Machine Learning, Optimization, and Data Science, 5 Conference, LOD 2019

Preface......Page 6
Organization......Page 8
Contents......Page 21
 1 Introduction......Page 27
 2 Related Work......Page 28
  3.2 Testing......Page 29
  3.3 Larger Models......Page 30
 4 Results......Page 31
  5.1 ``Good\'\' Paths......Page 34
  6.2 Partitions......Page 36
 7 Conclusion......Page 37
 References......Page 38
 1 Introduction......Page 39
 2 Deep Learning vs Neural Network......Page 41
  3.1 Volunteers and Equipment......Page 42
  3.2 Data......Page 43
 4 Deep Neural Network Results......Page 44
 5 Neural Network Results......Page 45
 6 Conclusion......Page 47
 References......Page 48
 1 Introduction......Page 50
 2 Case Study......Page 51
 3 Dataset Visualization......Page 52
 4 Preliminary Analysis......Page 54
  5.1 Independent Time Correlations Analysis......Page 56
  5.2 Time Correlations Analysis......Page 57
  5.3 Future Works......Page 59
 References......Page 60
 1 Introduction......Page 62
  2.1 Feature Analysis......Page 63
  2.2 Salient Information Visualization......Page 64
 3 Results and Discussion......Page 65
 References......Page 69
 1 Introduction......Page 71
 2 Motivation......Page 72
  3.1 Weight Reinitialization (WMM-WR)......Page 73
  3.2 Weight Shuffling (WMM-WS)......Page 74
 4 Experimental Setup......Page 76
 5 Results and Discussion......Page 77
 6 Conclusions......Page 81
 References......Page 82
 1 Introduction......Page 84
 2 Deep Learning for Computer Vision......Page 85
  3.1 The Black Box Problem......Page 86
  3.3 XAI Algorithms......Page 87
  4.1 Visual Comparison......Page 88
  4.2 Metrics......Page 89
  4.3 Quantitative Comparison......Page 91
 5 Linking Explanations with Structured Knowledge......Page 92
 6 Discussion......Page 94
 References......Page 95
 1 Introduction......Page 97
 2 A Gaussian Mixture Model......Page 99
 3 The Variance Law......Page 102
  3.1 Experimental Validation......Page 104
 4 Moments of the GMM Distribution......Page 106
 References......Page 107
 1 Introduction......Page 109
  2.1 Random Survival Forest......Page 111
  2.2 Random Forest Classifier for Histogram Data......Page 112
  3.1 Using Multiple Bins Together for Evaluating Node Split......Page 113
 4 Empirical Evaluation......Page 114
  4.2 Experiment Setup and Results......Page 115
 5 Conclusion and Future Work......Page 118
 References......Page 119
 1 Introduction......Page 121
  3.1 The Spectral Estimator......Page 122
  3.2 Theoretical Guarantees......Page 123
 4 Conclusion and Future Works......Page 125
  A.2  Bounds on......Page 126
  A.3   Perturbation of the Linear Term......Page 128
 References......Page 130
 1 Introduction......Page 132
 2 Literature Review......Page 134
 4 The Label-Independent Algorithm (LIA)......Page 135
  4.4 Time Complexity Analysis......Page 136
 6 Empirical Study......Page 138
  7.1 Performance Evaluation for Different Iterations of LIA......Page 139
  7.2 Performance Evaluation for the First Iteration of LIA......Page 140
 8 Conclusions and Future Work......Page 141
 References......Page 142
 1 Introduction......Page 144
 2 Previous Work......Page 145
 3 Metrics......Page 146
 4 Experimental Procedure......Page 149
 5 Results and Discussion......Page 151
 6 Conclusion......Page 153
 References......Page 154
 1 Introduction......Page 156
 3 Problem Definition......Page 158
  4.1 Obtaining Base Mapping......Page 159
  4.2 Higher Order Calculation......Page 161
 5 Experiments and Results......Page 162
  5.2 Importance of the Order Cell......Page 165
 References......Page 167
 1 Introduction......Page 169
  2.1 Financial Markets......Page 170
  3.1 Environment......Page 171
  3.2 Learning......Page 172
  3.3 Reward Function......Page 173
  4.2 Results and Discussion......Page 175
 5 Conclusion......Page 178
 References......Page 179
