## Luc De Raedt, Stefan Wrobel

*Proceedings of the 22nd International Conference on Machine Learning*

ICML, 2005.

@proceedings{ICML-2005, address = "Bonn, Germany", editor = "Luc De Raedt and Stefan Wrobel", isbn = "1-59593-180-5", publisher = "{ACM}", series = "{ACM International Conference Proceeding Series}", title = "{Proceedings of the 22nd International Conference on Machine Learning}", volume = 119, year = 2005, }

### Contents (134 items)

- ICML-2005-AbbeelN #learning
- Exploration and apprenticeship learning in reinforcement learning (PA, AYN), pp. 1–8.
- ICML-2005-AndersonM #algorithm #learning #markov #modelling
- Active learning for Hidden Markov Models: objective functions and algorithms (BA, AM), pp. 9–16.
- ICML-2005-AngelopoulosC
- Tempering for Bayesian C&RT (NA, JC), pp. 17–24.
- ICML-2005-Angiulli #nearest neighbour #performance
- Fast condensed nearest neighbor rule (FA), pp. 25–32.
- ICML-2005-BachJ #composition #kernel #predict #rank
- Predictive low-rank decomposition for kernel methods (FRB, MIJ), pp. 33–40.
- ICML-2005-BekkermanEM #clustering #interactive #multi
- Multi-way distributional clustering via pairwise interactions (RB, REY, AM), pp. 41–48.
- ICML-2005-BeygelzimerDHLZ #classification #fault #reduction
- Error limiting reductions between classification tasks (AB, VD, TPH, JL, BZ), pp. 49–56.
- ICML-2005-BlockeelPS #learning #multi
- Multi-instance tree learning (HB, DP, AS), pp. 57–64.
- ICML-2005-BowlingGW
- Action respecting embedding (MHB, AG, DFW), pp. 65–72.
- ICML-2005-BreitenbachG #clustering #ranking
- Clustering through ranking on manifolds (MB, GZG), pp. 73–80.
- ICML-2005-BridewellALT #induction #process
- Reducing overfitting in process model induction (WB, NBA, PL, LT), pp. 81–88.
- ICML-2005-BurgesSRLDHH #learning #rank #using
- Learning to rank using gradient descent (CJCB, TS, ER, AL, MD, NH, GNH), pp. 89–96.
- ICML-2005-BurgeL #learning #network
- Learning class-discriminative dynamic Bayesian networks (JB, TL), pp. 97–104.
- ICML-2005-CalinonB #framework #gesture #probability #recognition #using
- Recognition and reproduction of gestures using a probabilistic framework combining PCA, ICA and HMM (SC, AB), pp. 105–112.
- ICML-2005-CarneyCDL #network #predict #probability #using
- Predicting probability distributions for surf height using an ensemble of mixture density networks (MC, PC, JD, CL), pp. 113–120.
- ICML-2005-ChangK #learning
- Hedged learning: regret-minimization with learning experts (YHC, LPK), pp. 121–128.
- ICML-2005-ChengJSW #image #modelling
- Variational Bayesian image modelling (LC, FJ, DS, SW), pp. 129–136.
- ICML-2005-ChuG #learning #process
- Preference learning with Gaussian processes (WC, ZG), pp. 137–144.
- ICML-2005-ChuK
- New approaches to support vector ordinal regression (WC, SSK), pp. 145–152.
- ICML-2005-CortesMW #learning
- A general regression technique for learning transductions (CC, MM, JW), pp. 153–160.
- ICML-2005-CrandallG #game studies #learning
- Learning to compete, compromise, and cooperate in repeated general-sum games (JWC, MAG), pp. 161–168.
- ICML-2005-DaumeM #approximate #learning #optimisation #predict #scalability
- Learning as search optimization: approximate large margin methods for structured prediction (HDI, DM), pp. 169–176.
- ICML-2005-TorreK #analysis #multimodal
- Multimodal oriented discriminant analysis (FDlT, TK), pp. 177–184.
- ICML-2005-DrakeV #learning
- A practical generalization of Fourier-based learning (AD, DV), pp. 185–192.
- ICML-2005-DriessensD #first-order #learning #modelling
- Combining model-based and instance-based learning for first order regression (KD, SD), pp. 193–200.
- ICML-2005-EngelMM #learning #process
- Reinforcement learning with Gaussian processes (YE, SM, RM), pp. 201–208.
- ICML-2005-EspositoS #classification #comparison #monte carlo
