## Zoubin Ghahramani

*Proceedings of the 24th International Conference on Machine Learning*

ICML, 2007.

@proceedings{ICML-2007, address = "Corvallis, Oregon, USA", editor = "Zoubin Ghahramani", isbn = "978-1-59593-793-3", publisher = "{ACM}", series = "{ACM International Conference Proceeding Series}", title = "{Proceedings of the 24th International Conference on Machine Learning}", volume = 227, year = 2007, }

### Contents (150 items)

- ICML-2007-AimeurBG #algorithm #clustering #quantum
- Quantum clustering algorithms (EA, GB, SG), pp. 1–8.
- ICML-2007-AgarwalC #graph #learning #random #rank
- Learning random walks to rank nodes in graphs (AA, SC), pp. 9–16.
- ICML-2007-AmitFSU #classification #multi
- Uncovering shared structures in multiclass classification (YA, MF, NS, SU), pp. 17–24.
- ICML-2007-AndoZ #generative #learning
- Two-view feature generation model for semi-supervised learning (RKA, TZ), pp. 25–32.
- ICML-2007-AndrewG #modelling #scalability
- Scalable training of L1-regularized log-linear models (GA, JG), pp. 33–40.
- ICML-2007-AsharafMS #multi
- Multiclass core vector machine (SA, MNM, SKS), pp. 41–48.
- ICML-2007-Azran #algorithm #learning #markov #multi #random
- The rendezvous algorithm: multiclass semi-supervised learning with Markov random walks (AA), pp. 49–56.
- ICML-2007-BabariaNKSBM #crawling #scalability
- Focused crawling with scalable ordinal regression solvers (RB, JSN, SK, KRS, CB, MNM), pp. 57–64.
- ICML-2007-Bar-HillelW #distance #learning #similarity
- Learning distance function by coding similarity (ABH, DW), pp. 65–72.
- ICML-2007-BhattacharyaBC #classification #kernel
- Structural alignment based kernels for protein structure classification (SB, CB, NRC), pp. 73–80.
- ICML-2007-BickelBS #learning
- Discriminative learning for differing training and test distributions (SB, MB, TS), pp. 81–88.
- ICML-2007-BordesBGW #multi
- Solving multiclass support vector machines with LaRank (AB, LB, PG, JW), pp. 89–96.
- ICML-2007-BryanMSS
- Efficiently computing minimax expected-size confidence regions (BB, HBM, CMS, JGS), pp. 97–104.
- ICML-2007-BunescuM #learning #multi
- Multiple instance learning for sparse positive bags (RCB, RJM), pp. 105–112.
- ICML-2007-BusseOB #analysis #clustering #rank
- Cluster analysis of heterogeneous rank data (LMB, PO, JMB), pp. 113–120.
- ICML-2007-CaoSSYC #feature model #kernel
- Feature selection in a kernel space (BC, DS, JTS, QY, ZC), pp. 121–128.
- ICML-2007-CaoQLTL #approach #learning #rank
- Learning to rank: from pairwise approach to listwise approach (ZC, TQ, TYL, MFT, HL), pp. 129–136.
- ICML-2007-CazzantiG #analysis #similarity
- Local similarity discriminant analysis (LC, MRG), pp. 137–144.
- ICML-2007-ChanVL
- Direct convex relaxations of sparse SVM (ABC, NV, GRGL), pp. 145–153.
- ICML-2007-ChenJ #classification #feature model #set
- Minimum reference set based feature selection for small sample classifications (XwC, JCJ), pp. 153–160.
- ICML-2007-ChengV #image #learning
- Learning to compress images and videos (LC, SVNV), pp. 161–168.
- ICML-2007-CortesMR #algorithm #ranking
- Magnitude-preserving ranking algorithms (CC, MM, AR), pp. 169–176.
- ICML-2007-dAspremontBG #analysis #component
- Full regularization path for sparse principal component analysis (Ad, FRB, LEG), pp. 177–184.
- ICML-2007-DaiY #kernel
- Kernel selection forl semi-supervised kernel machines (GD, DYY), pp. 185–192.
