Unnamed: 0 int64 0 7.24k | id int64 1 7.28k | raw_text stringlengths 9 124k | vw_text stringlengths 12 15k |
|---|---|---|---|
900 | 1,824 | Four-Iegged Walking Gait Control Using a
Neuromorphic Chip Interfaced to a Support
Vector Learning Algorithm
Susanne Still
NEC Research Institute
4 Independence Way, Princeton NJ 08540, USA
sasa@research.nj.nec.com
Klaus Hepp
Institute of Theoretical Physics
ETH Zurich, Switzerland
Bernhard Scholkopf
Microsoft Resear... | 1824 |@word version:1 adrian:1 d2:2 locomotive:1 tr:1 solid:2 reduction:1 configuration:2 contains:3 liu:1 past:1 current:2 com:2 kondo:1 must:1 john:2 physiol:1 shape:1 enables:2 motor:10 plot:4 nervous:1 plane:1 smith:1 node:11 cpg:1 five:1 direct:1 scholkopf:2 consists:2 fitting:1 falsely:1 acquired:1 theoretically:... |
901 | 1,825 | Learning curves for Gaussian processes
regression: A framework for good
approximations
Dorthe Malzahn
Manfred Opper
Neural Computing Research Group
School of Engineering and Applied Science
Aston University, Birmingham B4 7ET, United Kingdom.
[malzahnd.opperm]~aston.ac.uk
Abstract
Based on a statistical mechanics app... | 1825 |@word version:1 polynomial:1 seems:3 simulation:3 bn:2 covariance:4 solid:2 phy:1 series:1 united:1 pub:1 yet:1 dx:1 subsequent:1 partition:7 xk:2 manfred:1 simpler:2 become:1 expected:1 ra:1 mechanic:5 decreasing:2 becomes:2 estimating:2 panel:4 developed:1 extremum:1 wrong:2 rm:1 uk:2 grant:1 yn:1 positive:1 en... |
902 | 1,826 | Algebraic Information Geometry for
Learning Machines with Singularities
Sumio Watanabe
Precision and Intelligence Laboratory
Tokyo Institute of Technology
4259 Nagatsuta, Midori-ku, Yokohama, 226-8503 J apan
swatanab@pi.titech.ac.jp
Abstract
Algebraic geometry is essential to learning theory. In hierarchical
learning... | 1826 |@word version:1 proportion:1 nd:1 open:2 xlw:3 electronics:1 itp:1 nt:1 rpi:3 si:1 dx:4 analytic:12 enables:1 midori:1 intelligence:1 fewer:1 selected:1 half:2 plane:1 math:3 clarified:3 ron:3 firstly:2 mathematical:1 direct:3 differential:1 prove:4 introduce:2 behavior:1 examine:1 nor:1 decomposed:1 decreasing:1... |
903 | 1,827 | A productive, systematic framework for
the representation of visual structure
Shimon Edelman
232 Uris Hall, Dept. of Psychology
Cornell University
Ithaca, NY 14853-7601
Nathan Intrator
Institute for Brain and Neural Systems
Box 1843, Brown University
Providence, RI 02912
se37@cornell.edu
N athan_Intrator@brown. edu... | 1827 |@word neurophysiology:1 version:2 middle:1 open:1 covariance:1 wisniewski:1 carry:1 moment:2 necessity:1 configuration:3 series:1 fragment:25 selecting:1 tuned:6 past:1 current:1 yet:1 must:1 stemming:1 shape:19 plot:1 alone:2 half:4 record:1 coarse:3 location:17 five:4 along:2 constructed:1 differential:1 edelma... |
904 | 1,828 | Spike-Timing-Dependent Learning for
Oscillatory Networks
Silvia Scarp etta
Dept. of Physics "E.R. Caianiello"
Salerno University 84081 (SA) Italy
and INFM, Sezione di Salerno Italy
scarpetta@na. infn. it
Zhaoping Li
Gatsby Compo Neurosci. Unit
University College, London, WCIN 3AR
United Kingdom
zhaoping@gatsby.ucl.ac... | 1828 |@word eliminating:1 hippocampus:3 seems:1 simulation:9 linearized:3 solid:1 vigorously:1 initial:1 efficacy:3 united:1 tuned:3 current:3 activation:2 written:1 john:1 numerical:1 distant:1 subsequent:1 plasticity:1 shape:1 gv:2 stationary:2 half:1 implying:1 plane:4 short:1 compo:1 wth:1 location:2 sigmoidal:1 co... |
905 | 1,829 | Learning winner-take-all competition between
groups of neurons in lateral inhibitory networks
Xiaohui Xie, Richard Hahnloser and H. Sebastian Seung
E25-21O, MIT, Cambridge, MA 02139
{xhxielrhlseung}@mit.edu
Abstract
It has long been known that lateral inhibition in neural networks can lead
to a winner-take-all compet... | 1829 |@word open:1 bn:1 initial:7 contains:5 bc:1 past:1 coactive:2 activation:1 dx:1 written:1 must:10 update:1 ith:1 unbounded:1 become:1 retrieving:1 qualitative:1 prove:3 behavior:1 inspired:2 globally:2 provided:4 discover:1 bounded:1 moreover:1 panel:5 circuit:1 interpreted:2 affirmative:1 guarantee:2 every:7 dem... |
906 | 183 | 348
Further Explorations in Visually-Guided
Reaching: Making MURPHY Smarter
Bartlett W. Mel
Center for Complex Systems Research
Beckman Institute, University of illinois
405 North Matheus Street
Urbana, IL 61801
ABSTRACT
MURPHY is a vision-based kinematic controller and path planner
based on a connectionist architect... | 183 |@word trial:2 briefly:1 joh:1 manageable:1 heuristically:1 tried:1 mammal:1 carry:2 phy:1 configuration:7 series:1 extrapersonal:1 contains:1 initial:1 tuned:2 current:2 activation:2 yet:1 must:1 grain:1 subsequent:2 shape:1 motor:10 designed:1 pursued:1 fewer:1 guess:1 plane:1 core:1 mental:11 coarse:1 quantized:... |
907 | 1,830 | Learning Segmentation by Random Walks
Marina Meila
University of Washington
Jianbo Shi
Carnegie Mellon University
mmp~stat.washington.edu
jshi~cs.cmu.edu
Abstract
We present a new view of image segmentation by pairwise similarities. We interpret the similarities as edge flows in a Markov
random walk and study the ... | 1830 |@word mild:1 version:1 open:1 adrian:1 current:1 dx:1 must:1 partition:4 shape:7 stationary:2 cue:4 intelligence:1 slh:3 gure:1 provides:6 node:4 lx:1 along:1 constructed:2 symposium:1 prove:1 consists:2 pairwise:5 roughly:1 examine:1 freeman:2 little:1 pf:3 lll:1 becomes:1 provided:1 underlying:1 moreover:1 line... |
908 | 1,831 | Balancing Multiple Sources of Reward in
Reinforcement Learning
Christian R. Shelton
Artificial Intelligence Lab
Massachusetts Institute of Technology
Cambridge, MA 02139
cshelton@ai.mit.edu
Abstract
For many problems which would be natural for reinforcement learning,
the reward signal is not a single scalar value but... | 1831 |@word seems:1 hu:1 r:1 pick:2 profit:1 series:1 must:3 christian:1 remove:1 designed:2 interpretable:1 aps:1 plot:2 v:15 sponsored:1 intelligence:3 leaf:1 weighing:1 fund:1 parameterization:1 record:1 provides:1 contribute:1 honda:1 preference:10 simpler:1 along:1 incorrect:1 manner:1 expected:1 behavior:2 nor:2 ... |
909 | 1,832 | Generalized Belief Propagation
Jonathan S. Yedidia
MERL
201 Broadway
Cambridge, MA 02139
Phone: 617-621-7544
William T. Freeman
MERL
201 Broadway
Cambridge, MA 02139
Phone: 617-621-7527
Yair Weiss
Computer Science Division
UC Berkeley, 485 Soda Hall
Berkeley, CA 94720-1776
Phone: 510-642-5029
yedidia@merl.com
free... | 1832 |@word exploitation:1 version:2 open:1 r:8 propagate:1 simulation:1 accounting:1 dramatic:1 minus:2 com:2 bd:2 must:2 written:2 reminiscent:1 designed:1 update:15 v:1 stationary:3 half:1 instantiate:1 yr:1 fewer:1 node:44 constructed:1 direct:5 qualitative:1 prove:2 introduce:3 pairwise:1 expected:1 freeman:2 litt... |
910 | 1,833 | Regularized Winnow Methods
Tong Zhang
Mathematical Sciences Department
IBM TJ. Watson Research Center
Yorktown Heights, NY 10598
tzhang@watson.ibm.com
Abstract
In theory, the Winnow multiplicative update has certain advantages over
the Perceptron additive update when there are many irrelevant attributes.
Recently, t... | 1833 |@word version:11 norm:5 seek:1 ld:2 initial:5 err:2 com:1 comparing:1 xiyi:7 john:1 numerical:1 additive:3 remove:1 update:28 discrimination:4 warmuth:1 mln:1 detecting:1 provides:1 hyperplanes:2 zhang:2 lor:1 height:1 mathematical:1 five:1 c2:1 direct:1 rc:1 scholkopf:1 introduce:3 ilxill:2 expected:2 behavior:1... |
911 | 1,834 | .
A new model of spatial representations In
multimodal brain areas.