 1 Introduction......Page 180
 2 State Graph......Page 182
 3 Beam Search Framework......Page 183
  3.1 Functions for Evaluating Nodes......Page 184
  3.2 A Heuristic Estimation of the Expected Length of an LCS......Page 185
  3.3 Expressing Existing Approaches in Terms of Our Framework......Page 186
 4 Experimental Evaluation......Page 187
 5 Conclusions and Future Work......Page 191
 References......Page 192
 1 Introduction......Page 194
 2 Hybrid Adaptive Particle Swarm Optimization......Page 195
 3 Results and Discussion......Page 197
 References......Page 204
 1 Human and Machine Cognition......Page 206
 2 Probability Measure of Regular Relations Recognition......Page 207
  2.2 Classification Task......Page 209
 3 Experimental Testing of the Measure of Regularities Recognition......Page 211
   3.1.2 Model Size......Page 213
 4 The Results of the Experiments......Page 214
 5 Conclusions......Page 215
 References......Page 216
 1 Introduction......Page 218
 2 Related Work......Page 219
 3 Preliminary Concepts......Page 220
 5 Generating the Complete Set of CARs......Page 221
  5.1 Finding the Frequent Itemsets Using Candidate Generation......Page 222
 6 Definition of Our Proposed Method......Page 223
 7 Experimental Evaluation......Page 226
 8 Conclusion and Future Work......Page 228
 References......Page 229
 1 Introduction......Page 231
 2 A Compound Hierarchical Model......Page 234
 3 Posterior Sampling......Page 236
  3.1 Inference via Markov Chain Monte Carlo Methods......Page 237
  3.2 Empirical Evaluation of Samples......Page 239
 4 Experimental Analyses......Page 240
 References......Page 243
 1 Introduction......Page 245
 2 Problem Formulation......Page 247
 3 The Robust Tensor Power Method......Page 248
 4 Proposed Method......Page 250
  5.1 Data Generation......Page 251
  5.2 Performance......Page 253
 6 Conclusion......Page 255
 References......Page 256
 1 Introduction......Page 258
 2 The Machine Learning Argument......Page 261
  2.1 Example......Page 265
 3 Discussion......Page 266
 4 Conclusion......Page 268
 References......Page 269
 1 Introduction......Page 270
 2 Spatial Rule Mining......Page 271
  3.1 Geoprocessing......Page 272
 4 Results and Discussion......Page 274
 5 Conclusions......Page 283
 References......Page 284
 1 Introduction......Page 285
 2 Background and Preliminaries......Page 286
 3 k-Richness Properties of k-Means Variants......Page 287
 4 k-Means-Random Not Probabilistically Rich......Page 291
 6 Discussion......Page 295
 References......Page 296
 1 Introduction......Page 298
  3.1 K-Medoids Clustering and Determining the Number of Clusters......Page 299
  3.2 COLD Algorithm......Page 300
  4.1 Artificial Examples......Page 302
  4.2 Training Procedure......Page 304
 5 Results......Page 305
 6 Conclusion......Page 307
 References......Page 308
 1 Introduction......Page 310
 2 Constructing Universal Approximate Centers......Page 312
  3.1 k-Clustering......Page 317
  3.3 Examples......Page 318
  3.4 Reduction to Discrete Euclidean Problems......Page 320
 References......Page 321
 1 Introduction......Page 322
  2.1 k-MXT Algorithm......Page 323
  2.2 k-MXT-W Algorithm......Page 324
  3.1 Methodology......Page 325
  3.2 The Blobs Type Data......Page 326
  3.3 The Circles Data......Page 328
  3.4 The Moons Data Type......Page 329
 4 Comparison of the Results of the k-MXT and k-MXT-W Algorithms on Real Data Sets......Page 330
 References......Page 332
 1 Introduction......Page 334
 2 CoPASample Framework......Page 335
  2.2 Total Covariance Preservation......Page 336
 3 Experimental Results......Page 340
 4 Conclusion......Page 344
 References......Page 345
 1 Introduction......Page 347
  2.1 Algorithm Selection......Page 349
 3 Method......Page 350
 4 Computational Results......Page 352
 5 Conclusion......Page 357
