File size: 83,675 Bytes
6fa4bc9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
{
    "paper_id": "P15-1013",
    "header": {
        "generated_with": "S2ORC 1.0.0",
        "date_generated": "2023-01-19T09:11:51.709299Z"
    },
    "title": "Low-Rank Regularization for Sparse Conjunctive Feature Spaces: An Application to Named Entity Classification",
    "authors": [
        {
            "first": "Audi",
            "middle": [],
            "last": "Primadhanty",
            "suffix": "",
            "affiliation": {
                "laboratory": "",
                "institution": "Universitat Polit\u00e8cnica de Catalunya",
                "location": {}
            },
            "email": "primadhanty@cs.upc.edu"
        },
        {
            "first": "Xavier",
            "middle": [],
            "last": "Carreras",
            "suffix": "",
            "affiliation": {},
            "email": "xavier.carreras@xrce.xerox.com"
        },
        {
            "first": "Ariadna",
            "middle": [],
            "last": "Quattoni",
            "suffix": "",
            "affiliation": {},
            "email": "ariadna.quattoni@xrce.xerox.com"
        }
    ],
    "year": "",
    "venue": null,
    "identifiers": {},
    "abstract": "Entity classification, like many other important problems in NLP, involves learning classifiers over sparse highdimensional feature spaces that result from the conjunction of elementary features of the entity mention and its context. In this paper we develop a low-rank regularization framework for training maxentropy models in such sparse conjunctive feature spaces. Our approach handles conjunctive feature spaces using matrices and induces an implicit low-dimensional representation via low-rank constraints. We show that when learning entity classifiers under minimal supervision, using a seed set, our approach is more effective in controlling model capacity than standard techniques for linear classifiers.",
    "pdf_parse": {
        "paper_id": "P15-1013",
        "_pdf_hash": "",
        "abstract": [
            {
                "text": "Entity classification, like many other important problems in NLP, involves learning classifiers over sparse highdimensional feature spaces that result from the conjunction of elementary features of the entity mention and its context. In this paper we develop a low-rank regularization framework for training maxentropy models in such sparse conjunctive feature spaces. Our approach handles conjunctive feature spaces using matrices and induces an implicit low-dimensional representation via low-rank constraints. We show that when learning entity classifiers under minimal supervision, using a seed set, our approach is more effective in controlling model capacity than standard techniques for linear classifiers.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Abstract",
                "sec_num": null
            }
        ],
        "body_text": [
            {
                "text": "Many important problems in NLP involve learning classifiers over sparse high-dimensional feature spaces that result from the conjunction of elementary features. For example, to classify an entity in a document, it is standard to exploit features of the left and right context in which the entity occurs as well as spelling features of the entity mention itself. These sets of features can be grouped into vectors which we call elementary feature vectors. In our example, there will be one elementary feature vector for the left context, one for the right context and one for the features of the mention. Observe that, when the elementary vectors consist of binary indicator features, the outer product of any pair of vectors represents all conjunctions of the corresponding elementary features.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "Ideally, we would like to train a classifier that can leverage all conjunctions of elementary features, since among them there might be some that are discriminative for the classification task at hand. However, allowing for such expressive high dimensional feature space comes at a cost: data sparsity becomes a key challenge and controlling the capacity of the model is crucial to avoid overfitting the training data.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "The problem of data sparsity is even more severe when the goal is to train classifiers with minimal supervision, i.e. small training sets. For example, in the entity classification setting we might be interested in training a classifier using only a small set of examples of each entity class. This is a typical scenario in an industrial setting, where developers are interested in classifying entities according to their own classification schema and can only provide a handful of examples of each class.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "A standard approach to control the capacity of a linear classifier is to use 1 or 2 regularization on the parameter vector. However, this type of regularization does not seem to be effective when dealing with sparse conjunctive feature spaces. The main limitation is that 1 and 2 regularization can not let the model give weight to conjunctions that have not been observed at training. Without such ability it is unlikely that the model will generalize to novel examples, where most of the conjunctions will be unseen in the training set.