Advances in Neural Information Processing Systems 15: Proceedings of the 2002 ConferenceSuzanna Becker, Sebastian Thrun, Klaus Obermayer MIT Press, 2003 - 1687 頁 Proceedings of the 2002 Neural Information Processing Systems Conference. The annual Neural Information Processing (NIPS) meeting is the flagship conference on neural computation. The conference draws a diverse group of attendees--physicists, neuroscientists, mathematicians, statisticians, and computer scientists--and the presentations are interdisciplinary, with contributions in algorithms, learning theory, cognitive science, neuroscience, vision, speech and signal processing, reinforcement learning and control, implementations, and applications. Only about thirty percent of the papers submitted are accepted for presentation at NIPS, so the quality is exceptionally high. This volume contains all the papers presented at the 2002 conference. |
內容
Preface | xvii |
NIPS Committees | xxi |
Reviewers | xxiii |
Cognitive ScienceArtificial Intelligence | 1 |
Fast Exact Inference with a Factored Model for Natural Language Parsing | 3 |
Prediction and Semantic Association | 11 |
Replay Repair and Consolidation | 19 |
A Minimal Intervention Principle for Coordinated Movement | 27 |
Manifold Parzen Windows | 847 |
Stochastic Neighbor Embedding | 855 |
Automatic Alignment of Local Representations | 863 |
Informed Projections | 871 |
Extracting Relevant Structures with Side Information | 879 |
Critical Lines in Symmetry of Mixture Models and its Application to Component Splitting | 887 |
Handling Missing Data with Variational Bayesian Learning of ICA | 903 |
From Large Margins To Small Covering Numbers | 911 |
A Unified MDL Account of Human Learning of Regular and Irregular Categories | 35 |
TheoryBased Causal Inference | 43 |
How the Poverty of the Stimulus Solves the Poverty of the Stimulus | 51 |
Bayesian Models of Inductive Generalization | 59 |
Combining Dimensions and Features in SimilarityBased Representations | 67 |
Modeling Midazolams Effect on the Hippocampus and Recognition Memory | 75 |
Dynamical Causal Learning | 83 |
Visual Development Aids the Acquisition of Motion Velocity Sensitivities | 91 |
Timing and Partial Observability in the Dopamine System | 99 |
Automatic Acquisition and Efficient Representation of Syntactic Structures | 107 |
Binary Coding in Auditory Cortex | 117 |
How Linear are Auditory Cortical Responses? | 125 |
Neural Decoding of Cursor Motion Using a Kalman Filter | 133 |
Embedding Spiking Neurons in InnerProduct Spaces | 141 |
SpectroTemporal Receptive Fields of Subthreshold Responses in Auditory Cortex | 149 |
Temporal Coherence Natural Image Sequences and the Visual Cortex | 157 |
Learning in Spiking Neural Assemblies | 165 |
ACh and NE in the Neocortex | 173 |
Dopamine Induced Bistability Enhances Signal Processing in Spiny Neurons | 181 |
Convergence Properties of some SpikeTriggered Analysis Techniques | 189 |
Branching Law for Axons | 197 |
Binary Tuning is Optimal for Neural Rate Coding with High Temporal Resolution | 205 |
An Information Theoretic Approach to the Functional Classification of Neurons | 213 |
MortonStyle Factorial Coding of Color in Primary Visual Cortex | 221 |
A Model for RealTime Computation in Generic Neural Microcircuits | 229 |
Adaptation and Unsupervised Learning | 237 |
Analysis and Design | 245 |
Derivation of the Learning Rule | 253 |
Selectivity and Metaplasticity in a Unified CalciumDependent Model | 261 |
Complex Cells Learn Disparity and Translation Invariance from Natural Images | 269 |
Analyzing Neural Responses to Natural Signals | 277 |
Dynamical Constraints on Computing with Spike Timing in the Cortex | 285 |
Interpreting Neural Response Variability as Monte Carlo Sampling of the Posterior | 293 |