- Experimental comparison between bagging and Monte Carlo ensemble classification (RE, LS), pp. 209–216.
- ICML-2005-FinleyJ #clustering
- Supervised clustering with support vector machines (TF, TJ), pp. 217–224.
- ICML-2005-FrohlichWSZ #graph #kernel
- Optimal assignment kernels for attributed molecular graphs (HF, JKW, FS, AZ), pp. 225–232.
- ICML-2005-GeurtsW #modelling
- Closed-form dual perturb and combine for tree-based models (PG, LW), pp. 233–240.
- ICML-2005-GirolamiR #kernel #learning #modelling
- Hierarchic Bayesian models for kernel learning (MG, SR), pp. 241–248.
- ICML-2005-GlocerET #classification #feature model #online
- Online feature selection for pixel classification (KAG, DE, JT), pp. 249–256.
- ICML-2005-GroisW #approach #comprehension #learning
- Learning strategies for story comprehension: a reinforcement learning approach (EG, DCW), pp. 257–264.
- ICML-2005-GuestrinKS #process
- Near-optimal sensor placements in Gaussian processes (CG, AK, APS), pp. 265–272.
- ICML-2005-GuptaG #clustering #hybrid #robust #using
- Robust one-class clustering using hybrid global and local search (GG, JG), pp. 273–280.
- ICML-2005-HeCM #analysis #locality #statistics
- Statistical and computational analysis of locality preserving projection (XH, DC, WM), pp. 281–288.
- ICML-2005-HeinA #estimation
- Intrinsic dimensionality estimation of submanifolds in Rd (MH, JYA), pp. 289–296.
- ICML-2005-HellerG #clustering
- Bayesian hierarchical clustering (KAH, ZG), pp. 297–304.
- ICML-2005-HerbsterPW #graph #learning #online
- Online learning over graphs (MH, MP, LW), pp. 305–312.
- ICML-2005-HillD #adaptation #classification #problem
- Adapting two-class support vector classification methods to many class problems (SIH, AD), pp. 313–320.
- ICML-2005-Ho #concept #data type #detection #framework
- A martingale framework for concept change detection in time-varying data streams (SSH), pp. 321–327.
- ICML-2005-IeWNL #adaptation #multi #recognition #using
- Multi-class protein fold recognition using adaptive codes (EI, JW, WSN, CSL), pp. 329–336.
- ICML-2005-IlghamiMNA #approximate #learning
- Learning approximate preconditions for methods in hierarchical plans (OI, HMA, DSN, DWA), pp. 337–344.
- ICML-2005-IresonCCFKL #information management #machine learning
- Evaluating machine learning for information extraction (NI, FC, MEC, DF, NK, AL), pp. 345–352.
- ICML-2005-JinCS #information retrieval #using
- Learn to weight terms in information retrieval using category information (RJ, JYC, LS), pp. 353–360.
- ICML-2005-JinZ #algorithm #probability #using
- A smoothed boosting algorithm using probabilistic output codes (RJ, JZ), pp. 361–368.
- ICML-2005-JingPR #classification #learning #naive bayes #network #performance
- Efficient discriminative learning of Bayesian network classifier via boosted augmented naive Bayes (YJ, VP, JMR), pp. 369–376.
- ICML-2005-Joachims #metric #multi #performance
- A support vector method for multivariate performance measures (TJ), pp. 377–384.
- ICML-2005-JoachimsH #bound #clustering #correlation #fault
- Error bounds for correlation clustering (TJ, JEH), pp. 385–392.
- ICML-2005-JodogneP #interactive #learning #visual notation
- Interactive learning of mappings from visual percepts to actions (SJ, JHP), pp. 393–400.
- ICML-2005-JonssonB #approach #composition
- A causal approach to hierarchical decomposition of factored MDPs (AJ, AGB), pp. 401–408.
- ICML-2005-KaariainenL #bound #comparison #fault
- A comparison of tight generalization error bounds (MK, JL), pp. 409–416.
- ICML-2005-Keerthi #classification #effectiveness #feature model
- Generalized LARS as an effective feature selection tool for text classification with SVMs (SSK), pp. 417–424.
- ICML-2005-KhoussainovHK #classification
- Ensembles of biased classifiers (RK, AH, NK), pp. 425–432.