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "Of course, one could impose a strong prior on the weight vector so that it assigns weight to unseen conjunctions, but how can we build such a prior? What kind of reasonable constraints can we put on unseen conjunctions?",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "Another common approach to handle high dimensional conjunctive feature spaces is to manually design the feature function so that it includes only a subset of \"relevant\" conjunctions. But designing such a feature function can be time consuming and one might need to design a new feature function for each classification task. Ideally, we would have a learning algorithm that does not require such feature engineering and that it can automatically leverage rich conjunctive feature spaces.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "In this paper we present a solution to this problem by developing a regularization framework specifically designed for sparse conjunctive feature spaces. Our approach results in a more effective way of controlling model capacity and it does not require feature engineering.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "Our strategy is based on:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "\u2022 Employing tensors to define the scoring function of a max-entropy model as a multilinear form that computes weighted inner products between elementary vectors.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "\u2022 Forcing the model to induce low-dimensional embeddings of elementary vectors via lowrank regularization on the tensor parameters.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "The proposed regularization framework is based on a simple conceptual trick. The standard approach to handle conjunctive feature spaces in NLP is to regard the parameters of the linear model as long vectors computing an inner product with a high dimensional feature representation that lists explicitly all possible conjunctions. Instead, the parameters of our the model will be tensors and the compatibility score between an input pattern and a class will be defined as the sum of multilinear functions over elementary vectors.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "We then show that the rank 1 of the tensor has a very natural interpretation. It can be seen as the intrinsic dimensionality of a latent embedding of the elementary feature vectors. Thus by imposing a low-rank penalty on the tensor parameters we are encouraging the model to induce a lowdimensional projection of the elementary feature vectors . Using the rank itself as a regularization constraint in the learning algorithm would result in a non-convex optimization. Instead, we follow a standard approach which is to use the nuclear norm as a convex relaxation of the rank.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "In summary the main contributions of this paper are:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "\u2022 We develop a new regularization framework for training max-entropy models in high-dimensional sparse conjunctive feature spaces. Since the proposed regularization implicitly induces a low dimensional embedding of feature vectors, our algorithm can also be seen as a way of implicitly learning a latent variable model.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "\u2022 We present a simple convex learning algorithm for training the parameters of the model.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "\u2022 We conduct experiments on learning entity classifiers with minimal supervision. Our results show that the proposed regularization framework is better for sparse conjunctive feature spaces than standard 2 and 1 regularization. These results make us conclude that encouraging the max-entropy model to operate on a low-dimensional space is an effective way of controlling the capacity of the model an ensure good generalization.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "The formulation we develop in this paper applies to any prediction task whose inputs are some form of tuple. We focus on classification of entity mentions, or entities in the context of a sentence. Formally, our input objects are tuples x = l, e, r consisting of an entity e, a left context l and a right context r. The goal is to classify x into one entity class in the set Y. We will use log-linear models of the form:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Entity Classification with Log-linear Models",
                "sec_num": "2"
            },
            {
                "text": "EQUATION",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [
                    {
                        "start": 0,
                        "end": 8,
                        "text": "EQUATION",
                        "ref_id": "EQREF",
                        "raw_str": "Pr(y | x; \u03b8) = exp{s \u03b8 (x, y)} y exp{s \u03b8 (x, y )}",
                        "eq_num": "(1)"
                    }
                ],
                "section": "Entity Classification with Log-linear Models",
                "sec_num": "2"
            },
            {
                "text": "where s \u03b8 : X \u00d7 Y \u2192 R is a scoring function of entity tuples with a candidate class, and \u03b8 are the parameters of this function, to be specified below.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Entity Classification with Log-linear Models",
                "sec_num": "2"
            },
            {
                "text": "In the literature it is common to employ a feature-based linear model. That is, one defines a feature function \u03c6 : X \u2192 {0, 1} n that represents entity tuples in an n-dimensional binary feature space 2 , and the model has a weight vector for each class,",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Entity Classification with Log-linear Models",
                "sec_num": "2"
            },
            {