- ICML-2007-DaiYXY #learning
- Boosting for transfer learning (WD, QY, GRX, YY), pp. 193–200.
- ICML-2007-DavidsonR #clustering #constraints
- Intractability and clustering with constraints (ID, SSR), pp. 201–208.
- ICML-2007-DavisKJSD #learning #metric
- Information-theoretic metric learning (JVD, BK, PJ, SS, ISD), pp. 209–216.
- ICML-2007-DavisCRP #approach #predict #process
- An integrated approach to feature invention and model construction for drug activity prediction (JD, VSC, SR, DP), pp. 217–224.
- ICML-2007-DelageM #markov #nondeterminism #optimisation #performance #process
- Percentile optimization in uncertain Markov decision processes with application to efficient exploration (ED, SM), pp. 225–232.
- ICML-2007-DietzBS #predict
- Unsupervised prediction of citation influences (LD, SB, TS), pp. 233–240.
- ICML-2007-DollarRB #algorithm #analysis #learning
- Non-isometric manifold learning: analysis and an algorithm (PD, VR, SJB), pp. 241–248.
- ICML-2007-DudikBS #estimation
- Hierarchical maximum entropy density estimation (MD, DMB, RES), pp. 249–256.
- ICML-2007-EspositoR #algorithm #classification #evaluation #named #performance
- CarpeDiem: an algorithm for the fast evaluation of SSL classifiers (RE, DPR), pp. 257–264.
- ICML-2007-FarahmandSA #adaptation #estimation
- Manifold-adaptive dimension estimation (AMF, CS, JYA), pp. 265–272.
- ICML-2007-GellyS #online
- Combining online and offline knowledge in UCT (SG, DS), pp. 273–280.
- ICML-2007-GerberTW #reduction #robust #using
- Robust non-linear dimensionality reduction using successive 1-dimensional Laplacian Eigenmaps (SG, TT, RTW), pp. 281–288.
- ICML-2007-GeurtsWd #kernel
- Gradient boosting for kernelized output spaces (PG, LW, FdB), pp. 289–296.
- ICML-2007-GhavamzadehE #algorithm
- Bayesian actor-critic algorithms (MG, YE), pp. 297–304.
- ICML-2007-GlobersonKCC #algorithm #predict
- Exponentiated gradient algorithms for log-linear structured prediction (AG, TK, XC, MC), pp. 305–312.
- ICML-2007-GriraH #clustering #heuristic
- Best of both: a hybridized centroid-medoid clustering heuristic (NG, MEH), pp. 313–320.
- ICML-2007-GuoHFX #approach #modelling #network
- Recovering temporally rewiring networks: a model-based approach (FG, SH, WF, EPX), pp. 321–328.
- ICML-2007-GuptaDS #clique #performance
- Efficient inference with cardinality-based clique potentials (RG, AAD, SS), pp. 329–336.
- ICML-2007-HeraultG #classification #probability
- Sparse probabilistic classifiers (RH, YG), pp. 337–344.
- ICML-2007-HaiderBS #clustering #detection #email #streaming
- Supervised clustering of streaming data for email batch detection (PH, UB, TS), pp. 345–352.
- ICML-2007-Hanneke #bound #complexity #learning
- A bound on the label complexity of agnostic active learning (SH), pp. 353–360.
- ICML-2007-HoiJL #constraints #kernel #learning #matrix #parametricity
- Learning nonparametric kernel matrices from pairwise constraints (SCHH, RJ, MRL), pp. 361–368.
- ICML-2007-Jaeger #learning #network #parametricity #relational
- Parameter learning for relational Bayesian networks (MJ), pp. 369–376.
- ICML-2007-JiC #optimisation
- Bayesian compressive sensing and projection optimization (SJ, LC), pp. 377–384.
- ICML-2007-JohnsM #approximate #graph
- Constructing basis functions from directed graphs for value function approximation (JJ, SM), pp. 385–392.
- ICML-2007-KerstingPPB #process