 References......Page 358
 1 Introduction......Page 361
 2 Framework......Page 363
  2.1 Using Statistical Learning to Bridge High and Low Fidelity Models......Page 364
  2.2 Algorithm Details and Behavior......Page 366
 3 Algorithm Analysis......Page 370
 References......Page 371
 1 Introduction......Page 373
 2 Differential Evolution for BLP......Page 375
 3 Review of Methods Using Surrogate Models in BLP......Page 376
  4.1 Surrogate Models......Page 377
  4.2 Description of the Method......Page 379
 5 Computational Experiments......Page 380
  5.1 Results and Discussions......Page 381
 References......Page 383
 1 Introduction......Page 386
  2.2 Polarity-Weighted Multi-hot Encoding......Page 388
 3 The BowTie Feedforward Neural Network......Page 389
 4 Training and Transfer Scenarios......Page 392
 5 Discussion and Next Steps......Page 395
 References......Page 397
 1 Introduction......Page 398
 2 Text Classification......Page 399
  2.2 Preprocessing and Feature Selection......Page 400
  2.3 Methods of Text Classification......Page 401
 4 Data Cleaning and Transformation......Page 402
  4.1 Model Development......Page 404
 5 Experimental Results......Page 405
 6 Conclusion......Page 407
 References......Page 408
 1 Introduction......Page 410
 2 A Natural Formulation for the Hotspot Problem......Page 412
 3 The path&cycle Formulation and a Re-optimization Framework......Page 415
 4 Computational Results......Page 418
 5 Conclusions......Page 420
 References......Page 421
 1 Introduction......Page 422
 2 Cartesian Genetic Programming......Page 423
  3.1 Guided Active Mutation......Page 425
 4 Computational Experiments......Page 428
 5 Concluding Remarks and Future Work......Page 432
 References......Page 433
 1 Introduction......Page 435
  2.2 Optimal Route Reflector Topology Design (ORRTD)......Page 437
 3 Border Routers and Route Reflectors......Page 439
  3.1 Extended ORRTD......Page 440
 4 Experimental Results......Page 442
 5 Conclusion......Page 445
 References......Page 446
 1 Motivation......Page 448
  2.1 GFP......Page 449
  2.3 SWGFP......Page 450
  3.2 Complexity Analysis......Page 451
  4.1 GRASP-GFP......Page 453
  4.2 GRASP-IS......Page 454
 6 Conclusions......Page 457
 References......Page 458
 1 Motivation......Page 460
 2 Definitions and Terminology......Page 461
  3.2 Elements......Page 462
  3.3 Finding Sparse UMRG......Page 463
  3.4 Finding Almost-Complete UMRG......Page 464
 4 The Graph H=C7 is UMRG......Page 465
 References......Page 469
 1 Introduction......Page 471
 2 Related Work......Page 473
 3 Merging Quality Estimation......Page 474
 4 Binary Classifiers......Page 476
 5 Computational Study......Page 479
 References......Page 482
 1 Introduction......Page 484
  2.1 DAGs and Linear Structural Equation Models......Page 486
  2.2 Completed Partially DAG and the PC Algorithm......Page 487
  2.3 Linear SEMs for Two-Set Partitions with Partial Ordering Information......Page 489
  2.4 Extension to K-set Partitions......Page 492
 3 Performance Evaluation......Page 493
 References......Page 496
 1 Introduction......Page 498
  2.1 Dictionary Learning......Page 500
  2.2 Atom Generation from Ancestral Atoms......Page 501
  2.3 Optimization of the Ancestors......Page 503
 3 Experimental Results......Page 504
  3.1 Experiments with Artificial Signals......Page 505
  3.2 Experiments with Natural Images......Page 507
 4 Conclusion and Future Work......Page 508
 References......Page 509
 1 Introduction......Page 510
  2.1 Model Specification......Page 512
  2.2 Model Inference of RKBMF......Page 514
  2.3 Construction of the Kernel......Page 517
  3.2 Video Example......Page 518
 References......Page 520
 1 Introduction......Page 522
 2 Related Work......Page 523
 3 Review of Support Vector Machine......Page 524
 4 Experiments on Parameters......Page 525
