provides:1 location:3 five:6 mathematical:1 along:8 constructed:2 ... |
483 | 1,440 | Two Approaches to Optimal Annealing
Todd K. Leen
Dept of Compo Sci. & Engineering
Oregon Graduate Institute of
Science and Technology
P.O.Box 91000, Portland,
Oregon 97291-1000
tleen@cse.ogi.edu
Bernhard Schottky and David Saad
Neural Computing Research Group
Dept of Compo Sci. & Appl. Math.
Aston University
Birmingh... | 1440 |@word achievable:1 confirms:1 invoking:1 minus:2 tr:1 accommodate:1 kappen:1 moment:4 initial:1 series:2 o2:1 activation:1 yet:2 perturbative:1 written:1 must:1 additive:1 numerical:2 enables:2 update:1 rjo:1 v:2 isotropic:2 compo:2 provides:2 math:2 node:5 cse:1 c2:1 differential:5 ik:1 consists:1 expected:6 als... |
484 | 1,441 | On the infeasibility of training neural
networks with small squared errors
Van H. Vu
Department of Mathematics, Yale University
vuha@math.yale.edu
Abstract
We demonstrate that the problem of training neural networks with
small (average) squared error is computationally intractable. Consider a data set of M points (Xi... | 1441 |@word polynomial:8 norm:12 open:3 mention:2 reduction:3 contains:5 chervonenkis:1 od:5 dx:2 bd:2 cruz:1 fn:5 half:2 warmuth:1 ith:1 emperical:2 completeness:1 math:1 node:13 sigmoidal:1 constructed:2 ucsc:1 prove:3 consists:1 polyhedral:1 hardness:2 behavior:1 roughly:2 inspired:1 freeman:1 decreasing:1 little:1 ... |
485 | 1,442 | Using Helmholtz Machines to analyze
multi-channel neuronal recordings
Virginia R. de Sa
desa@phy.ucsf.edu
R. Christopher deC harms
decharms@phy.ucsf.edu
Michael M. Merzenich
merz@phy.ucsf.edu
Sloan Center for Theoretical Neurobiology and
W. M. Keck Center for Integrative Neuroscience
University of California, San Fr... | 1442 |@word neurophysiology:2 trial:3 version:1 integrative:1 accounting:1 pick:1 schnitzer:3 reduction:6 phy:3 series:1 existing:1 current:1 activation:2 multineuron:1 treating:2 reproducible:2 stationary:1 greedy:2 generative:25 accordingly:1 detecting:1 simpler:1 constructed:2 consists:1 concerted:1 theoretically:1 ... |
486 | 1,443 | An Annealed Self-Organizing Map for Source
Channel Coding
Matthias Burger, Thore Graepel, and Klaus Obermayer
Department of Computer Science
Technical University of Berlin
FR 2-1, Franklinstr. 28/29, 10587 Berlin, Germany
{burger, graepel2, oby} @cs.tu-berlin.de
Abstract
We derive and analyse robust optimization sche... | 1443 |@word version:4 compression:3 open:2 covariance:3 solid:2 kappen:1 existing:1 recovered:1 com:1 analysed:1 assigning:1 numerical:2 hofmann:1 enables:1 drop:1 plot:2 v:1 ial:1 oblique:1 lr:2 quantizer:1 provides:1 codebook:10 successive:1 along:5 behavior:2 mechanic:1 lena:2 considering:1 increasing:1 becomes:1 bu... |
487 | 1,444 | Regression with Input-dependent Noise:
A Gaussian Process Treatment
Paul W. Goldberg
Department of Computer Science
University of Warwick
Coventry, CV 4 7AL, UK
pvgGdcs.varvick.ac.uk
Christopher K.I. Williams
Neural Computing Research Group
Aston University
Birmingham B4 7ET, UK
c.k.i.villiamsGaston.ac.uk
Christopher... | 1444 |@word inversion:1 simulation:1 covariance:10 solid:2 carry:2 current:1 com:1 written:1 plot:1 update:4 lky:1 xk:1 isotropic:2 ith:1 short:1 toronto:3 lx:2 along:2 vwv:1 fitting:2 introduce:1 manner:1 expected:1 wzd:1 becomes:1 matched:1 what:1 kind:1 every:2 ti:5 xd:1 uk:6 zl:2 control:1 unit:1 grant:1 gallinari:... |
488 | 1,445 | Structure Driven Image Database
Retrieval
Jeremy S. De Bonet &, Paul Viola
Artificial Intelligence Laboratory
Learning & Vision Group
545 Technology Square Massachusetts Institute of Technology
. Cambridge, MA 02139
EMAIL: jsdCOaLmit. edu & violaCOaLmit. edu
HOMEPAGE: http://www.ai . mit. edu/pro j ects/l v
Abstract