- Most likely heteroscedastic Gaussian process regression (KK, CP, PP, WB), pp. 393–400.
- ICML-2007-KimC #algorithm #geometry #message passing
- Neighbor search with global geometry: a minimax message passing algorithm (KHK, SC), pp. 401–408.
- ICML-2007-KimP #learning #recursion
- A recursive method for discriminative mixture learning (MK, VP), pp. 409–416.
- ICML-2007-KirshnerS #infinity
- Infinite mixtures of trees (SK, PS), pp. 417–423.
- ICML-2007-KlamiK #component
- Local dependent components (AK, SK), pp. 425–432.
- ICML-2007-KokD #statistics
- Statistical predicate invention (SK, PMD), pp. 433–440.
- ICML-2007-KramerB #kernel #performance
- Kernelizing PLS, degrees of freedom, and efficient model selection (NK, MLB), pp. 441–448.
- ICML-2007-KrauseG #approach #learning #process
- Nonmyopic active learning of Gaussian processes: an exploration-exploitation approach (AK, CG), pp. 449–456.
- ICML-2007-KropotovV #learning #on the
- On one method of non-diagonal regularization in sparse Bayesian learning (DK, DV), pp. 457–464.
- ICML-2007-KuzminW #kernel #matrix #online
- Online kernel PCA with entropic matrix updates (DK, MKW), pp. 465–472.
- ICML-2007-LarochelleECBB #architecture #empirical #evaluation #problem
- An empirical evaluation of deep architectures on problems with many factors of variation (HL, DE, ACC, JB, YB), pp. 473–480.
- ICML-2007-LawrenceM #modelling #process
- Hierarchical Gaussian process latent variable models (NDL, AJM), pp. 481–488.
- ICML-2007-LeeCVK #learning #multi
- Learning a meta-level prior for feature relevance from multiple related tasks (SIL, VC, DV, DK), pp. 489–496.
- ICML-2007-LeskovecF #graph #modelling #multi #scalability #using
- Scalable modeling of real graphs using Kronecker multiplication (JL, CF), pp. 497–504.
- ICML-2007-LiCFX #clustering
- Support cluster machine (BL, MC, JF, XX), pp. 505–512.
- ICML-2007-LiYW #distance #framework #learning #metric #reduction
- A transductive framework of distance metric learning by spectral dimensionality reduction (FL, JY, JW), pp. 513–520.
- ICML-2007-DingL #adaptation #analysis #clustering #reduction #using
- Adaptive dimension reduction using discriminant analysis and K-means clustering (CHQD, TL), pp. 521–528.
- ICML-2007-LiLL #learning #scalability
- Large-scale RLSC learning without agony (WL, KHL, KSL), pp. 529–536.
- ICML-2007-LiCLW #novel #orthogonal
- A novel orthogonal NMF-based belief compression for POMDPs (XL, WKWC, JL, ZW), pp. 537–544.
- ICML-2007-LiangJT #modelling
- A permutation-augmented sampler for DP mixture models (PL, MIJ, BT), pp. 545–552.
- ICML-2007-LiaoLC #classification #semistructured data
- Quadratically gated mixture of experts for incomplete data classification (XL, HL, LC), pp. 553–560.
- ICML-2007-LinWK #scalability #trust
- Trust region Newton methods for large-scale logistic regression (CJL, RCW, SSK), pp. 561–568.
- ICML-2007-LongZWY #clustering #relational #symmetry
- Relational clustering by symmetric convex coding (BL, Z(Z, XW, PSY), pp. 569–576.
- ICML-2007-MaLTK #analysis #correlation #similarity
- Discriminant analysis in correlation similarity measure space (YM, SL, ET, MK), pp. 577–584.
- ICML-2007-Mahadevan #3d #adaptation #learning #multi #using
- Adaptive mesh compression in 3D computer graphics using multiscale manifold learning (SM), pp. 585–592.
- ICML-2007-MannM #learning #robust #scalability