A Neural EdgeDetection Model for Enhanced Auditory Sensitivity in Modulated Noise | 301 |
An EstimationTheoretic Framework for the Presentation of Multiple Stimuli | 309 |
Evidence Optimization Techniques for Estimating StimulusResponse Functions | 317 |
Reconstructing StimulusDriven Neural Networks from Spike Times | 325 |
Theory | 333 |
DataDependent Bounds for Bayesian Mixture Methods | 335 |
A Statistical Mechanics Approach to Approximate Analytical Bootstrap Averages | 343 |
Maximum Likelihood and the Information Bottleneck | 351 |
Stable Fixed Points of Loopy Belief Propagation Are Minima of the Bethe Free Energy | 359 |
Concentration Inequalities for the Missing Mass and for Histogram Rule Error | 367 |
Dyadic Classification Trees via Structural Risk Minimization | 375 |
The Stability of Kernel Principal Components Analysis and its Relation to the Process Eigenspectrum | 383 |
Information Diffusion Kernels | 391 |
Scaling of ProbabilityBased Optimization Algorithms | 399 |
The Effect of Singularities in a Learning Machine when the True Parameters Do Not Lie on Such Singularities | 407 |
On the Complexity of Learning the Kernel Matrix | 415 |
A Toy Model | 423 |
Conditional Models on the Ranking Poset | 431 |
PACBayes Margins | 439 |
A Note on the Representational Incompatibility of Function Approximation and Factored Dynamics | 447 |
Fractional Belief Propagation | 455 |
Effective Dimension and Generalization of Kernel Learning | 471 |
Margin Analysis of the LVQ Algorithm | 479 |
MarginBased Algorithms for Information Filtering | 487 |
Hyperkernels | 495 |
Algorithms and Architectures | 501 |
Bayesian Monte Carlo | 503 |
MeanField Approach to a Probabilistic Model in Information Retrieval | 511 |
Distance Metric Learning with Application to Clustering with SideInformation | 519 |
Adapting Codes and Embeddings for Polychotomies | 527 |
KnowledgeBased Support Vector Machine Classifiers | 535 |
Gaussian Process Priors With Uncertain Inputs Application to MultipleStep Ahead Time Series Forecasting | 543 |
Kernel Design Using Boosting | 551 |
Generalizing Support Vector Machines via an Analogy to Electrostatic Systems | 559 |
Adaptive Scaling for Feature Selection in SVMs | 567 |
Support Vector Machines for MultipleInstance Learning | 575 |
Fast Kernels for String and Tree Matching | 583 |
Generalized Linear Models | 591 |
Cluster Kernels for SemiSupervised Learning | 599 |
Adaptive Nonlinear System Identification with Echo State Networks | 607 |
Rational Kernels | 615 |
The Informative Vector Machine | 623 |
StabilityBased Model Selection | 631 |
Feature Selection in MixtureBased Clustering | 639 |
String Kernels Fisher Kernels and Finite State Automata | 647 |
Boosting Density Estimation | 655 |
Independent Components Analysis through Product Density Estimation | 663 |
Learning Semantic Similarity | 671 |
Self Supervised Boosting | 679 |
The EM Family and Beyond | 687 |
Intrinsic Dimension Estimation Using Packing Numbers | 695 |
HalfLives of EigenFlows for Spectral Clustering | 703 |
On the Dirichlet Prior and Bayesian Regularization | 711 |
Global Versus Local Methods in Nonlinear Dimensionality Reduction | 719 |
Dynamic Bayesian Networks with Deterministic Latent Tables | 727 |
Parametric Mixture Models for MultiLabeled Text | 735 |
Clustering with the Fisher Score | 743 |
Adaptive Classification by Variational Kalman Filtering | 751 |
Boosted Dyadic Kernel Discriminants | 759 |
Regularized Greedy Importance Sampling | 767 |
OneClass LP Classifier for Dissimilarity Representations | 775 |
A Formulation for Minimax Probability Machine Regression | 783 |
A Variational Inference Engine for Bayesian Networks | 791 |
A Differential Semantics for Jointree Algorithms | 799 |