                "text": "\u03b8 = {w y } y\u2208Y . Then s \u03b8 (x, y) = \u03c6(x) \u2022 w y .",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Entity Classification with Log-linear Models",
                "sec_num": "2"
            },
            {
                "text": "In this section we propose a specific family of models for classifying entity tuples.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Low-rank Entity Classification Models",
                "sec_num": "3"
            },
            {
                "text": "We start from the observation that when representing tuple objects such as x = l, e, r with features, we often depart from a feature representation of each element of the tuple. Hence, let \u03c6 l and \u03c6 r be two feature functions representing left and right contexts, with binary dimensions d 1 and d 2 respectively. For now, we will define a model that ignores the entity mention e and makes predictions using context features. It is natural to define conjunctions of left and right features. Hence, in its most general form, one can define a matrix W y \u2208 R d 1 \u00d7d 2 for each class, such that \u03b8 = {W y } y\u2208Y and the score is:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "A Low-rank Model of Left-Right Contexts",
                "sec_num": "3.1"
            },
            {
                "text": "EQUATION",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [
                    {
                        "start": 0,
                        "end": 8,
                        "text": "EQUATION",
                        "ref_id": "EQREF",
                        "raw_str": "s \u03b8 ( l, e, r , y) = \u03c6 l (l) W y \u03c6 r (r) .",
                        "eq_num": "(2)"
                    }
                ],
                "section": "A Low-rank Model of Left-Right Contexts",
                "sec_num": "3.1"
            },
            {
                "text": "Note that this corresponds to a feature-based linear model operating in the product space of \u03c6 l and \u03c6 r , that is, the score has one term for each pair of features:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "A Low-rank Model of Left-Right Contexts",
                "sec_num": "3.1"
            },
            {
                "text": "i,j \u03c6 l (l)[i] \u03c6 r (r)[j] W y [i, j].",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "A Low-rank Model of Left-Right Contexts",
                "sec_num": "3.1"
            },
            {
                "text": "Note also that it is trivial to include elementary features of \u03c6 l and \u03c6 r , in addition to conjunctions, by having a constant dimension in each of the two representations set to 1.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "A Low-rank Model of Left-Right Contexts",
                "sec_num": "3.1"
            },
            {
                "text": "In all, the model in Eq. (2) is very expressive, with the caveat that it can easily overfit the data, specially when we work only with a handful of labeled examples. The standard way to control the capacity of a linear model is via 1 or 2 regularization.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "A Low-rank Model of Left-Right Contexts",
                "sec_num": "3.1"
            },
            {
                "text": "Regarding our parameters as matrices allows us to control the capacity of the model via regularizers that favor parameter matrices with low rank. To see the effect of these regularizers, consider that W y has rank k, and let W y = U y \u03a3 y V y be the singular value decomposition, where U y \u2208 R d 1 \u00d7k and V y \u2208 R d 2 \u00d7k are orthonormal projections and \u03a3 y \u2208 R k\u00d7k is a diagonal matrix of singular values. We can rewrite the score function as",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "A Low-rank Model of Left-Right Contexts",
                "sec_num": "3.1"
            },
            {
                "text": "s \u03b8 ( l, e, r , y) = (\u03c6 l (l) U y ) \u03a3 y (V y \u03c6 r (r)) .",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "A Low-rank Model of Left-Right Contexts",
                "sec_num": "3.1"
            },
            {
                "text": "(3) In words, the rank k is the intrinsic dimensionality of the inner product behind the score function. A low-rank regularizer will favor parameter matrices that have low intrinsic dimensionality. Below we describe a convex optimization for low-rank models using nuclear norm regularization.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "A Low-rank Model of Left-Right Contexts",
                "sec_num": "3.1"
            },
            {
                "text": "The model above classifies entities based only on the context. Here we propose an extension to make use of features of the entity. Let T be a set of possible entity feature tags, i.e. tags that describe an entity, such as ISCAPITALIZED, CONTAINSDIG-ITS, SINGLETOKEN, . . . Let \u03c6 e be a feature function representing entities. For this case, to simplify our expression, we will use a set notation and denote by \u03c6 e (e) \u2286 T the set of feature tags that describe e. Our model will be defined with one parameter matrix per feature tag and class label, i.e. \u03b8 = {W t,y } t\u2208T ,y\u2208Y . The model form is:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Adding Entity Features",
                "sec_num": "3.2"
            },
            {
                "text": "s \u03b8 ( l, e, r , y) = t\u2208\u03c6e(e) \u03c6 l (l) W t,y .\u03c6 r (r).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Adding Entity Features",