Sophie Deneve
Department of Brain and cognitive Science
University of Rochester
Rochester, NY 14620.
sdeneve@bcs.rochester.edu
Jean-Rene Duhamel
Institut des Sciences Cognitives
C.N.R.S
Bron, France 69675
jrd@isc.cnrs?fr
Alexandre Pouget
Department... | 1834 |@word neurophysiology:1 version:2 middle:1 seems:1 bf:1 extinction:4 grey:1 initial:1 tuned:4 must:1 readily:1 visible:1 realistic:1 subsequent:1 motor:17 cue:8 ith:1 fogassi:1 detecting:1 location:3 fixation:1 combine:4 behavior:1 themselves:1 nor:2 frequently:1 multi:7 brain:9 automatically:1 provided:3 retinot... |
912 | 1,835 | Keeping flexible active contours on track using
Metropolis updates
Trausti T. Kristjansson
University of Waterloo
tt kr i s tj @uwa te r l oo . ca
Brendan J. Frey
University of Waterloo
f r ey@uwate r l oo . ca
Abstract
Condensation, a form of likelihood-weighted particle filtering, has been
successfully used to inf... | 1835 |@word cox:1 compression:1 polynomial:2 kristjansson:1 dramatic:1 tr:1 initial:2 outperforms:1 current:2 surprising:1 written:1 kleen:2 shape:12 update:16 resampling:1 isard:3 fewer:2 toronto:1 successive:2 along:3 rnl:1 consists:1 fitting:4 presumed:1 roughly:1 behavior:1 examine:1 freeman:3 actual:2 little:4 pro... |
913 | 1,836 | Partially Observable SDE Models for
Image Sequence Recognition Tasks
Javier R. Movellan
Institute for Neural Computation
University of California San Diego
Paul Mineiro
Department of Cognitive Science
University of California San Diego
R. J. Williams
Department of Mathematics
University of California San Diego
Abstr... | 1836 |@word advantageous:1 calculus:1 simulation:2 tried:2 versatile:1 initial:2 tuned:1 current:2 activation:2 dx:1 reminiscent:1 realistic:2 shape:9 sdes:5 generative:1 node:7 rc:2 along:1 differential:5 become:1 consists:2 combine:1 manner:2 behavior:1 themselves:1 discretized:1 karatzas:2 inspired:1 encouraging:3 a... |
914 | 1,837 | Hierarchical Memory-Based
Reinforcement Learning
Natalia Hernandez-Gardio}
Artificial Intelligence Lab
Massachusetts Institute of Technology
Cambridge, MA 02139
nhg@ai.mit.edu
Sridhar Mahadevan
Department of Computer Science
Michigan State University
East Lansing, MI 48824
mahadeva@cse.msu.edu
Abstract
A key challeng... | 1837 |@word trial:1 version:1 middle:1 nd:1 open:1 termination:1 d2:1 propagate:1 decomposition:1 carry:1 contains:1 past:12 outperforms:1 current:7 comparing:2 nt:1 si:3 must:11 grain:1 realistic:1 informative:3 enables:1 update:5 smdp:3 v:1 intelligence:1 discovering:1 greedy:4 leaf:1 mccallum:2 hallway:5 short:15 re... |
915 | 1,838 | Feature Correspondence:
A Markov Chain Monte Carlo Approach
Frank Dellaert, Steven M. Seitz, Sebastian Thrun, and Charles Thorpe
Department of Computer Science &Robotics Institute
Carnegie Mellon University
Pittsburgh, PA 15213
{dellaert,seitz,thrun,cet }@cs.cmu.edu
Abstract
When trying to recover 3D structure from a... | 1838 |@word version:1 seitz:3 simplifying:1 initial:3 existing:3 current:1 recovered:2 written:1 realistic:1 shape:7 analytic:1 alone:1 intelligence:1 fewer:1 selected:1 isotropic:1 es:1 iterates:1 provides:2 along:2 direct:1 symposium:1 ik:1 ostland:1 introduce:1 theoretically:1 inter:1 expected:2 themselves:1 decreas... |
916 | 1,839 | A Neural Probabilistic Language Model
Yoshua Bengio; Rejean Ducharme and Pascal Vincent
Departement d'Informatique et Recherche Operationnelle
Centre de Recherche Mathematiques
Universite de Montreal
Montreal, Quebec, Canada, H3C 317
{bengioy,ducharme, vincentp }@iro.umontreal.ca
Abstract
A goal of statistical langua... | 1839 |@word version:1 bigram:3 compression:2 norm:1 seems:1 tried:2 decomposition:1 thereby:1 initial:1 score:3 tuned:1 document:1 denoting:1 current:1 must:1 partition:1 hash:2 v:1 mccallum:1 short:12 farther:1 recherche:2 granting:1 redone:1 lexicon:1 successive:1 tagger:1 along:1 direct:9 fitting:1 combine:1 paragra... |
917 | 184 | 169
DOES THE NEURON "LEARN" LIKE THE SYNAPSE?
RAOUL TAWEL
Jet Propulsion Laboratory
California Institute of Technology
Pasadena, CA 91109
Abstract. An improved learning paradigm that offers a significant reduction in computation time during the supervised learning phase is described. It is based on
extending the role ... | 184 |@word gradual:1 simulation:4 simplifying:1 tr:1 solid:2 reduction:3 initial:2 terminus:1 current:2 activation:8 written:1 j1:1 treating:1 drop:2 update:4 selected:4 device:1 fewer:1 ith:2 steepest:1 sigmoidal:2 mathematical:2 along:1 become:1 ouput:1 prove:1 pairwise:2 frequently:1 begin:2 project:1 funtion:1 unsp... |
918 | 1,840 | APRICODD: Approximate Policy Construction
using Decision Diagrams
Robert St-Aubin
Jesse Hoey
Craig Boutilier
Dept. of Computer Science
University of British Columbia
Vancouver, BC V6T lZA
Dept. of Computer Science
University of British Columbia
Vancouver, BC V6T lZA
Dept. of Computer Science
University of Toronto... | 1840 |@word cu:1 version:2 hu:2 confirms:1 tr:1 reduction:4 initial:2 series:1 pub:1 denoting:1 bc:2 unprimed:1 comparing:1 nt:1 must:2 gaona:1 randal:1 stationary:1 intelligence:2 discovering:1 leaf:9 selected:1 fewer:1 short:1 provides:1 node:7 toronto:3 direct:1 become:1 overhead:1 inter:1 abelardo:1 ra:1 planning:5... |
919 | 1,841 | Robust Reinforcement Learning
J un Morimoto
Graduate School of Information Science
Nara Institute of Science and Technology;
Kawato Dynamic Brain Project, JST
2-2 Hikaridai Seika-cho Soraku-gun
Kyoto 619-0288 JAPAN
xmorimo@erato.atr.co.jp
Kenji Doya
ATR International;
CREST, JST
2-2 Hikaridai Seika-cho Soraku-gun
Kyo... | 1841 |@word trial:2 briefly:1 norm:4 simulation:4 linearized:2 accommodate:1 initial:1 minmax:3 j1:1 shape:1 analytic:2 designed:2 update:1 fewer:1 mgl:1 glover:1 differential:2 introduce:2 acquired:2 behavior:2 seika:2 planning:2 brain:1 torque:1 td:1 actual:1 unpredictable:2 considering:1 project:1 estimating:2 bound... |
920 | 1,842 | Hippocampally-Dependent Consolidation in a
Hierarchical Model of Neocortex
Szabolcs Ka1i 1 ,2
Peter Dayan 1
1 Gatsby
Computational Neuroscience Unit
University College London
17 Queen Square, London, England, WCIN 3AR.
2Department of Brain and Cognitive Sciences
Massachusetts Institute of Technology
Cambridge, MA 02... | 1842 |@word implemented:1 establish:1 involves:2 come:1 indicate:1 hippocampus:9 damage:2 semantic:2 lobe:1 exploration:1 human:2 eg:1 adjacent:1 whereby:1 shot:1 require:1 entity:1 rat:1 hippocampal:2 efficacy:1 considers:1 complete:3 consensus:1 declarative:4 invisible:1 helping:1 code:1 relationship:1 activation:1 b... |
921 | 1,843 | From Mixtures of Mixtures to
Adaptive Transform Coding
Cynthia Archer and Todd K. Leen
Department of Computer Science and Engineering
Oregon Graduate Institute of Science & Technology
20000 N.W. Walker Rd, Beaverton, OR 97006-1000
E-mail: archer, tleen@cse.ogi.edu
Abstract
We establish a principled framework for adap... | 1843 |@word trial:4 mri:4 version:2 compression:11 advantageous:1 replicate:2 solid:1 reduction:8 substitution:2 envision:2 current:1 nowlan:2 dct:2 cottrell:1 partition:7 designed:7 concert:1 drop:1 greedy:1 half:2 haykin:5 quantizer:11 quantized:1 cse:1 provides:3 codebook:1 constructed:1 consists:2 fitting:6 ra:7 co... |
922 | 1,844 | A PAC-Bayesian Margin Bound for Linear
Classifiers: Why SVMs work
Ralf Herbrich
Statistics Research Group
Computer Science Department
Technical University of Berlin
ralfh@cs.tu-berlin.de
Thore Graepel
Statistics Research Group
Computer Science Department
Technical University of Berlin
guru@cs.tu-berlin.de
Abstract
W... | 1844 |@word repository:1 version:5 pw:5 polynomial:1 norm:1 tr:2 solid:1 series:1 chervonenkis:2 must:1 john:2 numerical:1 shawetaylor:1 plot:1 xex:1 half:1 maximised:1 vanishing:1 provides:2 boosting:2 coarse:1 draft:1 herbrich:3 scholkopf:1 eleventh:1 introduce:1 theoretically:1 ilxill:1 rapid:1 considering:2 provide... |
923 | 1,845 | Structure learning in human causal induction
Joshua B. Tenenbaum & Thomas L. Griffiths
Department of Psychology
Stanford University, Stanford, CA 94305
{jbt,gruffydd}@psych.stanfo rd .edu
Abstract
We use graphical models to explore the question of how people learn simple causal relationships from data. The two leadin... | 1845 |@word trial:7 judgement:1 stronger:1 seems:1 concise:1 outperforms:1 existing:2 must:1 written:1 numerical:2 designed:6 parameterization:2 ith:1 mental:1 provides:1 parameterizations:1 contribute:1 node:3 along:1 direct:1 introductory:1 behavioral:2 behavior:2 examine:2 inspired:1 decreasing:2 eil:1 becomes:2 pro... |
924 | 1,846 | The Missing Link - A Probabilistic Model of
Document Content and Hypertext Connectivity
David Cohn
Burning Glass Technologies
201 South Craig St, Suite 2W
Pittsburgh, PA 15213
david. cohn @burning-glass.com
Thomas Hofmann
Department of Computer Science
Brown University
Providence, RI 02192
th@cs.brown.edu
Abstract
W... | 1846 |@word trial:1 faculty:3 version:1 proportion:3 stronger:1 plsa:16 decomposition:9 ld:2 contains:2 score:1 selecting:1 document:83 existing:1 com:1 z2:2 crawling:4 must:1 unchanging:1 hofmann:2 treating:1 half:2 tenn:1 selected:2 greedy:2 mccallum:1 pointer:1 provides:2 authority:5 successive:2 along:1 symposium:1... |
925 | 1,847 | Stability and noise in biochemical switches
William Bialek
NEC Research Instit ute
4 Independence Way
Princeton, New Jersey 08540
bialek@research. nj. nec. com
Abstract
Many processes in biology, from the regulation of gene expression in
bacteria to memory in the brain, involve switches constructed from
networks of b... | 1847 |@word version:1 seems:3 nd:4 heuristically:1 simulation:3 pulse:1 carry:1 phosphorylation:1 electronics:1 liu:1 necessity:1 optically:2 genetic:5 reaction:20 current:1 com:1 comparing:1 activation:4 yet:1 dx:1 must:2 written:3 deniz:1 plasticity:2 analytic:1 fewer:1 disassembly:1 coleman:1 provides:1 complication... |
926 | 1,848 | .N-Body. Problems in Statistical Learning
Alexander G. Gray
Department of Computer Science
Carnegie Mellon University
agray@cs.cmu.edu
Andrew W. Moore
Robotics Inst. and Dept. Compo Sci.