A... | 1445 |@word duda:2 seems:1 rgb:2 tr:1 outperforms:2 current:1 comparing:3 com:1 si:4 yet:1 must:2 tot:1 john:1 shape:2 depict:1 petkovic:1 intelligence:1 half:2 indicative:1 ith:1 provides:1 contribute:1 hsv:2 location:2 along:1 constructed:2 become:1 ect:1 symposium:1 consists:1 roughly:1 examine:1 multi:1 detects:1 d... |
489 | 1,446 | Factorizing Multivariate Function Classes
Juan K. Lin*
Department of Physics
University of Chicago
Chicago, IL 60637
Abstract
The mathematical framework for factorizing equivalence classes of
multivariate functions is formulated in this paper. Independent
component analysis is shown to be a special case of this decom... | 1446 |@word grier:1 open:1 hu:1 simulation:1 seek:1 decomposition:13 tmg:1 moment:1 necessity:1 bc:2 current:1 discretization:1 si:8 scatter:1 must:1 chicago:3 numerical:2 additive:1 subsequent:1 analytic:8 remove:1 pursued:1 leaf:1 indicative:1 plane:1 vanishing:1 gpx:1 provides:1 location:1 attack:1 height:1 mathemat... |
490 | 1,447 | An Improved Policy Iteratioll Algorithm
for Partially Observable MDPs
Eric A. Hansen
Computer Science Department
University of Massachusetts
Amherst, MA 01003
hansen@cs.umass.edu
Abstract
A new policy iteration algorithm for partially observable Markov
decision processes is presented that is simpler and more efficien... | 1447 |@word version:4 briefly:1 polynomial:1 termination:1 fifteen:1 profit:1 initial:3 contains:1 uma:1 outperforms:3 current:1 must:2 subsequent:1 shlomo:1 update:15 stationary:2 fewer:1 twostate:1 accordingly:1 provides:1 detecting:1 loworder:1 math:1 preference:7 simpler:1 zhang:2 polyhedral:1 theoretically:1 marke... |
491 | 1,449 | Adaptation in Speech Motor Control
John F. Houde*
UCSF Keck Center
Box 0732
San Francisco, CA 94143
Michael I. Jordan
MIT Dept. of Brain and Cognitive Sci.
EI0-034D
Cambridge, MA 02139
houde~phy.ucsf.edu
jordan~psyche.mit.edu
Abstract
Human subjects are known to adapt their motor behavior to a
shift of the visual ... | 1449 |@word mild:1 open:1 r:1 eng:1 solid:4 phy:1 xform:2 past:1 synthesizer:1 must:1 john:2 motor:7 remove:1 designed:1 plot:8 v:2 half:3 selected:1 device:1 dissertation:1 sits:1 five:4 warmup:1 along:2 constructed:1 behavior:1 examine:3 formants:10 brain:1 actual:1 moreover:2 notation:1 kaufman:1 spoken:1 acoust:3 t... |
492 | 145 | 459
A BIFURCATION THEORY APPROACH TO THE
PROGRAMMING OF PERIODIC A TTRACTORS IN
NETWORK MODELS OF OLFACTORY CORTEX
Bill Baird
Department of Biophysics
U.C. Berkeley
ABSTRACT
A new learning algorithm for the storage of static
and periodic attractors in biologically inspired
recurrent analog neural networks is introduc... | 145 |@word grey:2 rhesus:1 linearized:1 simulation:1 r:9 initial:1 contains:1 emn:2 imaginary:2 bsj:3 written:1 numerical:1 additive:3 analytic:1 v2s:3 alone:1 leaf:1 liapunov:1 xk:2 core:1 node:1 location:1 hyperplanes:1 mathematical:3 along:2 constructed:1 differential:1 hopf:2 behavioral:1 olfactory:7 expected:1 rou... |
493 | 1,450 | A Non-parametric Multi-Scale Statistical
Model for Natural Images
Jeremy S. De Bonet & Paul Viola
Artificial Intelligence Laboratory
Learning & Vision Group
545 Technology Square Massachusetts Institute of Technology
Cambridge, MA 02139
EMAIL: jsd@ai.mit.edu & viola@ai.mit.edu
HOMEPAGE: http://www.ai .mit. edu/project... | 1450 |@word version:1 compression:1 duda:2 pick:1 reduction:2 series:1 contains:3 ours:2 rightmost:2 current:2 si:1 yet:1 must:1 john:1 realistic:1 hypothesize:1 designed:1 interpretable:1 remove:1 resampling:1 intelligence:2 generative:5 provides:1 lx:2 constructed:1 direct:1 initiative:1 fmn:1 combine:1 expected:2 no... |
494 | 1,452 | Radial Basis Functions:
a Bayesian treatment
David Barber*
Bernhard Schottky
Neural Computing Research Group
Department of Applied Mathematics and Computer Science
Aston University, Birmingham B4 7ET, U.K.