 5 Key Findings......Page 526
  5.2 A Fast Estimation of an Appropriate......Page 527
  5.3 Regulator C......Page 529
  6.1 Devised Algorithm......Page 530
  6.2 An Alterable Tolerance......Page 531
 7 Conclusion and Future Scope......Page 532
 References......Page 533
 1 Introduction......Page 534
  2.1 Predictive Accuracy Estimation......Page 536
  2.2 Estimation of Three Types of Overfitting......Page 537
  3.1 Predictive Accuracy and Overfit Results......Page 538
  3.2 Selected Models......Page 542
  3.3 Statistical Analysis......Page 544
 References......Page 545
 1 Introduction......Page 547
 2 Related Work......Page 548
  3.2 L-SHADE......Page 549
  3.3 Adaptive Random Search......Page 550
  4.2 The Search Space......Page 551
  4.3 Datasets......Page 552
  5.1 Measuring Performance......Page 553
  5.2 Comparison......Page 554
 6 Conclusion......Page 555
 References......Page 556
 1 Introduction......Page 557
 2 Background......Page 559
 3 Conditional Generalization Error and Its Large-Sample Approximation......Page 560
 4 Optimal Trade-Off Between Sample Size and Precision of Supervision for the Fixed Effects Panel Data Model Under the Large-Sample Approximation......Page 562
 5 Discussion......Page 564
 References......Page 568
 1 Introduction......Page 569
  2.1 Description......Page 570
  2.2 Data Preprocessing......Page 571
  3.2 Predicting Future Apartments\' Prices......Page 572
 4 Experiments......Page 573
  4.2 ARIMA Model Applied to Base Time Series......Page 574
  4.4 LSTM Model Applied to All Time Series......Page 575
 5 Conclusion......Page 576
 References......Page 577
 1 Introduction......Page 579
  2.1 System Overview and Problem Definition......Page 581
  2.2 Dataset......Page 582
  2.3 Models......Page 583
  2.4 Performance Measures......Page 585
 3 Results......Page 586
 4 Conclusion and Ongoing Work......Page 589
 References......Page 590
 1 Motivation......Page 592
 2 Heterogeneity of Data Quality......Page 594
 3 Forecasting Models......Page 595
  3.1 A Power Forecasting Model from Weather Data......Page 596
  3.2 A Mapping Model Learning from Reference Power Forecasts......Page 598
 4 Evaluation......Page 599
 References......Page 603
Trading-off Data Fit and Complexity in Training Gaussian Processes with Multiple Kernels......Page 605
 1 Introduction......Page 606
 2 Gaussian Processes for Regression......Page 607
  2.1 Kernels......Page 608
 3 Multi-objective Training of Gaussian Processes......Page 609
  3.1 Training with Multiple Kernels......Page 611
 4 Numerical Experiments......Page 612
 References......Page 615
 1 Introduction......Page 618
 3 Cartesian Genetic Programming......Page 620
  3.1 Objective Functions......Page 621
 5 The Proposed Method......Page 622
 6 Computational Experiments......Page 624
  6.2 Scenario 2 – Approximate Design of CLCs from Scratch......Page 625
  6.3 Scenario 3 – Approximate CLCs Using Conventional Models......Page 627
 7 Concluding Remarks and Future Work......Page 629
 References......Page 630
 1 Introduction......Page 631
 2 Pareto Optimality for MultiTask Learning......Page 632
 3 Experimental Protocol......Page 633
  3.2 Evolutionary Strategy......Page 634
 4 Results......Page 635
 5 Discussions and Conclusions......Page 642
 References......Page 643
 1 Introduction......Page 645
 2 Materials and Methods......Page 647
 3 Results......Page 651
 4 Conclusion......Page 654
 References......Page 655
On the Role of Hub and Orphan Genes in the Diagnosis of Breast Invasive Carcinoma......Page 657
 1 Introduction......Page 658
  2.1 Classification......Page 659
  2.2 Data......Page 661
 3 Results and Discussion......Page 662
 References......Page 667
 1 Introduction......Page 669
 2 Stochastic Surgery Scheduling with Ward Restrictions......Page 671
 3 MIP Formulation......Page 672
 5 Optimization and Learning ORDS......Page 674