- ICML-2005-KoivistoS #aspect-oriented #modelling
- Computational aspects of Bayesian partition models (MK, KS), pp. 433–440.
- ICML-2005-KokD #learning #logic #markov #network
- Learning the structure of Markov logic networks (SK, PMD), pp. 441–448.
- ICML-2005-KolterM #concept #using
- Using additive expert ensembles to cope with concept drift (JZK, MAM), pp. 449–456.
- ICML-2005-KulisBDM #approach #clustering #graph #kernel
- Semi-supervised graph clustering: a kernel approach (BK, SB, ISD, RJM), pp. 457–464.
- ICML-2005-LalSHP #feedback #interface #online
- A brain computer interface with online feedback based on magnetoencephalography (TNL, MS, NJH, HP, TH, JM, MB, WR, TH, NB, BS), pp. 465–472.
- ICML-2005-LangfordZ #classification #learning #performance
- Relating reinforcement learning performance to classification performance (JL, BZ), pp. 473–480.
- ICML-2005-LavioletteM #bound #classification
- PAC-Bayes risk bounds for sample-compressed Gibbs classifiers (FL, MM), pp. 481–488.
- ICML-2005-LeSC #process
- Heteroscedastic Gaussian process regression (QVL, AJS, SC), pp. 489–496.
- ICML-2005-LeiteB #classification #performance #predict
- Predicting relative performance of classifiers from samples (RL, PB), pp. 497–503.
- ICML-2005-LiaoXC #data flow
- Logistic regression with an auxiliary data source (XL, YX, LC), pp. 505–512.
- ICML-2005-LiuXC #graph #predict #using
- Predicting protein folds with structural repeats using a chain graph model (YL, EPX, JGC), pp. 513–520.
- ICML-2005-LongVGTS #integration
- Unsupervised evidence integration (PML, VV, SG, MT, RAS), pp. 521–528.
- ICML-2005-LowdD #estimation #modelling #naive bayes #probability
- Naive Bayes models for probability estimation (DL, PMD), pp. 529–536.
- ICML-2005-MacskassyPR #empirical #evaluation
- ROC confidence bands: an empirical evaluation (SAM, FJP, SR), pp. 537–544.
- ICML-2005-MadsenKE #modelling #using #word
- Modeling word burstiness using the Dirichlet distribution (REM, DK, CE), pp. 545–552.
- ICML-2005-Mahadevan #learning
- Proto-value functions: developmental reinforcement learning (SM), pp. 553–560.
- ICML-2005-MannorPR #classification
- The cross entropy method for classification (SM, DP, RYR), pp. 561–568.
- ICML-2005-McMahanLG #bound #performance #programming #realtime
- Bounded real-time dynamic programming: RTDP with monotone upper bounds and performance guarantees (HBM, ML, GJG), pp. 569–576.
- ICML-2005-Meila #axiom #clustering #perspective
- Comparing clusterings: an axiomatic view (MM), pp. 577–584.
- ICML-2005-MenchettiCF #composition #kernel
- Weighted decomposition kernels (SM, FC, PF), pp. 585–592.
- ICML-2005-MichelsSN #learning #using
- High speed obstacle avoidance using monocular vision and reinforcement learning (JM, AS, AYN), pp. 593–600.
- ICML-2005-NatarajanT #learning #multi
- Dynamic preferences in multi-criteria reinforcement learning (SN, PT), pp. 601–608.
- ICML-2005-NatarajanTADFR #first-order #learning #modelling #probability
- Learning first-order probabilistic models with combining rules (SN, PT, EA, TGD, AF, ACR), pp. 609–616.
- ICML-2005-NguyenH #performance
- An efficient method for simplifying support vector machines (DN, TBH), pp. 617–624.
- ICML-2005-Niculescu-MizilC #learning #predict
- Predicting good probabilities with supervised learning (ANM, RC), pp. 625–632.
- ICML-2005-OntanonP #learning #multi
- Recycling data for multi-agent learning (SO, EP), pp. 633–640.
- ICML-2005-PaiementEBB #embedded #visual notation
- A graphical model for chord progressions embedded in a psychoacoustic space (JFP, DE, SB, DB), pp. 641–648.
- ICML-2005-PalettaFS #recognition #visual notation
- Q-learning of sequential attention for visual object recognition from informative local descriptors (LP, GF, CS), pp. 649–656.
- ICML-2005-PernkopfB #classification #generative #learning #network #parametricity