Carnegie Mellon University
awm@cs.cmu.edu
Abstract
We present efficient algorithms for all-point-pairs problems , or 'Nbody '-like... | 1848 |@word rightchild:6 briefly:1 version:2 middle:1 nd:2 disk:2 twelfth:2 open:1 simulation:2 r:2 covariance:3 dramatic:1 harder:2 recursively:1 disappointingly:1 contains:1 loc:1 ours:1 existing:1 current:3 arest:1 etwork:1 intelligence:2 leaf:6 unacceptably:1 short:1 record:3 compo:1 pointer:1 coarse:1 characteriza... |
927 | 1,849 | Kernel-Based Reinforcement Learning in
Average-Cost Problems: An Application
to Optimal Portfolio Choice
Dirk Ormoneit
Department of Computer Science
Stanford University
Stanford, CA 94305-9010
ormoneit@cs.stanford.edu
Peter Glynn
EESOR
Stanford University
Stanford, CA 94305-4023
Abstract
Many approaches to reinforc... | 1849 |@word middle:1 km:8 simulation:1 thereby:2 recursively:1 initial:1 series:3 ours:1 past:1 existing:1 current:1 written:2 readily:1 realistic:2 additive:1 numerical:2 j1:6 update:4 stationary:1 iterates:1 provides:1 location:2 preference:1 direct:1 differential:3 prove:1 combine:1 ica:1 market:3 frequently:1 ol:1 ... |
928 | 185 | 215
Consonant Recognition by Modular Construction of
Large Phonemic Time-Delay Neural Networks
Alex Waibel
Carnegie-Mellon University
Pittsburgh, PA 15213,
ATR Interpreting Telephony Research Laboratories
Osaka, Japan
Abstract
In this paperl we show that neural networks for speech recognition can be constructed in
a ... | 185 |@word exploitation:1 middle:1 version:1 proportion:1 retraining:2 glue:8 alliant:1 tr:2 initial:1 configuration:1 contains:1 score:3 liquid:1 existing:1 contextual:1 activation:1 yet:1 lang:2 must:2 subsequent:1 subcomponent:5 pertinent:1 wanted:1 discrimination:13 alone:1 half:1 selected:1 v:1 coarse:4 provides:1... |
929 | 1,850 | Feature Selection for SVMs
J. Weston t, S. Mukherjee tt , O. Chapelle*, M. Pontil tt
T. Poggiott, V. Vapnik*,ttt
t Barnhill Biolnformatics.com, Savannah, Georgia, USA.
tt CBCL MIT, Cambridge, Massachusetts, USA.
* AT&T Research Laboratories, Red Bank, USA.
ttt Royal Holloway, University of London, Egham, Surrey, UK.
A... | 1850 |@word trial:2 eliminating:1 polynomial:1 norm:1 smirnov:4 tamayo:2 tried:1 gish:1 myeloid:1 solid:2 wrapper:7 score:10 existing:1 err:1 bradley:1 com:1 must:3 john:1 realize:1 remove:1 discrimination:3 selected:1 yr:1 provides:1 downing:1 consists:1 introduce:4 indeed:1 expected:2 morphology:2 considering:1 minim... |
930 | 1,851 | Dendritic compartmentalization could underlie
competition and attentional biasing of
simultaneous visual stimuli
Kevin A. Archie
Neuroscience Program
University of Southern California
Los Angeles, CA 90089-2520
Bartlett W. Mel
Department of Biomedical Engineering
University of Southern California
Los Angeles, CA 9008... | 1851 |@word simplecell:1 open:1 cm2:4 integrative:1 simulation:3 attended:5 carry:1 extrastriate:4 initial:1 efficacy:1 disparity:1 mainen:2 reynolds:5 happen:1 progressively:1 v:1 alone:8 half:1 selected:2 postnatal:1 record:1 location:3 preference:1 mathematical:1 along:1 direct:1 roughly:1 brain:1 spherical:1 chap:2... |
931 | 1,852 | Explaining Away in Weight Space
Peter Dayan
Sham Kakade
Gatsby Computational Neuroscience Unit, UCL
17 Queen Square London WCIN 3AR
da y a n @ga t sb y.u c l. ac . uk
sham@ga t sby.u c l. ac .uk
Abstract
Explaining away has mostly been considered in terms of inference of
states in belief networks. We show how it ca... | 1852 |@word trial:23 judgement:1 stronger:1 d2:1 seek:1 crucially:1 bn:6 covariance:24 accounting:1 heteroassociative:1 r:1 solid:3 initial:3 l__:1 current:1 com:1 activation:2 yet:1 must:1 additive:1 update:6 sby:1 lx:1 become:1 manner:1 notably:1 ra:1 expected:1 behavior:2 frequently:1 growing:2 nor:1 window:1 provid... |
932 | 1,853 | Automatic choice of dimensionality for peA
Thomas P. Minka
MIT Media Lab
20 Ames St, Cambridge, MA 02139
tpminka@media.mit.edu
Abstract
A central issue in principal component analysis (PCA) is choosing the
number of principal components to be retained. By interpreting PCA as
density estimation, we show how to use Bay... | 1853 |@word determinant:1 nd:2 d2:3 r:1 simulation:1 covariance:5 decomposition:2 rayner:1 pick:2 tr:8 phy:1 contains:3 exclusively:2 score:1 zij:1 series:1 fragment:4 selecting:1 pub:2 recovered:1 ka:1 com:1 yet:1 must:2 numerical:2 informative:1 j1:1 noninformative:2 drop:1 generative:1 leaf:1 intelligence:1 paramete... |
933 | 1,854 | Algorithmic Stability and Generalization
Performance
Olivier Bousquet
CMAP
Ecole Polytechnique
F-91128 Palaiseau cedex
FRANCE
bousquet@cmapx.polytechnique?fr
Andre Elisseeff'"
Barnhill Technologies
6709 Waters Avenue
Savannah, GA 31406
USA
andre@barnhilltechnologies.com
Abstract
We present a novel way of obtaining PA... | 1854 |@word briefly:1 norm:1 tried:1 elisseeff:2 pick:1 initial:1 ecole:1 rkhs:2 com:1 si:5 written:2 girosi:2 intelligence:1 math:1 readability:1 ron:2 mcdiarmid:3 prove:2 consists:1 introduce:1 deteriorate:1 indeed:1 eurocolt:1 lyon:2 little:1 begin:1 notation:4 bounded:2 moreover:1 null:1 what:3 kind:1 minimizes:1 u... |
934 | 1,855 | Fast Training of Support Vector Classifiers
F. Perez-Cruzt, P. L. Alarc6n-Dianat, A. Navia-Vazquez:j:and A. Artes-Rodriguez:j:.
tDpto. Teoria de la Seiial y Com., Escuela Politecnica, Universidad de Alcala.