- Simple, robust, scalable semi-supervised learning via expectation regularization (GSM, AM), pp. 593–600.
- ICML-2007-Marthi #automation #composition
- Automatic shaping and decomposition of reward functions (BM), pp. 601–608.
- ICML-2007-Masnadi-ShiraziV #symmetry
- Asymmetric boosting (HMS, NV), pp. 609–619.
- ICML-2007-McNeillV #generative #linear #modelling #probability
- Linear and nonlinear generative probabilistic class models for shape contours (GM, SV), pp. 617–624.
- ICML-2007-MihalkovaM #bottom-up #learning #logic #markov #network
- Bottom-up learning of Markov logic network structure (LM, RJM), pp. 625–632.
- ICML-2007-MimnoLM #topic
- Mixtures of hierarchical topics with Pachinko allocation (DMM, WL, AM), pp. 633–640.
- ICML-2007-MnihH #modelling #statistics #visual notation
- Three new graphical models for statistical language modelling (AM, GEH), pp. 641–648.
- ICML-2007-MoschittiZ #effectiveness #kernel #learning #performance #relational
- Fast and effective kernels for relational learning from texts (AM, FMZ), pp. 649–656.
- ICML-2007-MosciRV #reduction
- Dimensionality reduction and generalization (SM, LR, AV), pp. 657–664.
- ICML-2007-MylonakisSH #estimation #modelling
- Unsupervised estimation for noisy-channel models (MM, KS, RH), pp. 665–672.
- ICML-2007-NelsonC #clustering #constraints #modelling #probability
- Revisiting probabilistic models for clustering with pair-wise constraints (BN, IC), pp. 673–680.
- ICML-2007-NguyenG #algorithm #sequence
- Comparisons of sequence labeling algorithms and extensions (NN, YG), pp. 681–688.
- ICML-2007-NiCD #learning #multi #process
- Multi-task learning for sequential data via iHMMs and the nested Dirichlet process (KN, LC, DBD), pp. 689–696.
- ICML-2007-NilssonSJ #kernel #reduction #using
- Regression on manifolds using kernel dimension reduction (JN, FS, MIJ), pp. 697–704.
- ICML-2007-OsentoskiM #learning
- Learning state-action basis functions for hierarchical MDPs (SO, SM), pp. 705–712.
- ICML-2007-YoganandaMG #linear #performance
- A fast linear separability test by projection of positive points on subspaces (APY, MNM, LG), pp. 713–720.
- ICML-2007-PandeyCA #multi #problem
- Multi-armed bandit problems with dependent arms (SP, DC, DA), pp. 721–728.
- ICML-2007-ParkerFT #learning #performance #query #retrieval
- Learning for efficient retrieval of structured data with noisy queries (CP, AF, PT), pp. 729–736.
- ICML-2007-ParrPLL #approximate #generative
- Analyzing feature generation for value-function approximation (RP, CPW, LL, MLL), pp. 737–744.
- ICML-2007-PetersS #learning
- Reinforcement learning by reward-weighted regression for operational space control (JP, SS), pp. 745–750.
- ICML-2007-PhuaF #approximate #learning #linear
- Tracking value function dynamics to improve reinforcement learning with piecewise linear function approximation (CWP, RF), pp. 751–758.
- ICML-2007-RainaBLPN #learning #self
- Self-taught learning: transfer learning from unlabeled data (RR, AB, HL, BP, AYN), pp. 759–766.
- ICML-2007-RakhlinAB #online #similarity
- Online discovery of similarity mappings (AR, JA, PLB), pp. 767–774.
- ICML-2007-RakotomamonjyBCG #kernel #learning #multi #performance
- More efficiency in multiple kernel learning (AR, FRB, SC, YG), pp. 775–782.
- ICML-2007-RattiganMJ #clustering #graph #network
- Graph clustering with network structure indices (MJR, MEM, DJ), pp. 783–790.
- ICML-2007-SalakhutdinovMH #collaboration #strict