- Discriminative versus generative parameter and structure learning of Bayesian network classifiers (FP, JAB), pp. 657–664.
- ICML-2005-Pietraszek #analysis #classification #optimisation #using
- Optimizing abstaining classifiers using ROC analysis (TP), pp. 665–672.
- ICML-2005-PoczosL #analysis #independence #using
- Independent subspace analysis using geodesic spanning trees (BP, AL), pp. 673–680.
- ICML-2005-RamakrishnanCKB #approximate #classification
- A model for handling approximate, noisy or incomplete labeling in text classification (GR, KPC, RK, PB), pp. 681–688.
- ICML-2005-RasmussenQ
- Healing the relevance vector machine through augmentation (CER, JQC), pp. 689–696.
- ICML-2005-RayC #comparison #empirical #learning #multi
- Supervised versus multiple instance learning: an empirical comparison (SR, MC), pp. 697–704.
- ICML-2005-RayP
- Generalized skewing for functions with continuous and nominal attributes (SR, DP), pp. 705–712.
- ICML-2005-RennieS #collaboration #matrix #performance #predict
- Fast maximum margin matrix factorization for collaborative prediction (JDMR, NS), pp. 713–719.
- ICML-2005-RohanimaneshM #approach #concurrent #generative #markov #named #process
- Coarticulation: an approach for generating concurrent plans in Markov decision processes (KR, SM), pp. 720–727.
- ICML-2005-RosellHRP #learning #why
- Why skewing works: learning difficult Boolean functions with greedy tree learners (BR, LH, SR, DP), pp. 728–735.
- ICML-2005-RothY #integer #linear #programming #random
- Integer linear programming inference for conditional random fields (DR, WtY), pp. 736–743.
- ICML-2005-RousuSSS #classification #learning #modelling #multi
- Learning hierarchical multi-category text classification models (JR, CS, SS, JST), pp. 744–751.
- ICML-2005-SalojarviPK #algorithm
- Expectation maximization algorithms for conditional likelihoods (JS, KP, SK), pp. 752–759.
- ICML-2005-SajamaO #distance #metric
- Estimating and computing density based distance metrics (S, AO), pp. 760–767.
- ICML-2005-SajamaO05a #modelling #reduction #using
- Supervised dimensionality reduction using mixture models (S, AO), pp. 768–775.
- ICML-2005-ScholkopfSB #machine learning #problem
- Object correspondence as a machine learning problem (BS, FS, VB), pp. 776–783.
- ICML-2005-ShaS #analysis #reduction
- Analysis and extension of spectral methods for nonlinear dimensionality reduction (FS, LKS), pp. 784–791.
- ICML-2005-ShashuaH #statistics
- Non-negative tensor factorization with applications to statistics and computer vision (AS, TH), pp. 792–799.
- ICML-2005-SiddiqiM #learning #performance
- Fast inference and learning in large-state-space HMMs (SMS, AWM), pp. 800–807.
- ICML-2005-SilvaS #identification #learning #modelling
- New d-separation identification results for learning continuous latent variable models (RBdAeS, RS), pp. 808–815.
- ICML-2005-SimsekWB #clustering #graph #identification #learning
- Identifying useful subgoals in reinforcement learning by local graph partitioning (ÖS, APW, AGB), pp. 816–823.
- ICML-2005-SindhwaniNB #learning
- Beyond the point cloud: from transductive to semi-supervised learning (VS, PN, MB), pp. 824–831.
- ICML-2005-SinghPGBB #analysis #learning
- Active learning for sampling in time-series experiments with application to gene expression analysis (RS, NP, DKG, BB, ZBJ), pp. 832–839.
- ICML-2005-SnelsonG #approximate #predict
- Compact approximations to Bayesian predictive distributions (ES, ZG), pp. 840–847.
- ICML-2005-SonnenburgRS #classification #scalability #sequence
- Large scale genomic sequence SVM classifiers (SS, GR, BS), pp. 848–855.
- ICML-2005-StrehlL #analysis #estimation #modelling
- A theoretical analysis of Model-Based Interval Estimation (ALS, MLL), pp. 856–863.
- ICML-2005-SunD #approach #learning