28871-Alcala de Henares (Madrid) Spain. e-mail: fernando@tsc.uc3m.es
:j:Dpto. Tecnologias de las comunicaciones,... | 1855 |@word trial:2 version:1 seems:1 decomposition:1 eng:1 reduction:1 contains:1 rkhs:4 com:1 si:6 must:4 readily:1 john:1 cruz:2 numerical:1 girosi:2 sampl:1 selected:1 short:2 haykin:1 simpler:1 scholkopf:2 inside:1 cpu:6 inappropriate:1 solver:2 considering:1 cardinality:2 spain:3 cache:1 moreover:2 lowest:1 what:... |
935 | 1,856 | Gaussianization
Scott Shaobing Chen
Renaissance Technologies
East Setauket, NY 11733
schen@rentec.com
Ramesh A. Gopinath
IBM TJ. Watson Research Center
Yorktown Heights, NY 10598
rameshg@us.ibm.com
Abstract
High dimensional data modeling is difficult mainly because the so-called
"curse of dimensionality". We propose... | 1856 |@word d2:1 covariance:4 tr:1 ld:1 com:2 negentropy:9 numerical:3 update:4 parametrization:2 height:1 along:1 direct:1 become:3 differential:3 prove:1 advocate:1 huber:3 ica:4 indeed:2 decomposed:1 td:1 curse:9 jm:4 becomes:2 estimating:1 underlying:1 argmin:2 finding:4 transformation:2 guarantee:1 every:1 xd:1 no... |
936 | 1,857 | Tree-Based Modeling and Estimation of
Gaussian Processes on Graphs with Cycles
Martin J. Wainwright, Erik B. Sudderth, and Alan S. Willsky
Laboratory for Information and Decision Systems
Department of Electrical Engineering and Computer Science
Massachusetts Institute of Technology
Cambridge, MA 02139
{ mjwain, esudde... | 1857 |@word inversion:2 open:1 covariance:25 decomposition:1 initial:1 series:4 xiy:1 contains:1 selecting:1 current:1 comparing:1 must:2 finest:1 fn:1 subsequent:1 numerical:2 designed:1 update:1 intelligence:1 leaf:1 sys:1 ith:1 esuddert:1 short:1 provides:1 iterates:3 node:22 coarse:1 dn:1 direct:1 incorrect:1 rem:1... |
937 | 1,858 | Stagewise processing in error-correcting
codes and image restoration
K. Y. Michael Wong
Department of Physics, Hong Kong University of Science and Technology,
Clear Water Bay, Kowloon, Hong Kong
phkywong@ust.hk
Hidetoshi Nishimori
Department of Physics, Tokyo Institute of Technology,
Oh-Okayama, Meguro-ku, Tokyo 152-... | 1858 |@word kong:3 version:2 simulation:7 covariance:1 reduction:1 tuned:1 okayama:1 perturbative:2 ust:1 reminiscent:1 realistic:2 analytic:1 progressively:1 implying:1 weighing:1 ith:2 hamiltonian:4 provides:1 nishi:1 qualitative:1 consists:4 prove:1 introduce:3 expected:3 rapid:1 inappropriate:1 becomes:1 bounded:1 ... |
938 | 1,859 | Accumulator networks: Suitors of local
probability propagation
Brendan J. Frey and Anitha Kannan
Intelligent Algorithms Lab, University of Toronto, www. cs. toronto. edu/ "-'frey
Abstract
One way to approximate inference in richly-connected graphical
models is to apply the sum-product algorithm (a.k.a. probability pr... | 1859 |@word accounting:1 brightness:1 configuration:3 series:1 current:1 ixj:5 si:34 yet:1 kleen:1 update:1 intelligence:2 selected:1 toronto:2 sits:1 allerton:1 along:2 constructed:1 direct:1 ray:6 introduce:1 behavior:1 discretized:1 freeman:2 encouraging:1 becomes:2 mass:1 what:1 every:1 oscillates:1 zl:10 control:1... |
939 | 186 | 272
NEURAL NET RECEIVERS IN
MULTIPLE-ACCESS COMMUNICATIONS
Bernd-Peter Paris, Geoffrey Orsak, Mahesh Varanasi, Behnaam Aazhang
Department of Electrical and Computer Engineering
Rice University
Houston, TX 77251-1892
ABSTRACT
The application of neural networks to the demodulation of
spread-spectrum signals in a multip... | 186 |@word trial:3 version:1 manageable:1 polynomial:1 instrumental:2 simulation:8 pulse:1 subscriber:1 shot:1 reduction:1 configuration:5 ntc:1 slotted:1 multiuser:13 outperforms:1 comparing:1 nt:2 com:5 written:2 additive:3 numerical:1 drop:1 ith:2 filtered:1 node:1 unacceptable:3 direct:2 become:3 consists:1 introdu... |
940 | 1,860 | Natural sound statistics and divisive
normalization in the auditory system
Odelia Schwartz
Center for Neural Science
New York University
odelia@cns.nyu.edu
Eero P. Simoncelli
Howard Hughes Medical Institute
Center for Neural Science, and
Courant Institute of Mathematical Sciences
New York University
eero.simoncelli@n... | 1860 |@word eliminating:1 advantageous:1 seems:2 simulation:3 pressure:4 tuned:2 current:1 comparing:1 neurophys:1 must:3 additive:1 shape:1 remove:3 designed:1 plot:1 v:1 alone:1 tone:16 readability:1 five:1 mathematical:1 consists:1 fitting:1 dan:1 manner:1 ravindran:1 mask:6 expected:1 roughly:4 abscissa:1 examine:1... |
941 | 1,861 | Algorithms for Non-negative Matrix
Factorization
Daniel D. Lee*
*BelJ Laboratories
Lucent Technologies
Murray Hill, NJ 07974
H. Sebastian Seung*t
tDept. of Brain and Cog. Sci.
Massachusetts Institute of Technology
Cambridge, MA 02138
Abstract
Non-negative matrix factorization (NMF) has previously been shown to
be a ... | 1861 |@word h:2 version:1 compression:1 decomposition:1 att:4 daniel:1 ours:1 current:1 ka:1 yet:1 written:1 additive:3 numerical:4 update:32 stationary:2 warmuth:1 constructed:1 prove:3 consists:1 manner:1 indeed:1 themselves:1 nor:1 brain:2 considering:1 increasing:1 discover:1 bounded:2 factorized:1 linda:1 what:2 k... |
942 | 1,862 | The Kernel Trick for Distances
Bernhard SchOikopf
Microsoft Research
1 Guildhall Street
Cambridge, UK
bs@kyb.tuebingen.mpg.de
Abstract
A method is described which, like the kernel trick in support vector machines (SVMs), lets us generalize distance-based algorithms to operate
in feature spaces, usually nonlinearly re... | 1862 |@word norm:2 hannonic:1 pick:1 thatfor:1 contains:1 series:1 ka:4 surprising:1 yet:1 attracted:1 john:1 cruz:1 girosi:1 kyb:1 qiyi:1 drop:2 plot:2 short:2 math:1 herbrich:1 mathematical:2 ucsc:1 direct:1 indeed:1 mpg:1 automatically:1 actual:1 window:2 considering:2 becomes:1 project:1 underlying:3 moreover:3 wha... |
943 | 1,863 | An Information Maximization Approach to
Overcomplete and Recurrent Representations
Oren Shriki and Haim Sompolinsky
Racah Institute of Physics and
Center for Neural Computation
Hebrew University
Jerusalem, 91904, Israel
Daniel D. Lee
Bell Laboratories
Lucent Technologies
Murray Hill, NJ 07974
Abstract
The principle ... | 1863 |@word version:1 nd:1 proportionality:1 decomposition:1 tr:1 initial:1 series:1 daniel:1 recovered:2 si:1 wx:3 shape:1 update:2 generative:4 discovering:1 ith:1 sigmoidal:1 uxj:1 along:1 consists:1 fitting:1 yst:1 inside:1 ica:5 decomposed:1 jm:1 considering:1 underlying:1 suffice:1 maximizes:1 israel:2 interprete... |
944 | 1,864 | A New Approximate Maximal Margin
Classification Algorithm
Claudio Gentile
DSI, Universita' di Milano,
Via Comelico 39,
20135 Milano, Italy
gentile@dsi.unimi.it
Abstract
A new incremental learning algorithm is described which approximates
the maximal margin hyperplane w.r.t. norm p ~ 2 for a set of linearly
separable ... | 1864 |@word trial:10 version:3 briefly:1 polynomial:2 norm:27 seems:5 suitably:1 open:1 grey:1 bn:1 u11:1 pick:1 incurs:1 recursively:2 initial:2 att:1 current:4 comparing:1 com:1 fn:1 numerical:1 limp:1 designed:1 update:8 v:1 fewer:1 nips2000:1 warmuth:3 boosting:1 hyperplanes:1 mathematical:2 along:1 ik:1 scholkopf:... |
945 | 1,865 | Model Complexity, Goodness of Fit and
Diminishing Returns
Igor V. Cadez
Information and Computer Science
University of California
Irvine, CA 92697-3425, U.S.A.
Padhraic Smyth
Information and Computer Science
University of California
Irvine, CA 92697-3425, U.S.A.
Abstract
We investigate a general characteristic of th... | 1865 |@word proportion:2 proportionality:1 covariance:2 tr:1 initial:1 contains:1 score:2 series:4 selecting:1 cadez:4 com:1 numerical:1 alone:2 selected:1 record:1 characterization:1 parameterizations:1 simpler:2 height:1 constructed:1 consists:2 prove:1 specialize:1 fitting:1 expected:1 decomposed:3 actual:2 increasi... |
946 | 1,866 |
!
"
#
$
'
(
)
*
! ! "
# !$
&% )' ( *+ !
% &
,.-0/2Y135Z\47[67]_8:^`a]c9;bO-=]_d</?eGf>)gi@;hiABj:@Dk+CElT]m^`a1n35FGd$FH13JILk2Z:KMj!4$opIOd7NQbOPSgqk2RrQstITf67uv6B-=lO135bOI:]m4Vk]mU=^`af<TZ <TWX<T/
w dB[Z:]i^`an?pd7^gm`ar5[x... | 1866 |@word cu:2 pw:1 d2:1 llo:1 p0:1 k7:2 tr:1 o2:1 ka:1 od:3 gv:1 cfo:1 n0:1 v:1 dcfe:1 yr:1 nq:1 xk:1 lx:2 c6:1 zii:1 ik:1 x0:1 pkb:1 xz:2 uz:2 ol:2 td:1 jm:1 lxz:1 z:1 gid:1 k2:8 uk:1 zl:1 ly:1 t13:1 t1:1 ak:1 fpe:1 au:1 co:1 uy:1 vu:2 sq:3 sca:3 ga:3 nb:1 py:1 yt:1 l:1 kqr:1 oh:1 jtd:2 gm:2 pa:1 jk:2 q7:1 ep:1 ft:... |
947 | 1,867 | What can a single neuron compute?
Blaise Agiiera y Areas, l Adrienne L. Fairhall, 2 and William Bialek2
1 Rare Books Library, Princeton University, Princeton, New Jersey 08544
2NEC Research Institute, 4 Independence Way, Princeton, New Jersey 08540
blaisea@prineeton. edu {adrienne, bialek} @researeh. nj. nee. com
Abs... | 1867 |@word cm2:2 simulation:2 pulse:3 simplifying:1 covariance:13 carry:2 moment:2 reduction:1 current:12 com:1 discretization:1 surprising:1 must:1 physiol:1 realistic:1 interspike:1 mulated:1 reproducible:1 fewer:1 device:3 short:1 filtered:2 hodgkinhuxley:1 provides:1 completeness:1 successive:1 attack:1 dn:1 along... |
948 | 1,868 | Sparsity of data representation of optimal kernel
machine and leave-one-out estimator
A. Kowalczyk
Chief Technology Office, Telstra
770 Blackburn Road, Clayton, Vic. 3168, Australia
(adam.kowalczy k@team.telstra.com)
Abstract
Vapnik's result that the expectation of the generalisation error ofthe optimal hyperplane is... | 1868 |@word cpe:1 rreg:2 determinant:1 cox:1 polynomial:3 norm:1 stronger:1 achievable:1 open:3 necessity:1 celebrated:1 series:2 rkhs:2 com:1 xlr:1 additive:1 subsequent:1 girosi:1 analytic:7 greedy:1 selected:1 sys:1 ith:2 lr:1 characterization:1 math:1 location:1 hyperplanes:2 firstly:1 mathematical:1 direct:1 becom... |
949 | 1,869 | Text Classification using String Kernels
HUlna Lodhi
John Shawe-Taylor
N ello Cristianini
Chris Watkins
Department of Computer Science Royal Holloway, University of London
Egham, Surrey TW20 OEX, UK
{huma, john, nello, chrisw}Cdcs.rhbnc.ac.uk
Abstract
We introduce a novel kernel for comparing two text documents.