- Restricted Boltzmann machines for collaborative filtering (RS, AM, GEH), pp. 791–798.
- ICML-2007-Shah #bound
- Sample compression bounds for decision trees (MS), pp. 799–806.
- ICML-2007-Shalev-ShwartzSS #named
- Pegasos: Primal Estimated sub-GrAdient SOlver for SVM (SSS, YS, NS), pp. 807–814.
- ICML-2007-SongSGB #clustering #dependence
- A dependence maximization view of clustering (LS, AJS, AG, KMB), pp. 815–822.
- ICML-2007-SongSGBB #dependence #estimation #feature model
- Supervised feature selection via dependence estimation (LS, AJS, AG, KMB, JB), pp. 823–830.
- ICML-2007-SriperumbudurTL #programming
- Sparse eigen methods by D.C. programming (BKS, DAT, GRGL), pp. 831–838.
- ICML-2007-SternHG #game studies #learning
- Learning to solve game trees (DHS, RH, TG), pp. 839–846.
- ICML-2007-SunKR #fault #metric #robust
- Robust mixtures in the presence of measurement errors (JS, AK, SR), pp. 847–854.
- ICML-2007-SunJSF #algorithm #kernel #learning
- A kernel-based causal learning algorithm (XS, DJ, BS, KF), pp. 855–862.
- ICML-2007-SuttonM #performance #pseudo #random
- Piecewise pseudolikelihood for efficient training of conditional random fields (CAS, AM), pp. 863–870.
- ICML-2007-SuttonKS #on the
- On the role of tracking in stationary environments (RSS, AK, DS), pp. 871–878.
- ICML-2007-TaylorS #learning
- Cross-domain transfer for reinforcement learning (MET, PS), pp. 879–886.
- ICML-2007-TitovH #incremental #network #predict
- Incremental Bayesian networks for structure prediction (IT, JH), pp. 887–894.
- ICML-2007-TomiokaA #matrix
- Classifying matrices with a spectral regularization (RT, KA), pp. 895–902.
- ICML-2007-TsampoukaS #algorithm #approximate
- Approximate maximum margin algorithms with rules controlled by the number of mistakes (PT, JST), pp. 903–910.
- ICML-2007-TsangKK
- Simpler core vector machines with enclosing balls (IWT, AK, JTK), pp. 911–918.
- ICML-2007-Tsuda #graph
- Entire regularization paths for graph data (KT), pp. 919–926.
- ICML-2007-UrtasunD #classification #process
- Discriminative Gaussian process latent variable model for classification (RU, TD), pp. 927–934.
- ICML-2007-HulseKN #learning
- Experimental perspectives on learning from imbalanced data (JVH, TMK, AN), pp. 935–942.
- ICML-2007-WachmanK #kernel #learning #order
- Learning from interpretations: a rooted kernel for ordered hypergraphs (GW, RK), pp. 943–950.
- ICML-2007-WangYL #algorithm #kernel
- A kernel path algorithm for support vector machines (GW, DYY, FHL), pp. 951–958.
- ICML-2007-WangZQ #learning #metric #towards
- Dirichlet aggregation: unsupervised learning towards an optimal metric for proportional data (HYW, HZ, HQ), pp. 959–966.
- ICML-2007-WangYHLT
- Transductive regression piloted by inter-manifold relations (HW, SY, TSH, JL, XT), pp. 967–974.
- ICML-2007-WangFH #modelling #multi #process
- Multifactor Gaussian process models for style-content separation (JMW, DJF, AH), pp. 975–982.
- ICML-2007-WangZZ #array #classification #hybrid
- Hybrid huberized support vector machines for microarray classification (LW, JZ, HZ), pp. 983–990.
- ICML-2007-WangYF #difference #learning #on the
- On learning with dissimilarity functions (LW, CY, JF), pp. 991–998.
- ICML-2007-Warmuth
- Winnowing subspaces (MKW), pp. 999–1006.
- ICML-2007-Werner #algorithm #consistency #question #what
- What is decreased by the max-sum arc consistency algorithm? (TW), pp. 1007–1014.