- Explanation-Augmented SVM: an approach to incorporating domain knowledge into SVM learning (QS, GD), pp. 864–871.
- ICML-2005-SunTLW #framework
- Unifying the error-correcting and output-code AdaBoost within the margin framework (YS, ST, JL, DW), pp. 872–879.
- ICML-2005-SzepesvariM #bound #finite
- Finite time bounds for sampling based fitted value iteration (CS, RM), pp. 880–887.
- ICML-2005-TannerS #network
- TD(λ) networks: temporal-difference networks with eligibility traces (BT, RSS), pp. 888–895.
- ICML-2005-TaskarCKG #approach #learning #modelling #predict #scalability
- Learning structured prediction models: a large margin approach (BT, VC, DK, CG), pp. 896–903.
- ICML-2005-ToussaintV #learning #modelling
- Learning discontinuities with products-of-sigmoids for switching between local models (MT, SV), pp. 904–911.
- ICML-2005-TsangKL #problem #scalability
- Core Vector Regression for very large regression problems (IWT, JTK, KTL), pp. 912–919.
- ICML-2005-Tsuda
- Propagating distributions on a hypergraph by dual information regularization (KT), pp. 920–927.
- ICML-2005-VeeramachaneniSA #classification #documentation
- Hierarchical Dirichlet model for document classification (SV, DS, PA), pp. 928–935.
- ICML-2005-WalderCS #modelling #problem
- Implicit surface modelling as an eigenvalue problem (CW, OC, BS), pp. 936–939.
- ICML-2005-WangS #classification #kernel
- New kernels for protein structural motif discovery and function classification (CW, SDS), pp. 940–947.
- ICML-2005-WangWGSC #markov #modelling #random #semantics
- Exploiting syntactic, semantic and lexical regularities in language modeling via directed Markov random fields (SW, SW, RG, DS, LC), pp. 948–955.
- ICML-2005-WangLBS #online #optimisation
- Bayesian sparse sampling for on-line reward optimization (TW, DJL, MHB, DS), pp. 956–963.
- ICML-2005-Wiewiora #learning #predict
- Learning predictive representations from a history (EW), pp. 964–971.
- ICML-2005-WilliamsLXC #classification #using
- Incomplete-data classification using logistic regression (DW, XL, YX, LC), pp. 972–979.
- ICML-2005-WolfeJS #learning #predict
- Learning predictive state representations in dynamical systems without reset (BW, MRJ, SPS), pp. 980–987.
- ICML-2005-WuMR #classification #detection #linear #symmetry
- Linear Asymmetric Classifier for cascade detectors (JW, MDM, JMR), pp. 988–995.
- ICML-2005-WuSB #classification #scalability
- Building Sparse Large Margin Classifiers (MW, BS, GHB), pp. 996–1003.
- ICML-2005-XuTYYK #learning #relational
- Dirichlet enhanced relational learning (ZX, VT, KY, SY, HPK), pp. 1004–1011.
- ICML-2005-YuTS #learning #multi #process
- Learning Gaussian processes from multiple tasks (KY, VT, AS), pp. 1012–1019.
- ICML-2005-ZhangJS #naive bayes #ranking
- Augmenting naive Bayes for ranking (HZ, LJ, JS), pp. 1020–1027.
- ICML-2005-ZhouLZ #clustering #distance #metric
- A new Mallows distance based metric for comparing clusterings (DZ, JL, HZ), pp. 1028–1035.
- ICML-2005-ZhouHS #graph #learning
- Learning from labeled and unlabeled data on a directed graph (DZ, JH, BS), pp. 1036–1043.
- ICML-2005-ZhuNWZM #2d #information management #random #web
- 2D Conditional Random Fields for Web information extraction (JZ, ZN, JRW, BZ, WYM), pp. 1044–1051.
- ICML-2005-ZhuL #graph #induction #learning #modelling #scalability
- Harmonic mixtures: combining mixture models and graph-based methods for inductive and scalable semi-supervised learning (XZ, JDL), pp. 1052–1059.
- ICML-2005-ZienC #scalability
- Large margin non-linear embedding (AZ, JQC), pp. 1060–1067.

45 ×#learning

21 ×#classification

18 ×#modelling

15 ×#using

11 ×#predict

10 ×#clustering

9 ×#multi

9 ×#performance

7 ×#analysis

7 ×#graph

21 ×#classification

18 ×#modelling

15 ×#using

11 ×#predict

10 ×#clustering

9 ×#multi

9 ×#performance

7 ×#analysis

7 ×#graph