T... | 1869 |@word uee:1 sri:2 version:2 seems:1 lodhi:1 tr:1 efficacy:1 att:1 document:16 current:1 comparing:1 com:1 surprising:1 si:1 must:1 readily:1 john:2 cruz:1 fn:2 informative:1 wanted:1 remove:1 designed:1 prohibitive:1 selected:1 beginning:1 short:1 lr:1 normalising:1 ucsc:1 direct:3 prove:1 combine:1 inside:1 intr... |
950 | 187 | 671
PROGRAMMABLE ANALOG PULSE-FIRING
NEURAL NETWORKS
Alan F. Murray
Alister Hamilton
Dept. of Elec. Eng.,
Dept. of Elec. Eng.,
University of Edinburgh, University of Edinburgh,
Mayfield Road,
Mayfield Road,
Edinburgh, EH9 3JL
Edinburgh, EH9 3JL
United Kingdom.
United Kingdom.
Lionel Tarassenko
Dept. of Eng. Science,... | 187 |@word inversion:1 proportion:2 chopping:5 simulation:3 pulse:40 eng:3 reduction:1 united:3 activation:1 subsequent:1 remove:3 msb:1 intelligence:1 device:3 signalling:1 smith:2 burst:2 constructed:1 supply:2 consists:1 resistive:1 mayfield:2 simulator:1 integrator:2 ol:1 actual:1 little:1 begin:1 linearity:1 circu... |
951 | 1,870 | From Margin To Sparsity
Thore Graepel, Ralf Herbrich
Computer Science Department
Technical University of Berlin
Berlin, Germany
{guru, ralfh)@cs.tu-berlin.de
Robert C. Williamson
Department of Engineering
Australian National University
Canberra, Australia
Bob. Williamson@anu.edu.au
Abstract
We present an improvement... | 1870 |@word version:1 achievable:3 compression:6 advantageous:1 norm:4 stronger:1 crucially:1 llo:1 tr:1 series:1 outperforms:1 current:1 dx:8 intriguing:1 john:1 cruz:1 subsequent:1 update:1 greedy:1 warmuth:2 short:1 normalising:1 ron:7 herbrich:4 hyperplanes:1 org:1 ironically:1 mathematical:1 symposium:1 reinterpre... |
952 | 1,871 | Factored Semi-Tied Covariance Matrices
M.J.F. Gales
Cambridge University Engineering Department
Trumpington Street, Cambridge. CB2 IPZ
United Kingdom
mjfg@eng.cam.ac.uk
Abstract
A new form of covariance modelling for Gaussian mixture models and
hidden Markov models is presented. This is an extension to an efficient
f... | 1871 |@word determinant:1 version:3 r:2 covariance:34 eng:2 reduction:4 initial:5 united:1 selecting:1 current:4 must:3 written:4 john:1 update:2 generative:7 greedy:1 selected:1 ith:1 consists:1 themselves:1 decomposed:1 increasing:4 becomes:1 estimating:2 underlying:1 null:1 tying:1 developed:1 spoken:1 transformatio... |
953 | 1,872 | Dopamine Bonuses
Sham Kakade
Peter Dayan
Gatsby Computational Neuroscience Unit
17 Queen Square, London, England, WC1N 3AR.
sham@gat sby.u c l. ac . uk
da y a n @gat sby.u c l. ac .uk
Abstract
Substantial data support a temporal difference (TO) model of
dopamine (OA) neuron activity in which the cells provide a globa... | 1872 |@word trial:37 exploitation:3 seems:3 instrumental:1 nd:1 open:1 r:2 jacob:3 initial:6 horvitz:1 current:1 activation:9 must:2 subsequent:1 benign:1 motor:2 plot:15 sby:2 cue:6 leaf:1 unfamiliarity:1 rc:1 become:2 persistent:2 incorrect:1 consists:1 behavioral:1 theoretically:1 expected:4 indeed:2 behavior:9 them... |
954 | 1,873 | On a Connection between Kernel PCA
and Metric Multidimensional Scaling
Christopher K. I. WilliaIns
Division of Informatics
The University of Edinburgh
5 Forrest Hill, Edinburgh EH1 2QL, UK
c.k.i.williams~ed.ac.uk
http://anc.ed.ac.uk
Abstract
In this paper we show that the kernel peA algorithm of Sch6lkopf
et al (199... | 1873 |@word cox:6 polynomial:2 sammon:3 crucially:1 covariance:7 kent:1 carry:1 configuration:7 series:1 current:1 scatter:1 written:1 realize:1 analytic:1 plot:2 stationary:7 intelligence:1 fewer:1 item:1 isotropic:9 short:1 postal:1 location:2 hah:1 scholkopf:1 decreasing:1 increasing:1 what:1 kind:3 interpreted:3 mi... |
955 | 1,874 | Machine Learning for Video-Based
Rendering
Arno Schadl
arno@schoedl. org
Irfan Essa
irjan@cc.gatech.edu
Georgia Institute of Technology
GVU Center / College of Computing
Atlanta, GA 30332-0280, USA.
Abstract
We present techniques for rendering and animation of realistic
scenes by analyzing and training on short vid... | 1874 |@word manageable:1 series:1 animated:1 current:5 must:1 realistic:2 fewer:1 guess:1 beginning:1 short:2 record:1 location:1 org:1 unacceptable:1 predecessor:1 ik:8 fitting:1 introduce:1 manner:1 expected:1 animator:1 automatically:3 little:1 insure:1 advent:1 lowest:2 arno:2 finding:2 every:6 iearning:1 estimat:1... |
956 | 1,875 | An Adaptive Metric Machine for Pattern
Classification
Carlotta Domeniconi, Jing Peng+, Dimitrios Gunopulos
Dept. of Computer Science, University of California, Riverside, CA 92521
+ Dept. of Computer Science, Oklahoma State University, Stillwater, OK 74078
{ carlotta, dg} @cs.ucr.edu, jpeng@cs.okstate.edu
Abstract
Nea... | 1875 |@word repository:2 duda:1 proportion:2 seems:2 simulation:1 carolina:1 thereby:2 recursively:1 carry:1 reduction:1 myles:1 contains:2 efficacy:1 series:1 dx:1 must:1 john:1 numerical:3 informative:1 plot:1 discrimination:1 intelligence:1 isotropic:1 ith:1 provides:1 location:4 five:3 adamenn:13 along:12 consists:... |
957 | 1,876 | Vicinal Risk Minimization
Olivier Chapelle, Jason Weston* , Leon Bottou and Vladimir Vapnik
AT&T Research Labs, 100 Schultz drive, Red Bank, NJ, USA
* Barnhill BioInformatics.com, Savannah, GA, USA.
{chapelle, weston,leonb, vlad}@research.att.com
Abstract
The Vicinal Risk Minimization principle establishes a bridge b... | 1876 |@word norm:1 suitably:2 covariance:2 carry:1 initial:4 contains:2 att:2 existing:4 com:3 comparing:1 yet:1 dx:1 must:1 distant:1 shape:2 offunctions:1 asymptote:1 update:2 discrimination:1 generative:9 selected:4 intelligence:1 provides:2 successive:1 five:1 mathematical:1 scholkopf:2 consists:1 combine:1 underfi... |
958 | 1,877 | A comparison of Image Processing
Techniques for Visual Speech Recognition
Applications
Michael S. Gray
Computational Neurobiology Laboratory
The Salk Institute
San Diego, CA 92186-5800
Terrence J. Sejnowski
Javier R. Movellan*
Computational Neurobiology Laboratory
The Salk Institute
San Diego, CA 92186-5800
Departm... | 1877 |@word version:4 tried:2 speechreading:1 accounting:1 decomposition:6 covariance:3 hager:1 reduction:1 selecting:2 hereafter:1 current:1 yet:1 cottrell:2 informative:1 half:1 selected:3 plane:1 filtered:2 provides:1 location:18 consists:1 ica:10 themselves:2 examine:1 decreasing:1 automatically:3 lll:1 matched:1 u... |
959 | 1,878 | Learning Sparse Image Codes using a
Wavelet Pyramid Architecture
Bruno A. Olshausen
Department of Psychology and
Center for Neuroscience, UC Davis
1544 Newton Ct.
Davis, CA 95616
baolshausen@uedavis.edu
Phil Sallee
Department of Computer Science
UC Davis
Davis, CA 95616
sallee@es.uedavis.edu
Michael S. Lewicki
Depar... | 1878 |@word version:1 compression:2 norm:1 contains:1 pub:1 shape:1 designed:1 update:2 v:1 generative:1 plane:2 iso:1 quantized:1 successive:1 simpler:1 mathematical:1 along:2 differential:1 become:1 warehouse:2 introduce:1 manner:1 upenn:1 mask:1 examine:1 multi:1 freeman:1 decreasing:1 jm:1 increasing:2 moreover:1 n... |
960 | 1,879 | On Reversing Jensen's Inequality
Tony Jebara
MIT Media Lab
Cambridge, MA 02139
jebam@media.mit.edu
Alex Pentland
MIT Media Lab
Cambridge, MA 02139
sandy@media.mit.edu
Abstract
Jensen's inequality is a powerful mathematical tool and one of the
workhorses in statistical learning. Its applications therein include the E... | 1879 |@word version:2 seems:1 nd:3 km:1 cml:11 simplifying:2 covariance:1 b39:1 invoking:1 configuration:2 current:2 yet:1 additive:1 shape:1 analytic:2 plot:1 depict:1 discrimination:5 v:1 generative:1 short:1 provides:1 sigmoidal:1 mathematical:2 direct:1 prove:1 introduce:1 indeed:1 roughly:1 themselves:1 proliferat... |
961 | 188 | 643
LEARNING SEQUENTIAL STRUCTURE
IN SIMPLE RECURRENT NETWORKS
David Servan-Schreiber. Axel Cleeremans. and James L. McClelland
Departtnents of Computer Science and Psycholgy
Carnegie Mellon University
Pittsburgh, PA 15213
ABSTRACT
We explore a network architecture introduced by Elman (1988) for
predicting successive... | 188 |@word trial:1 fmite:1 pick:1 initial:3 past:1 current:9 contextual:1 comparing:1 activation:29 si:2 interrupted:1 subsequent:1 progressively:1 selected:3 leaf:1 discovering:2 beginning:1 record:1 accepting:1 provides:1 node:9 contribute:1 successive:2 simpler:2 five:1 constructed:1 direct:1 become:1 predecessor:1 ... |
962 | 1,880 | Sparse Greedy
Gaussian Process Regression
Alex J. Smola?