                "sec_num": "3.2"
            },
            {
                "text": "(4)",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Adding Entity Features",
                "sec_num": "3.2"
            },
            {
                "text": "In this section we describe a convex procedure to learn models of the above form that have low rank. We will define an objective that combines a loss and a regularization term.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Learning with Low-rank Constraints",
                "sec_num": "3.3"
            },
            {
                "text": "Our first observation is that our parameters are a tensor with up to four axes, namely left and right context representations, entity features, and entity classes. While a matrix has a clear definition of rank, it is not the case for general tensors, and there exist various definitions in the literature. The technique that we use is based on matricization of the tensor, that is, turning the tensor into a matrix that has the same parameters as the tensor but organized in two axes. This is done by partitioning the tensor axes into two sets, one for matrix rows and another for columns. Once the tensor has been turned into a matrix, we can use the standard definition of matrix rank. A main advantage of this approach is that we can make use of standard routines like singular value decomposition (SVD) to decompose the matricized tensor. This is the main reason behind our choice.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Learning with Low-rank Constraints",
                "sec_num": "3.3"
            },
            {
                "text": "In general, different ways of partitioning the tensor axes will lead to different notions of intrinsic dimensions. In our case we choose the left context axes as the row dimension, and the rest of axes as the column dimension. 3 In this section, we will denote as W the matricized version of the parameters \u03b8 of our models.",
                "cite_spans": [
                    {
                        "start": 227,
                        "end": 228,
                        "text": "3",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Learning with Low-rank Constraints",
                "sec_num": "3.3"
            },
            {
                "text": "The second observation is that minimizing the rank of a matrix is a non-convex problem. We make use of a convex relaxation based on the nuclear norm (Srebro and Shraibman, 2005) . The nuclear norm 4 of a matrix W, denoted W , is the sum of its singular values:",
                "cite_spans": [
                    {
                        "start": 149,
                        "end": 177,
                        "text": "(Srebro and Shraibman, 2005)",
                        "ref_id": "BIBREF15"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Learning with Low-rank Constraints",
                "sec_num": "3.3"
            },
            {
                "text": "W = i \u03a3 i,i",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Learning with Low-rank Constraints",
                "sec_num": "3.3"
            },
            {
                "text": "where W = U\u03a3V is the singular value decomposition of W. This norm has been used in several applications in machine learning as a convex surrogate for imposing low rank, e.g. (Srebro et al., 2004) .",
                "cite_spans": [
                    {
                        "start": 174,
                        "end": 195,
                        "text": "(Srebro et al., 2004)",
                        "ref_id": "BIBREF16"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Learning with Low-rank Constraints",
                "sec_num": "3.3"
            },
            {
                "text": "Thus, the nuclear norm is used as a regularizer. With this, we define our objective as follows:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Learning with Low-rank Constraints",
                "sec_num": "3.3"
            },
            {
                "text": "EQUATION",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [
                    {
                        "start": 0,
                        "end": 8,
                        "text": "EQUATION",
                        "ref_id": "EQREF",
                        "raw_str": "argmin W L(W) + \u03c4 R(W) ,",
                        "eq_num": "(5)"
                    }
                ],
                "section": "Learning with Low-rank Constraints",
                "sec_num": "3.3"
            },
            {
                "text": "where L(W) is a convex loss function, R(W) is a regularizer, and \u03c4 is a constant that trades off error and capacity. In experiments we will compare nuclear norm regularization with 1 and 2 regularizers. In all cases we use the negative log-likelihood as loss function, denoting the training data as D:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Learning with Low-rank Constraints",
                "sec_num": "3.3"
            },
            {
                "text": "L(W) = ( l,e,r ,y)\u2208D \u2212 log Pr(y | l, e, r ; W) .",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Learning with Low-rank Constraints",
                "sec_num": "3.3"
            },
            {
                "text": "(6) To solve the objective in Eq. (5) we use a simple optimization scheme known as forward-backward splitting (FOBOS) (Duchi and Singer, 2009) . In a series of iterations, this algorithm performs a gradient update followed by a proximal projection of the parameters. Such projection depends on the regularizer used: for 1 it thresholds the parameters; for 2 it scales them; and for nuclearnorm regularization it thresholds the singular values. This means that, for nuclear norm regularization, each iteration requires to decompose W using SVD. See (Madhyastha et al., 2014) for details about this optimization for a related application.",