- ICML-2007-WilsonFRT #approach #learning #multi
- Multi-task reinforcement learning: a hierarchical Bayesian approach (AW, AF, SR, PT), pp. 1015–1022.
- ICML-2007-WipfN #using
- Beamforming using the relevance vector machine (DPW, SSN), pp. 1023–1030.
- ICML-2007-WoznicaKH #learning
- Learning to combine distances for complex representations (AW, AK, MH), pp. 1031–1038.
- ICML-2007-WuYYS #learning
- Local learning projections (MW, KY, SY, BS), pp. 1039–1046.
- ICML-2007-XuF #learning #linear #on the #ranking
- On learning linear ranking functions for beam search (YX, AF), pp. 1047–1054.
- ICML-2007-XuanM #dependence #modelling #multi
- Modeling changing dependency structure in multivariate time series (XX, KPM), pp. 1055–1062.
- ICML-2007-XueDC #flexibility #learning #matrix #multi #process
- The matrix stick-breaking process for flexible multi-task learning (YX, DBD, LC), pp. 1063–1070.
- ICML-2007-Yairi #locality #reduction
- Map building without localization by dimensionality reduction techniques (TY), pp. 1071–1078.
- ICML-2007-YamazakiKWSM #fault
- Asymptotic Bayesian generalization error when training and test distributions are different (KY, MK, SW, MS, KRM), pp. 1079–1086.
- ICML-2007-Ye #analysis #linear
- Least squares linear discriminant analysis (JY), pp. 1087–1093.
- ICML-2007-YeCJ #kernel #learning #parametricity #programming
- Discriminant kernel and regularization parameter learning via semidefinite programming (JY, JC, SJ), pp. 1095–1102.
- ICML-2007-YuTY #learning #multi #robust
- Robust multi-task learning with t-processes (SY, VT, KY), pp. 1103–1110.
- ICML-2007-ZhangY #classification #consistency #constraints #on the
- On the value of pairwise constraints in classification and consistency (JZ, RY), pp. 1111–1118.
- ICML-2007-ZhangTK #clustering
- Maximum margin clustering made practical (KZ, IWT, JTK), pp. 1119–1126.
- ICML-2007-ZhangC #analysis #component #independence
- Nonlinear independent component analysis with minimal nonlinear distortion (KZ, LC), pp. 1127–1134.
- ICML-2007-ZhangXSGL #classification #metric
- Optimal dimensionality of metric space for classification (WZ, XX, ZS, YFG, HL), pp. 1135–1142.
- ICML-2007-ZhangAV #learning #multi #random
- Conditional random fields for multi-agent reinforcement learning (XZ, DA, SVNV), pp. 1143–1150.
- ICML-2007-ZhaoL #feature model #learning
- Spectral feature selection for supervised and unsupervised learning (ZZ, HL), pp. 1151–1157.
- ICML-2007-ZhouB #clustering #learning #multi
- Spectral clustering and transductive learning with multiple views (DZ, CJCB), pp. 1159–1166.
- ICML-2007-ZhouX #learning #multi #on the
- On the relation between multi-instance learning and semi-supervised learning (ZHZ, JMX), pp. 1167–1174.
- ICML-2007-ZhuNZW #markov #random #web
- Dynamic hierarchical Markov random fields and their application to web data extraction (JZ, ZN, BZ, JRW), pp. 1175–1182.
- ICML-2007-ZienBS
- Transductive support vector machines for structured variables (AZ, UB, TS), pp. 1183–1190.
- ICML-2007-ZienO #kernel #learning #multi
- Multiclass multiple kernel learning (AZ, CSO), pp. 1191–1198.

50 ×#learning

20 ×#multi

15 ×#kernel

13 ×#algorithm

13 ×#clustering

11 ×#modelling

10 ×#classification

9 ×#performance

9 ×#process

8 ×#analysis

20 ×#multi

15 ×#kernel

13 ×#algorithm

13 ×#clustering

11 ×#modelling

10 ×#classification

9 ×#performance

9 ×#process

8 ×#analysis