RSISE and Department of Engineering
Australian National University
Canberra, ACT, 0200
Peter Bartlett
RSISE
Australian National University
Canberra, ACT, 0200
Alex.Smola@anu.edu.au
Peter.Bartlett@anu.edu.au
Abstract
We present a simple sparse greedy techniqu... | 1880 |@word repository:1 briefly:1 inversion:3 km:1 seek:2 pold:8 covariance:8 decomposition:3 dramatic:1 carry:1 contains:2 selecting:1 diagonalized:1 ka:3 yet:4 must:1 numerical:3 girosi:1 plot:1 update:1 v:1 infant:1 greedy:17 selected:1 location:2 bopt:5 zhang:1 direct:1 become:1 combine:1 inside:1 rapid:1 lrmxm:1 ... |
963 | 1,881 | Generalizable Singular Value
Decomposition for Ill-posed Datasets
Ulrik Kjerns
Lars K. Hansen
Department of Mathematical Modelling
Technical University of Denmark
DK-2800 Kgs. Lyngby, Denmark
uk, lkhansen@imm. dtu. dk
Stephen C. Strother
PET Imaging Service
VA medical center
Minneapolis
steve@pet. med. va. gov
Abst... | 1881 |@word determinant:2 advantageous:1 open:3 covariance:9 decomposition:6 thereby:1 tr:12 solid:2 contains:3 series:1 comparing:1 activation:2 john:1 subsequent:2 remove:1 plot:3 joy:1 lx:2 mathematical:1 direct:1 qij:1 consists:1 inside:3 manner:1 introduce:2 inter:1 mask:2 expected:1 ica:1 brain:6 decomposed:2 gov... |
964 | 1,882 | Temporally Dependent Plasticity:
An Information Theoretic Account
Gal Chechik
and
N aft ali Tishby
School of Computer Science and Engineering
and the Interdisciplinary Center for Neural Computation
The Hebrew University, Jerusalem, Israel
{ggal,tishby}@cs.huji.ac.il
Abstract
The paradigm of Hebbian learning has recen... | 1882 |@word hippocampus:3 mehta:1 additively:1 simulation:2 covariance:1 q1:2 solid:1 reduction:1 moment:3 series:1 efficacy:6 denoting:1 interestingly:1 amp:1 current:2 comparing:2 si:2 dx:1 aft:1 must:1 realize:1 plasticity:9 plot:1 aps:1 stationary:1 ith:1 short:1 supplying:1 zhang:1 lor:1 mathematical:1 along:3 con... |
965 | 1,883 | Position Variance, Recurrence and Perceptual
Learning
Zhaoping Li
Peter Dayan
Gatsby Computational Neuroscience Unit
17 Queen Square, London, England, WCIN 3AR.
zhaoping @g a t s by.u c l. a c.u k
da y a n @gat sby.u c l. ac .uk
Abstract
Stimulus arrays are inevitably presented at different positions on the
retina in ... | 1883 |@word h:1 trial:3 version:1 seems:1 seek:1 solid:2 accommodate:1 harder:1 interestingly:1 outperforms:1 wd:1 ixj:1 surprising:1 yet:3 import:1 must:2 additive:1 blur:1 plasticity:1 shape:1 drop:2 sby:1 discrimination:24 v:3 selected:1 core:1 short:1 manfred:1 provides:1 location:3 zhang:1 height:1 mathematical:1 ... |
966 | 1,884 | Whence Sparseness?
C. van Vreeswijk
Gatsby Computational Neuroscience Unit
University College London
17 Queen Square, London WCIN 3AR, United Kingdom
Abstract
It has been shown that the receptive fields of simple cells in VI can be explained by assuming optimal encoding, provided that an extra constraint
of sparsenes... | 1884 |@word version:1 simulation:1 pressure:2 solid:1 awij:7 vigorously:1 initial:1 necessity:1 united:1 denoting:1 optican:1 recovered:1 yet:1 subsequent:1 happen:1 analytic:1 cheap:1 smith:1 short:2 nom:1 wijsj:1 become:1 uwm:3 manner:1 inter:2 notably:1 roughly:2 p1:1 brain:2 automatically:1 window:4 becomes:1 provi... |
967 | 1,885 | One Microphone Source Separation
Sam T. Roweis
Gatsby Unit, University College London
roweis@gatsby.ucl. a c.uk
Abstract
Source separation, or computational auditory scene analysis , attempts to extract
individual acoustic objects from input which contains a mixture of sounds from
different sources, altered by the ac... | 1885 |@word middle:1 version:2 cleanly:1 grey:1 crucially:1 mitsubishi:1 simplifying:1 covariance:3 decomposition:1 carry:1 initial:1 configuration:1 contains:2 score:7 exclusively:1 dff:1 existing:2 kmk:1 recovered:3 comparing:1 must:1 written:1 realistic:1 subsequent:1 visible:1 designed:1 plot:1 v:1 cue:5 pursued:1 ... |
968 | 1,886 | Rate-coded Restricted Boltzmann Machines for
Face Recognition
Vee WhyeTeh
Department of Computer Science
University of Toronto
Toronto M5S 2Z9 Canada
Geoffrey E. Hinton
Gatsby Computational Neuroscience UnitUniversity College London
London WCIN 3AR u.K.
ywteh@cs.toronto.edu
hinton@ gatsby. ucl.ac. uk
Abstract
We d... | 1886 |@word version:1 seems:2 simulation:1 tried:1 contrastive:3 q1:2 tr:1 harder:1 contains:5 score:6 comparing:1 activation:6 si:4 yet:1 visible:11 partition:1 shape:3 remove:2 update:2 generative:6 half:6 intelligence:1 tone:3 toronto:3 five:2 unbounded:1 incorrect:1 consists:1 mask:1 expected:4 nor:1 inspired:2 pro... |
969 | 1,887 | Support Vector Novelty Detection
Applied to Jet Engine Vibration Spectra
Paul Hayton
Department of Engineering Science
University of Oxford, UK
pmh@robots.ox.ac.uk
Bernhard SchOlkopf
Microsoft Research
1 Guildhall Street, Cambridge, UK
bsc@scientist.com
Lionel Tarassenko
Department of Engineering Science
University ... | 1887 |@word msr:2 seek:1 eng:1 thereby:1 tr:2 score:1 tachometer:1 current:1 com:1 yet:1 written:1 john:2 shape:8 analytic:1 fewer:1 beginning:1 short:2 prespecified:1 characterization:1 detecting:1 provides:4 sits:1 along:2 direct:1 become:2 scholkopf:4 shorthand:1 consists:2 inside:1 introduce:2 indeed:2 expected:1 l... |
970 | 1,888 | Using Free Energies to Represent Q-values in a
Multiagent Reinforcement Learning Task
Brian Sallans
Geoffrey E. Hinton
Department of Computer Science
University of Toronto
Toronto M5S 2Z9 Canada
sallam'@cs,toronto,edu
Gatsby Computational Neuroscience Unit
University College London
London WCIN 3AR u.K.
hinton @ gat... | 1888 |@word trial:3 exploitation:2 middle:1 advantageous:1 contrastive:1 minus:1 tr:1 initial:2 selecting:3 ours:1 past:1 current:3 ka:1 recovered:1 activation:1 must:4 belmont:1 additive:5 partition:1 designed:1 update:5 stationary:1 intelligence:2 selected:1 parameterization:1 ith:1 short:1 coarse:1 toronto:3 ik:2 co... |
971 | 1,889 | Adaptive Object Representation with
Hierarchically-Distributed Memory Sites
Bosco S. Tjan
Department of Psychology
University of Southern California
btjan@usc.edu
Abstract
Theories of object recognition often assume that only one representation scheme is used within one visual-processing pathway. Versatility
of the v... | 1889 |@word bosco:1 trial:3 middle:1 version:1 fusiform:2 seems:2 simulation:1 minus:4 solid:2 extrastriate:1 contains:1 exclusively:1 existing:2 current:1 emory:1 activation:3 yet:2 issuing:3 exposing:1 shape:1 designed:1 progressively:1 v:1 alone:2 fewer:1 item:5 provides:1 disoriented:1 simpler:1 along:4 constructed... |
972 | 189 | 769
A SELF-LEARNING NEURAL NETWORK
A. Hartstein and R. H. Koch
IBM - Thomas J. Watson Research Center
Yorktown Heights, New York
ABSTRACf
We propose a new neural network structure that is compatible
with silicon technology and has built-in learning capability. The
thrust of this network work is a new synapse function... | 189 |@word open:1 simulation:7 initial:2 t7:1 current:1 readily:1 realize:1 partition:1 thrust:2 fewer:1 device:5 guess:1 provides:2 node:1 height:1 direct:1 become:1 viable:1 retrieving:1 behavior:3 decreasing:2 little:1 becomes:3 erase:1 what:2 easiest:1 gutfreund:1 finding:6 every:1 shed:1 appear:1 dropped:1 local:1... |
973 | 1,890 | Second order approximations for probability
models
Hilbert J. Kappen
Department of Biophysics
Nijmegen University
Nijmegen, the Netherlands
bert@mbfys.kun.nl
Wim Wiegerinck
Department of Biophysics
Nijmegen University
Nijmegen, the Netherlands
wimw@mbfys.kun.nl
Abstract
In this paper, we derive a second order mean f... | 1890 |@word polynomial:1 tedious:1 open:1 grey:1 solid:3 kappen:9 series:2 contains:2 mag:1 si:12 yet:1 written:1 must:2 numerical:3 partition:1 intelligence:1 xk:13 node:22 lor:1 direct:1 differential:1 combine:1 introduce:1 mbfys:2 p1:2 themselves:1 cpu:2 becomes:1 notation:1 factorized:10 lowest:2 interpreted:1 mini... |
974 | 1,891 | Convergence of Large Margin Separable Linear
Classification
Tong Zhang
Mathematical Sciences Department
IBM TJ. Watson Research Center
Yorktown Heights, NY 10598
tzhang@watson.ibm.com
Abstract
Large margin linear classification methods have been successfully applied to many applications. For a linearly separable prob... | 1891 |@word ia2:2 version:1 briefly:1 eliminating:1 norm:1 seek:1 thatfor:1 moment:1 contains:1 existing:1 com:1 z2:2 surprising:1 comparing:1 xiyi:10 must:1 john:4 additive:1 numerical:3 update:1 alone:2 warmuth:1 mln:1 zhang:2 height:1 mathematical:1 rc:1 direct:1 become:1 scholkopf:1 prove:1 interscience:1 expected:... |
975 | 1,892 | Learning Switching Linear Models of Human
Motion
Vladimir Pavlovic and James M. Rehg
Compaq - Cambridge Research Lab
Cambridge, MA 02139
{vladimir.pavlovic,jim.rehg}@compaq.com
John MacCormick
Compaq - System Research Center
Palo Alto, CA 94301
{john.maccormick} @compaq.com
Abstract
The human figure exhibits complex... | 1892 |@word seems:1 simulation:1 pick:1 tr:1 initial:3 series:1 ours:1 outperforms:1 com:2 surprising:2 written:2 e01:1 john:2 readily:1 realistic:2 update:3 cue:1 selected:2 isard:1 discovering:1 rts:3 intelligence:1 plane:2 iso:1 provides:5 constructed:2 xtl:2 dpr:1 indeed:2 expected:1 roughly:1 behavior:2 ldss:1 mul... |
976 | 1,893 | Sparse Representation for Gaussian Process
Models
Lehel Csat6 and Manfred Opper
Neural Computing Research Group
School of Engineering and Applied Sciences
B4 7ET Birmingham, United Kingdom
{csat o l, oppe r m} @as t o n. ac .uk
Abstract
We develop an approach for a sparse representation for Gaussian Process
(GP) mode... | 1893 |@word briefly:1 inversion:3 polynomial:2 seems:3 open:2 grey:1 simulation:2 covariance:3 decomposition:2 moment:1 phy:1 contains:2 score:7 united:1 att:1 existing:1 john:1 numerical:1 kleen:2 update:14 v:1 leaf:1 smith:1 manfred:1 postal:1 org:1 scholkopf:1 combine:1 ra:1 preclude:1 project:1 ti:1 bernardo:1 exac... |
977 | 1,894 | Universality and individuality in a neural
code
Elad Schneidman,1,2 Naama Brenner,3 Naftali Tishby,1,3
Rob R. de Ruyter van Steveninck, 3 William Bialek3
ISchool of Computer Science and Engineering, Center for Neural Computation and
2Department of Neurobiology, Hebrew University, Jerusalem 91904, Israel
3NEC Research I... | 1894 |@word trial:2 middle:2 version:1 seems:1 carry:4 com:1 universality:3 yet:2 must:1 written:1 physiol:1 informative:1 motor:1 reproducible:1 plot:1 v:4 alone:2 nervous:2 beginning:2 sys:1 ith:2 short:1 record:1 iog2:1 provides:4 contribute:1 codebook:4 mathematical:1 qualitative:2 elads:1 expected:1 behavior:4 nor... |