                "cite_spans": [
                    {
                        "start": 118,
                        "end": 142,
                        "text": "(Duchi and Singer, 2009)",
                        "ref_id": "BIBREF4"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Learning with Low-rank Constraints",
                "sec_num": "3.3"
            },
            {
                "text": "The main aspect of our approach is the use of a spectral penalty (i.e., the rank) to control the capacity of multilinear functions parameterized by matrices or tensors. used nuclear-norm regularization to learn latentvariable max-margin sequence taggers. Madhyastha et al. (2014) defined bilexical distribu- Lei et al. (2014) also use low-rank tensor learning in the context of dependency parsing, where like in our case dependencies are represented by conjunctive feature spaces. While the motivation is similar, their technical solution is different. We use the technique of matricization of a tensor combined with a nuclear-norm relaxation to obtain a convex learning procedure. In their case they explicitly look for a low-dimensional factorization of the tensor using a greedy alternating optimization. Also recently, have framed entity classification as a low-rank matrix completion problem. The idea is based on the fact that if two entities (in rows) have similar descriptions (in columns) they should have similar classes. The low-rank structure of the matrix defines intrinsic representations of entities and feature descriptions. The same idea was applied to relation extraction , using a matrix of entity pairs times descriptions that corresponds to a matricization of an entity-entity-description tensor. Very recently Singh et al. (2015) explored alternative ways of applying low-rank constraints to tensor-based relation extraction.",
                "cite_spans": [
                    {
                        "start": 308,
                        "end": 325,
                        "text": "Lei et al. (2014)",
                        "ref_id": "BIBREF6"
                    },
                    {
                        "start": 1332,
                        "end": 1351,
                        "text": "Singh et al. (2015)",
                        "ref_id": "BIBREF14"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Related Work",
                "sec_num": "4"
            },
            {
                "text": "Another aspect of this paper is training entity classification models using minimal supervision, which has been addressed by multiple works in the literature. A classical successful approach for this problem is to use co-training (Blum and Mitchell, 1998) : learn two classifiers that use different views of the data by using each other's predictions. In the same line, Collins and Singer (1999) trained entity classifiers by bootstraping from an initial set of seeds, using a boosting version of co-training. Seed sets have also been exploited by graphical model approaches. Haghighi and Klein (2006) define a graphical model that is soft-constrained such that the prediction for an unlabeled example agrees with the labels of seeds that are distributionally similar. Li et al. (2010) present a Bayesian approach to expand an initial seed set, with the goal of creating a gazetteer.",
                "cite_spans": [
                    {
                        "start": 230,
                        "end": 255,
                        "text": "(Blum and Mitchell, 1998)",
                        "ref_id": "BIBREF1"
                    },
                    {
                        "start": 370,
                        "end": 395,
                        "text": "Collins and Singer (1999)",
                        "ref_id": "BIBREF3"
                    },
                    {
                        "start": 576,
                        "end": 601,
                        "text": "Haghighi and Klein (2006)",
                        "ref_id": "BIBREF5"
                    },
                    {
                        "start": 769,
                        "end": 785,
                        "text": "Li et al. (2010)",
                        "ref_id": "BIBREF7"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Related Work",
                "sec_num": "4"
            },
            {
                "text": "Another approach to entity recognition that, like in our case, learns projections of contextual features is the method by Ando and Zhang (2005) . : For each entity class, the seed of entities for the 10-30 set, together with the number of mentions in the training data that involve entities in the seed for various sizes of the seeds.",
                "cite_spans": [
                    {
                        "start": 122,
                        "end": 143,
                        "text": "Ando and Zhang (2005)",
                        "ref_id": "BIBREF0"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Related Work",
                "sec_num": "4"
            },
            {