Constraint Classification for Multiclass Classification and Ranking | 807 |
Nash Propagation for Loopy Graphical Games | 815 |
Using Tarjans Red Rule for Fast Dependency Tree Construction | 823 |
Exact MAP Estimates by Hypertree Agreement | 831 |
Denoising Pairwise Data | 839 |
Learning with Multiple Labels | 919 |
Robust Novelty Detection with SingleClass MPM | 927 |
Artefactual Structure from Least Squares Multidimensional Scaling | 935 |
The Decision List Machine | 943 |
Using Manifold Structure for Partially Labelled Classification | 951 |
Ranking with Large Margin Principle Two Approaches | 959 |
Multiclass Learning by Probabilistic Embeddings | 967 |
Transductive and Inductive Methods for Approximate Gaussian Process Regression | 975 |
Charting a Manifold | 983 |
Annealing and the Rate Distortion Problem | 991 |
Discriminative Learning for Label Sequences via Boosting | 999 |
Discriminative Densities from Maximum Contrast Estimation | 1007 |
FloatBoost Learning for Classification | 1015 |
Incremental Gaussian Processes | 1023 |
Learning Graphical Models with Mercer Kernels | 1031 |
Multiple Cause Vector Quantization | 1039 |
Information Regularization with Partially Labeled Data | 1047 |
Derivative observations in Gaussian Process Models of Dynamic Systems | 1055 |
Multiplicative Updates for Nonnegative Quadratic Programming in Support Vector Machines | 1063 |
Location Estimation with a Differential Update Network | 1071 |
Realtime Particle Filters | 1079 |
Emerging Technologies | 1087 |
Optoelectronic Implementation of a FitzHughNagumo Neural Model | 1089 |
Circuit Model of ShortTerm Synaptic Dynamics | 1097 |
Adaptive Quantization and Density Estimation in Silicon | 1105 |
Neuromorphic Bistable VLSI Synapses with SpikeTimingDependent Plasticity | 1113 |
Retinal Processing Emulation in a Programmable 2Layer Analog Array Processor CMOS Chip | 1121 |
A Case Study | 1129 |
Combining Features for BCI | 1137 |
Classifying Patterns of Visual Motion a Neuromorphic Approach | 1145 |
Developing Topography and Ocular Dominance Using two a VLSI Vision Sensors and a Neurotrophic Model of Plasticity | 1153 |
Topographic Map Formation by Silicon Growth Cones | 1161 |
Spike TimingDependent Plasticity in the Address Domain | 1169 |
FieldProgrammable Learning Arrays | 1177 |
Speech and Signal Processing | 1185 |
ForwardDecoding KernelBased Phone Sequence Recognition | 1187 |
A Probabilistic Approach to Single Channel Blind Signal Separation | 1195 |
Toward Autonomous Agents with Perfect Pitch | 1203 |
Analysis of Information in Speech based on MANOVA | 1211 |
Bayesian Estimation of TimeFrequency Coefficients for Audio Signal Enhancement | 1219 |
Source Separation with a Sensor Array Using Graphical Models and Subband Filtering | 1227 |
An Asynchronous Hidden Markov Model for AudioVisual Speech Recognition | 1235 |
Monaural Speech Separation | 1243 |
Discriminative Binaural Sound Localization | 1251 |
Application of Variational Bayesian Approach to Speech Recognition | 1259 |
Visual Processing | 1267 |
Learning to Perceive Transparency from the Statistics of Natural Scenes | 1269 |
Learning to Detect Natural Image Boundaries Using Brightness and Texture | 1277 |
Fast TransformationInvariant Factor Analysis | 1285 |
A Prototype for Automatic Recognition of Spontaneous Facial Actions | 1293 |
Bayesian Image SuperResolution | 1301 |
A Bilinear Model for Sparse Coding | 1309 |
Dynamic Structure SuperResolution | 1317 |
Unsupervised Color Constancy | 1325 |
Recovering Articulated Model Topology from Observed Rigid Motion | 1333 |
Linear Combinations of Optic Flow Vectors for Estimating SelfMotion a RealWorld Test of a Neural Model | 1341 |
Learning Sparse Multiscale Image Representations | 1349 |