978 | 1,895 | Overfitting in Neural Nets: Backpropagation,
Conjugate Gradient, and Early Stopping
Rich Caruana
CALD,CMU
5000 Forbes Ave.
Pittsburgh, PA 15213
caruana@cs.cmu.edu
Steve Lawrence
NEC Research Institute
4 Independence Way
Princeton, NJ 08540
lawrence@ research. nj. nec. com
Lee Giles
Information Sciences
Penn State Un... | 1895 |@word trial:2 polynomial:8 nd:1 hu:26 minus:1 reduction:1 initial:1 com:1 subcomponents:1 yet:1 written:1 must:1 distant:1 plot:1 update:5 v:3 hallway:1 node:3 five:1 fitting:1 underfitting:1 ra:1 expected:3 overtrain:1 behavior:4 examine:1 multi:1 ol:1 simulator:1 brain:1 little:3 becomes:2 xx:1 linearity:11 und... |
979 | 1,896 | Data clustering by Markovian relaxation
and the Information Bottleneck Method
N aft ali Tishby
and
N oam Slonim
School of Computer Science and Engineering and Center for Neural Computation *
The Hebrew University, Jerusalem, 91904 Israel
email: {tishby.noamm}ees.huji.ae.il
Abstract
We introduce a new, non-parametric a... | 1896 |@word middle:1 compression:1 diametrically:1 gish:1 initial:9 denoting:2 ixj:2 yet:1 aft:1 john:1 additive:1 partition:1 hofmann:1 enables:2 shape:1 plot:2 xex:2 v:1 stationary:4 greedy:1 eshkol:1 noamm:1 plane:1 provides:1 node:1 location:2 lx:5 allerton:1 simpler:1 direct:3 become:2 consists:1 combine:2 introdu... |
980 | 1,897 | The Use of MDL to Select among
Computational Models of Cognition
In J. Myung, Mark A. Pitt & Shaobo Zhang Vijay Balasubramanian
Department of Psychology
David Rittenhouse Laboratories
Ohio State University
University of Pennsylvania
Columbus, OH 43210
Philadelphia, PA 19103
{myung.l, pitt.2}@osu.edu
vijay@endiv.hep.u... | 1897 |@word compression:1 seems:1 proportion:1 simulation:3 uncovers:1 phy:1 contains:1 series:1 selecting:3 recovered:2 od:3 must:4 written:3 distant:1 enables:2 fund:1 v:2 half:1 selected:3 provides:5 zhang:2 mathematical:1 along:1 enterprise:1 differential:5 become:1 symposium:1 fitting:2 manner:1 introduce:1 theore... |
981 | 1,898 | Learning Joint Statistical Models for
Audio-Visual Fusion and Segregation
John W. Fisher 111*
Massachusetts Institute of Technology
Cambridge, MA 02139
fisher@ai.mit.edu
Trevor Darrell
Massachusetts Institute of Technology
Cambridge, MA 02139
trevor@ai.mit.edu
William T. Freeman
Mitsubishi Electric Research Laborato... | 1898 |@word version:1 underperform:1 mitsubishi:1 simplifying:1 solid:1 reduction:2 initial:2 series:1 efficacy:1 selecting:1 ours:1 recovered:3 com:1 yet:1 must:2 john:2 realistic:1 kleen:1 informative:4 analytic:1 cue:1 selected:1 short:1 compo:1 coarse:1 constructed:1 prove:2 manner:1 periodograms:1 multi:6 freeman:... |
982 | 1,899 | Some new bounds on the generalization error of
combined classifiers
Vladimir Koltchinskii
Department of Mathematics and Statistics
University of New Mexico
Albuquerque, NM 87131-1141
vlad@math.unm.edu
Dmitriy Panchenko
Department of Mathematics and Statistics
University of New Mexico
Albuquerque, NM 87131-1141
panche... | 1899 |@word version:1 norm:2 seems:1 nd:2 recursively:1 chervonenkis:1 dpn:1 nt:1 plot:2 intelligence:1 lr:5 provides:1 boosting:11 math:2 simpler:1 prove:1 consists:1 introduce:1 expected:1 roughly:1 multi:1 eurocolt:1 cpu:2 actual:1 increasing:1 conv:5 provided:2 becomes:1 bounded:1 what:1 substantially:1 voting:3 xd... |
983 | 19 | 474
OPTIMIZAnON WITH ARTIFICIAL NEURAL NETWORK SYSTEMS:
A MAPPING PRINCIPLE
AND
A COMPARISON TO GRADIENT BASED METHODS t
Harrison MonFook Leong
Research Institute for Advanced Computer Science
NASA Ames Research Center 230-5
Moffett Field, CA, 94035
ABSTRACT
General formulae for mapping optimization problems into sys... | 19 |@word erate:1 version:2 seems:1 r:1 simulation:10 shading:5 moment:17 electronics:1 initial:12 contains:1 selecting:1 genetic:2 ala:1 lapedes:2 current:1 comparing:3 com:1 written:1 must:2 john:1 numerical:6 subsequent:1 j1:1 half:6 instantiate:2 device:3 guess:3 nervous:1 beginning:1 ith:1 short:1 dissertation:2 c... |
984 | 190 | 282
Kanerva
Contour-Map Encoding of Shape for Early Vision
Pentti Kanerva
Research Institute for Advanced Computer Science
Mail Stop 230-5, NASA Ames Research Center
Moffett Field, California 94035
ABSTRACT
Contour maps provide a general method for
recognizing two-dimensional shapes. All but
blank images give rise t... | 190 |@word middle:1 mammal:1 minus:2 initial:1 blank:1 comparing:2 must:1 slanted:1 shape:8 discrimination:1 leaf:2 guess:1 tone:1 accordingly:1 beginning:1 short:1 coarse:1 location:2 ames:1 lbo:1 along:1 constructed:1 combine:1 recognizable:1 manner:1 indeed:1 rapid:1 roughly:3 inspired:1 detects:1 automatically:1 ma... |
985 | 1,900 | Mixtures of Gaussian Processes
Volker Tresp
Siemens AG, Corporate Technology, Department of Neural Computation
Otto-Hahn-Ring 6,81730 Miinchen, Germany
Volker. Tresp@mchp.siemens.de
Abstract
We introduce the mixture of Gaussian processes (MGP) model which is
useful for applications in which the optimal bandwidth of a... | 1900 |@word briefly:1 inversion:1 covariance:1 jacob:4 pressure:2 contains:1 selecting:1 current:1 nowlan:1 si:3 additive:1 kdd:1 hofmann:4 cheap:1 plot:4 update:2 intelligence:2 xk:17 dissertation:1 provides:1 miinchen:1 preference:2 location:1 introduce:3 expected:1 jm:1 underlying:1 notation:1 lowest:1 minimizes:1 a... |
986 | 1,901 | Interactive Parts Model: an Application
to Recognition of On-line Cursive Script
Predrag Neskovic, Philip C Davis' and Leon N Cooper
Physics Department and Institute for Brain and Neural Systems
Brown University, Providence, RI 02912
Abstract
In this work, we introduce an Interactive Parts (IP) model as an
alternativ... | 1901 |@word version:1 retraining:1 propagate:2 mention:1 solid:2 configuration:3 contains:2 selecting:1 past:1 contextual:1 erms:2 must:1 written:1 shape:8 v:1 selected:1 beginning:2 provides:1 location:9 constructed:1 supply:1 symposium:1 consists:1 introduce:3 pairwise:9 expected:3 multi:1 brain:3 considering:1 incre... |
987 | 1,902 | Noise suppression based on
neurophysiologically-motivated SNR
estimation for robust speech recognition
J iirgen Tcharz
Medical Physics Group
Oldenburg University
26111 Oldenburg
Germany
tch@medi.physik.uni-oldenburg.de
Michael Kleinschmidt
Medical Physics Group
Oldenburg University
26111 Oldenburg
Germany
Birger Kal... | 1902 |@word compression:2 advantageous:1 annoying:1 physik:2 simulation:2 meansquare:1 mammal:1 reduction:1 oldenburg:7 medi:1 yet:1 must:1 physiol:1 additive:1 speakerindependent:1 shape:3 update:1 discrimination:1 stationary:4 cue:2 imitate:1 tone:1 short:2 supplying:1 filtered:1 compo:1 supply:1 incorrect:1 consists... |
988 | 1,903 | Processing of Time Series by Neural Circuits
with Biologically Realistic Synaptic Dynamics
Thomas NatschIager & Wolfgang Maass
Institute for Theoretical Computer Science
Technische Universitat Graz, Austria
{tna t schl,maass }@i g i.tu -gra z. ac . a t
Eduardo D. Sontag
Anthony Zador
Dept. of Mathematics
Rutgers Uni... | 1903 |@word trial:2 dtk:1 version:1 hippocampus:1 open:1 r:1 simulation:1 pressure:1 series:12 efficacy:4 contains:1 past:1 current:1 nt:1 od:1 john:1 numerical:1 realistic:3 subsequent:1 plasticity:5 offunctions:1 motor:1 designed:3 alone:2 realism:1 short:3 weierstrass:1 characterization:3 provides:1 org:1 sigmoidal:... |