  6.1 Instance Generation......Page 675
  6.2 Results......Page 677
 References......Page 679
  1.1 The Problem of Principle Component Dimension Selection......Page 681
  1.2 The Principles of MDL and NML......Page 682
  1.4 Outline......Page 683
 2 NML Reduction via Elimination of Optimization Parameters......Page 684
 3 Reducing PCA NML to Linear Regression NML......Page 687
  A.1  Number of Quantized Unitary Matrices......Page 689
 References......Page 691
 2 Previous Work......Page 693
 3 Learning Algorithm......Page 694
 4 Experimental Setup......Page 696
 5 Results......Page 697
 References......Page 700
 1 Introduction......Page 702
  2.1 Similarity Learning as Pairwise Bipartite Ranking......Page 704
  2.2 Recursive ROC Curve Optimization - The TreeRank Algorithm......Page 706
  3.1 A Similarity-Learning Version of TreeRank......Page 708
  3.2 Generalization Ability - Rate Bound Analysis......Page 711
 4 Illustrative Numerical Experiments......Page 712
 References......Page 713
 1 Introduction......Page 715
 2 Assumptions and Methods......Page 716
  2.2 Assumptions......Page 717
  2.3 Background on Used Methods......Page 718
  3.2 Experiments......Page 719
  3.3 Experimental Results......Page 722
 4 Conclusion and Outlook......Page 723
 References......Page 724
Federated Learning of Deep Neural Decision Forests......Page 726
 1 Introduction......Page 727
 2 Deep Neural Decision Forests......Page 728
  2.1 Learning Deep Neural Decision Forests by Back-Propagation......Page 730
 3 Federated Learning......Page 731
 4 Experimental Results......Page 732
 5 Conclusion and Future Work......Page 734
 References......Page 735
 1 Introduction......Page 737
 2 Problem Statement......Page 738
 3 Overall Methodology......Page 739
  4.1 Solder Paste Printing Measurements......Page 740
  5.1 Background of Variational AutoEncoders......Page 741
  6.1 Datasets Description......Page 743
  6.3 Network Optimization and Hyperparameter Selection......Page 744
  6.5 Performance Metrics......Page 745
  7.1 Performance Analysis......Page 746
 8 Conclusion and Future Work......Page 748
 References......Page 749
 1 Introduction and Related Work......Page 751
 2 Methodology......Page 752
  2.1 Data Encoding Techniques......Page 753
  2.3 Machine Learning Techniques......Page 754
  2.4 Performance Metrics......Page 755
 3 Evaluation and Analysis......Page 756
  3.1 Anonymization of Textual Movie Reviews......Page 757
  3.2 Anonymization of Measurements from Satellite Sensors......Page 760
 4 Conclusions and Future Work......Page 761
 References......Page 762
 1 Introduction......Page 764
 3 The Service Point Distribution Problem......Page 766
 4 Cooperative Optimization Algorithm......Page 767
  4.1 Solution Management Component......Page 768
  4.3 Evaluation Component......Page 769
  4.4 Optimization Component......Page 770
  5.1 Benchmark Scenarios......Page 771
  5.2 Computational Results......Page 772
 6 Conclusion and Future Work......Page 774
 References......Page 775
 1 Introduction and Related Works......Page 777
  2.2 Network......Page 778
 4 Discussion and Conclusions......Page 779
 References......Page 780
 1 Introduction......Page 781
 3 Methodology......Page 782
 4 Evaluation and Results......Page 783
 References......Page 784
 1 Introduction......Page 785
 2 Preliminary Observations on Neural Net Learning......Page 786
 3 The Generalized Abs-Normal and Abs-Linear Forms......Page 788
 4 Three Optimization Strategies for the Inner Loop......Page 789
  4.1 Coordinate Global Descent......Page 790
  4.2 Target Oriented Averaging Trajectory Search......Page 791
  4.3 Mixed Integer Bilinear Optimization......Page 793
 5 Numerical Experiments and Comparison......Page 794
 6 Summary and Outlook......Page 795
 References......Page 796
Author Index......Page 797