Scene Representations that Refer to the Image | 1357 |
Recovering Intrinsic Images from a Single Image | 1365 |
Feature Selection by Maximum Marginal Diversity | 1373 |
Learning Sparse Topographic Representations with Products of Studentt Distributions | 1381 |
A Model for Learning Variance Components of Natural Images | 1389 |
Kernels do the Trick | 1397 |
Concurrent Object Recognition and Segmentation by Graph Partitioning | 1405 |
Factorial Learning without Factorial Search | 1413 |
Applications | 1421 |
Identity Uncertainty and Citation Matching | 1423 |
Prediction of Rheumatoid Joint Inflammation Based on Laser Imaging | 1431 |
Mismatch String Kernels for SVM Protein Classification | 1439 |
GraphDriven Features Extraction from Microarray Data using Diffusion Kernels and Kernel CCA | 1447 |
RealTime Monitoring of Complex Industrial Processes with Particle Filters | 1455 |
A Maximum Entropy Approach To Collaborative Filtering in Dynamic Sparse HighDimensional Domains | 1463 |
Prediction of Protein Topologies Using Generalized IOHMMs and RNNs | 1471 |
Approximate Inference and ProteinFolding | 1479 |
Adaptive Caching by Refetching | 1487 |
Inferring a Semantic Representation of Text via CrossLanguage Correlation Analysis | 1495 |
Improving a Page Classifier with Anchor Extraction and Link Analysis | 1503 |
A Hierarchical Bayesian Markovian Model for Motifs in Biopolymer Sequences | 1511 |
Learning to Classify Galaxy Shapes using the EM Algorithm | 1519 |
A Probabilistic Approach to Querying on Music and Text | 1527 |
A Probabilistic Model for Learning Concatenative Morphology | 1535 |
Reinforcement Learning and Control | 1543 |
Learning Attractor Landscapes for Learning Motor Primitives | 1545 |
Learning a Forward Model of a Reflex | 1553 |
An Application to Robust Biped Walking | 1561 |
BiasOptimal Incremental Problem Solving | 1569 |
ValueDirected Compression of POMDPs | 1577 |
Optimality of Reinforcement Learning Algorithms with Linear Function Approximation | 1585 |
Speeding up the PartiGame Algorithm | 1593 |
Reinforcement Learning to Play an Optimal Nash Equilibrium in Team Markov Games | 1601 |
Convergent Combinations of Reinforcement Learning with Linear Function Approximation | 1609 |
Approximate Linear Programming for AverageCost Dynamic Programming | 1617 |
A Convergent Form of Approximate Policy Iteration | 1625 |
Efficient Learning Equilibrium | 1633 |
A TrajectoryBased Approach | 1641 |
Learning to Take Concurrent Actions | 1649 |
Learning in ZeroSum Team Markov Games Using Factored Value Functions | 1657 |
Exponential Family PCA for Belief Compression in POMDPs | 1665 |
Index of Authors | 1673 |
| 1679 | |
常見字詞
Advances in Neural algorithm analysis approach approximation average Bayesian Bayesian networks binary boosting bound cells classifier clustering complex components compute constraint convergence correlation corresponding cortex cortical covariance data set defined denote density derived dimension dimensional distribution dynamics eigenvalues eigenvectors equation error estimate example factor feature space Figure filter firing rate Fisher kernel Gaussian Gaussian process given graph inference input Isomap iteration Kalman filter kernel labels linear loss function Machine Learning mapping margin Markov matrix method metric Midazolam minimization Neural Information Processing Neural Networks neurons node noise nonlinear obtained optimal output parameters patterns performance points posterior prediction prior probabilistic probability problem random receptive fields regression representation response sample semantic sequence similar simulations spike statistical stimulus structure support vector machines Theorem theory training data tree unsupervised learning update variables variance weights