http://www.ncrg.aston.ac.uk/
{D.Barber,B.Schottky}~aston.ac.uk
Abstract
Bayesian methods have been successfull... | 1452 |@word achievable:1 advantageous:1 grooteplein:1 seek:1 r:1 covariance:3 tr:3 solid:1 shot:1 carry:1 tuned:1 expositional:1 assigning:2 additive:3 analytic:4 plot:2 hts:1 update:1 alone:1 intelligence:1 leaf:1 normalising:1 provides:1 lx:1 simpler:2 mla:1 symposium:1 combine:1 fitting:1 manner:1 mbfys:1 examine:2 ... |
495 | 1,453 | Asymptotic Theory for Regularization:
One-Dimensional Linear Case
Petri Koistinen
Rolf Nevanlinna Institute, P.O. Box 4, FIN-00014 University of Helsinki,
Finland. Email: PetrLKoistinen@rnLhelsinkLfi
Abstract
The generalization ability of a neural network can sometimes be
improved dramatically by regularization. To an... | 1453 |@word polynomial:6 open:1 covariance:1 solid:1 moment:3 series:2 selecting:1 wj2:1 xiyi:2 written:1 john:2 v:1 cjx:1 mathematical:2 specialize:2 introduce:1 expected:2 little:1 provided:3 bounded:2 notation:2 easiest:1 what:1 interpreted:1 aln:2 textbook:1 titterington:2 ghosh:2 nonrandom:1 guarantee:1 exactly:1 ... |
496 | 1,454 | Graph Matching with Hierarchical
Discrete Relaxation
Richard C. Wilson and Edwin R. Hancock
Department of Computer Science, University of York
York, YOl 5DD, UK.
Abstract
Our aim in this paper is to develop a Bayesian framework for matching hierarchical relational models. The goal is to make discrete label assignments... | 1454 |@word grey:1 propagate:1 configuration:2 contains:1 exclusively:1 quadrilateral:1 current:2 comparing:2 si:3 must:2 shape:1 update:2 alone:1 selected:1 item:2 argm:1 provides:1 node:20 organising:1 firstly:1 along:2 descendant:3 consists:1 sustained:1 incorrect:1 manner:3 inter:9 multi:1 grade:1 freeman:1 resolve... |
497 | 1,455 | Asymptotic Theory for Regularization:
One-Dimensional Linear Case
Petri Koistinen
Rolf Nevanlinna Institute, P.O. Box 4, FIN-00014 University of Helsinki,
Finland. Email: PetrLKoistinen@rnLhelsinkLfi
Abstract
The generalization ability of a neural network can sometimes be
improved dramatically by regularization. To an... | 1455 |@word version:2 polynomial:6 loading:2 nd:1 open:1 orf:1 covariance:1 lobe:1 brightness:1 solid:1 moment:3 initial:3 configuration:2 series:2 contains:1 selecting:2 wj2:1 past:1 current:15 xiyi:2 written:2 must:1 john:2 realize:1 visible:2 plot:1 v:1 selected:1 imitate:1 plane:6 capitalizes:1 realizing:3 cjx:1 de... |
498 | 1,456 | On the Separation of Signals from
Neighboring Cells in Tetrode Recordings
Maneesh Sahani, John S. Pezaris and Richard A. Andersen
maneesh@caltech.edu, pz@caltech.edu, andersen@vis.caltech.edu
Computation and Neural Systems
California Institute of Technology
216-76 Caltech, Pasadena, CA 91125 USA
Abstract
We discuss a... | 1456 |@word version:1 briefly:1 eliminating:1 seems:1 nd:1 heuristically:1 rhesus:1 pulse:1 covariance:5 electronics:1 selecting:1 current:2 neurophys:1 nowlan:2 must:1 john:2 additive:2 visible:1 subsequent:2 shape:7 drop:1 implying:1 generative:3 ajd:1 selected:1 isotropic:1 short:1 compo:1 filtered:2 provides:1 node... |
499 | 1,457 | Prior Knowledge in Support Vector Kernels
Bernhard Scholkopf*t, Patrice Simard t , Alex Smola t, & Vladimir Vapnikt
* Max-Planck-Institut fur biologische Kybernetik, Tiibingen, Gennany
t GMD FIRST, Rudower Chaussee 5, 12489 Berlin, Gennany
t AT&T Research, 100 Schulz Drive, Red Bank, NJ, USA
bS@first.gmd.de
Abstract
... | 1457 |@word polynomial:9 proportion:1 retraining:1 open:1 d2:5 covariance:4 uon:1 carry:1 reduction:2 initial:1 series:1 zij:2 chervonenkis:2 bc:1 diagonalized:1 assigning:1 written:2 oldenbourg:1 seelen:1 qiyi:2 drop:2 aside:1 leaf:1 selected:1 kyk:1 xk:1 beginning:1 short:1 provides:1 mulier:1 postal:1 location:1 hyp... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.