989 | 1,904 | Competition and Arbors in Ocular Dominance
Peter Dayan
Gatsby Computational Neuroscience Unit, UCL
17 Queen Square, London, England, WCIN 3AR.
d a y a n @gat sby.u c l.a c .uk
Abstract
Hebbian and competitive Hebbian algorithms are almost ubiquitous in
modeling pattern formation in cortical development. We analyse in ... | 1904 |@word middle:1 polynomial:5 stronger:1 simulation:1 seek:1 commute:1 solid:9 harder:1 initial:6 exclusively:1 analysed:1 activation:1 must:2 bd:1 john:1 piepenbrock:7 asymptote:3 plot:6 update:1 sby:1 footing:1 location:3 organising:2 ofo:1 lor:1 hermite:2 manner:1 indeed:1 expected:1 behavior:1 roughly:2 aliasin... |
990 | 1,905 | A silicon primitive for competitive learning
David Usu
Miguel Figueroa
Chris Diorio
Computer Science and Engineering
The University of Washington
114 Sieg Hall, Box 352350
Seattle, W A 98195-2350 USA
hsud, miguel, diorio@cs.washington.edu
Abstract
Competitive learning is a technique for training classification and... | 1905 |@word trial:1 version:1 briefly:1 simulation:2 covariance:1 cp2:1 existing:1 current:15 ihei:25 refresh:1 subsequent:1 shape:1 enables:1 designed:1 update:5 v:1 device:8 floatinggate:1 dfl:1 provides:2 cse:1 location:1 sieg:1 along:3 differential:4 m7:2 combine:1 paragraph:1 behavior:3 nor:1 brain:1 terminal:2 m8... |
991 | 1,906 | Weak Learners and Improved Rates of
Convergence in Boosting
Shie Mannor and Ron Meir
Department of Electrical Engineering
Technion, Haifa 32000, Israel
{shie,rmeir }@{techunix,ee}.technion.ac.il
Abstract
The problem of constructing weak classifiers for boosting algorithms is studied. We present an algorithm that pro... | 1906 |@word briefly:1 version:2 polynomial:1 achievable:1 seems:1 nd:1 open:1 simulation:5 mention:1 ld:2 offering:1 yet:1 additive:1 numerical:4 partition:6 greedy:1 plane:1 xk:2 characterization:1 provides:2 mannor:2 boosting:16 ron:1 hyperplanes:1 completeness:1 zhang:1 constructed:2 become:1 ik:2 consists:1 combine... |
992 | 1,907 | Propagation Algorithms for Variational
Bayesian Learning
Zoubin GhahraIllani and Matthew J. Beal
Gatsby Computational Neuroscience Unit
University College London
17 Queen Square, London WC1N 3AR, England
{zoubin,m.beal}~gatsby.ucl.ac.uk
Abstract
Variational approximations are becoming a widespread tool for
Bayesian l... | 1907 |@word manageable:1 unif:1 calculus:1 covariance:8 pressure:1 mention:1 tr:2 initial:1 series:6 contains:3 recovered:2 surprising:1 yet:1 dx:1 written:1 readily:1 cruz:1 visible:2 analytic:1 update:1 progressively:1 v:1 discrimination:1 maximised:1 sys:2 provides:2 node:3 ssm:5 simpler:1 become:1 manner:1 yllxl:1 ... |
993 | 1,908 | Speech Denoising and Dereverberation Using
Probabilistic Models
Hagai Attias
John C. Platt
Alex Acero
Li Deng
Microsoft Research
1 Microsoft Way
Redmond, WA 98052
{hagaia,jplatt,alexac,deng} @microsoft.com
Abstract
This paper presents a unified probabilistic framework for denoising and
dereverberation of speech s... | 1908 |@word msr:1 version:1 brandstein:1 bigram:1 open:1 r:3 covariance:1 invoking:1 tr:1 reduction:1 initial:1 series:1 dff:5 past:1 existing:1 current:2 com:2 si:1 artijiciallntelligence:1 must:2 john:1 realistic:1 shape:1 remove:1 update:4 v:4 stationary:5 xk:4 dembo:1 simpler:1 mathematical:3 become:1 autocorrelati... |
994 | 1,909 | Color Opponency Constitutes A Sparse
Representation For the Chromatic
Structure of Natural Scenes
Te-Won Lee; Thomas Wachtler and Terrence Sejnowski
Institute for Neural Computation, University of California, San Diego &
Computational Neurobiology Laboratory, The Salk Institute
10010 N. Torrey Pines Road
La Jolla, Cal... | 1909 |@word version:1 norm:4 decomposition:1 brightness:2 gjb:1 interestingly:1 activation:2 si:6 tilted:2 informative:1 shape:1 plot:1 alone:1 selected:1 plane:4 provides:1 location:1 along:9 inside:1 manner:1 pairwise:1 twer:1 ica:6 roughly:2 krauskopf:2 decreasing:3 estimating:1 moreover:1 lowest:1 what:1 minimizes:... |
995 | 191 | 348
Farotimi, Demho and Kailath
Neural Network Weight Matrix Synthesis Using
Optimal Control Techniques
O. Farotimi
A. Dembo
Information Systems Lab.
Electrical Engineering Dept.
Stanford University,
Stanford, CA 94305
T. Kailath
ABSTRACT
Given a set of input-output training samples, we describe a procedure for d... | 191 |@word version:1 heuristically:1 seek:1 simulation:5 initial:3 activation:5 discovering:1 dembo:4 hamiltonian:1 lr:1 mathematical:1 along:1 differential:3 become:1 replication:1 consists:1 behavior:1 abscissa:1 examine:1 chi:1 considering:1 becomes:1 discover:1 underlying:5 what:1 developed:1 finding:1 transformati... |
996 | 1,910 | Foundations for a Circuit Complexity Theory of
Sensory Processing*
Robert A. Legenstein & Wolfgang Maass
Institute for Theoretical Computer Science
Technische Universitat Graz, Austria
{Iegi, maass }@igi.tu-graz.ac.at
Abstract
We introduce total wire length as salient complexity measure for an analysis of the circuit... | 1910 |@word middle:2 polynomial:2 seems:1 nd:2 grey:2 km:2 simulation:1 bn:1 concise:1 solid:1 prefix:1 rightmost:2 existing:3 savage:3 subsequent:1 underly:1 realistic:6 leaf:5 nervous:1 plane:8 xk:1 provides:4 node:9 location:16 ofo:3 mathematical:4 along:1 direct:1 become:2 loll:1 consists:1 manner:2 introduce:1 g4:... |
997 | 1,911 | Reinforcement Learning with Function
Approximation Converges to a Region
Geoffrey J. Gordon
ggordon@es.emu.edu
Abstract
Many algorithms for approximate reinforcement learning are not
known to converge. In fact, there are counterexamples showing
that the adjustable weights in some algorithms may oscillate within
a reg... | 1911 |@word version:2 stronger:1 norm:6 twelfth:1 open:3 crucially:1 contraction:3 pick:2 carry:1 initial:2 contains:1 fragment:2 selecting:1 series:2 interestingly:1 current:5 wd:1 must:5 written:2 subsequent:1 happen:2 wll:1 update:23 greedy:9 leaf:1 beginning:2 iterates:2 provides:1 successive:1 simpler:1 prove:1 in... |
998 | 1,912 | Smart Vision Chip Fabricated Using Three
Dimensional Integration Technology
H.Kurino, M.Nakagawa, K.W .Lee, T.Nakamura,
Y.Yamada, K.T.Park and M.Koyanagi
Dept. of Machine Intelligence
and Systems Engineering,
Tohoku University
01, Aza-Aramaki, Aoba-ku, Sendai
980-8579, Japan
kurino@sd.mech.tohoku.ac.jp
Abstract
The s... | 1912 |@word aramaki:1 implemented:1 proportion:1 direction:1 realized:2 receptive:1 lou:1 preliminary:1 biological:4 consumer:1 image:7 considered:1 si:1 aoba:1 great:1 recently:1 difficult:1 must:1 bump:1 realize:3 jp:1 purpose:2 designed:1 design:1 analog:2 intelligence:1 vertical:3 realizing:2 yamada:1 sensor:1 succ... |
999 | 1,913 | Shape Context: A new descriptor for
shape matching and object recognition
Serge Belongie, Jitendra Malik and Jan Puzicha
Department of Electrical Engineering and Computer Sciences
University of California at Berkeley
Berkeley, CA 94720, USA
{sjb, malik,puzicha} @cs.berkeley.edu
Abstract
We develop an approach to obje... | 1913 |@word version:2 smirnov:2 tried:1 decomposition:1 brightness:6 moment:2 configuration:1 score:3 att:1 selecting:1 document:1 past:1 comparing:1 com:1 yet:1 written:1 must:1 shape:97 enables:1 medial:1 v:1 alone:1 greedy:1 selected:3 plane:2 core:1 lr:1 coarse:2 iterates:1 node:1 location:2 mathematical:1 consists... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.