                "text": "They define a set of auxiliary tasks, which can be supervised using unlabeled data, and find a projection of the data that works well as input representation for the auxiliary tasks. This representation is then used for the target task. More recently Neelakantan and Collins (2014) presented another approach to gazetteer expansion using an initial seed. A novel aspect is the use of Canonical Correlation Analysis (CCA) to compute embeddings of entity contexts, that are used by the named entity classifier. Like in our case, their method learns a compressed representation of contexts that helps prediction.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Related Work",
                "sec_num": "4"
            },
            {
                "text": "In this section we evaluate our regularization framework for training models in highdimensional sparse conjunctive feature spaces. We run experiments on learning entity classifiers with minimal supervision. We focus on classification of unseen entities to highlight the ability of the regularizer to generalize over conjunctions that are not observed at training. We simulate minimal supervision using the CoNLL-2003 Shared Task data (Tjong Kim Sang and De Meulder, 2003) , and compare the performance to 1 and 2 regularizers.",
                "cite_spans": [
                    {
                        "start": 434,
                        "end": 471,
                        "text": "(Tjong Kim Sang and De Meulder, 2003)",
                        "ref_id": "BIBREF17"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Experiments",
                "sec_num": "5"
            },
            {
                "text": "We use a minimal supervision setting where we provide the algorithm a seed of entities for each class, that is, a list of entities that is representative for that class. The assumption is that any mention of an entity in the seed is a positive example for the corresponding class. Given unlabeled data and a seed of entities for each class, the goal is to learn a model that correctly classifies mentions of entities that are not in the seed. In addition to standard entity classes, we also consider a special non-entity class, which is part of the classification but is excluded from evaluation.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Minimal Supervision Task",
                "sec_num": "5.1"
            },
            {
                "text": "Note that named entity classification for unseen entities is a challenging problem. Even in the standard fully-supervised scenario, when we measure the performance of state-of-the-art methods on unseen entities, the F1 values are in the range of 60%. This represents a significant drop with respect to the standard metrics for named entity recognition, which consider all entity mentions of the test set irrespective of whether they appear in the training data or not, and where F1 values at 90% levels are obtained (e.g. (Ratinov and Roth, 2009) ). This suggests that part of the success of state-of-the-art models is in storing known entities together with their type (in the form of gazetteers or directly in lexicalized parameters of the model).",
                "cite_spans": [
                    {
                        "start": 522,
                        "end": 546,
                        "text": "(Ratinov and Roth, 2009)",
                        "ref_id": "BIBREF12"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Minimal Supervision Task",
                "sec_num": "5.1"
            },
            {
                "text": "We use the CoNLL-2003 English data, which is annotated with four types: person (PER), location (LOC), organization (ORG), and miscellaneous (MISC). In addition, the data is tagged with partsof-speech (PoS), and we compute word clusters running the Brown clustering algorithm (Brown et al., 1992) on the words in the training set.",
                "cite_spans": [
                    {
                        "start": 275,
                        "end": 295,
                        "text": "(Brown et al., 1992)",
                        "ref_id": "BIBREF2"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Setting",
                "sec_num": "5.2"
            },
            {
                "text": "We consider annotated entity phrases as candidate entities, and all single nouns that are not part of an entity as candidate non-entities (O). Both candidate entities and non-entities will be referred to as candidates in the remaining of this section. We lowercase all candidates and remove the am- biguous ones (i.e., those with more than one label in different mentions). 5 To simulate a minimal supervision, we create supervision seeds by picking the n most frequent training candidates for entity types, and the m most frequent candidate non-entities. We create seeds of various sizes n-m, namely 10-30, 40-120, 640-1920, as well as all of the candidates. For each seed, the training set consists of all training mentions that involve entities in the seed. Table 1 shows the smaller seed, as well as the number of mentions for each seed size.",
                "cite_spans": [
                    {
                        "start": 374,
                        "end": 375,
                        "text": "5",
                        "ref_id": null
                    }
                ],
                "ref_spans": [
                    {
                        "start": 761,
                        "end": 768,
                        "text": "Table 1",
                        "ref_id": "TABREF1"
                    }
                ],
                "eq_spans": [],
                "section": "Setting",
                "sec_num": "5.2"
            },
            {
                "text": "For evaluation we use the development and test sections of the data, but we remove the instances of candidates in the training data (i.e., that are in the all seed). We do not remove instances that are ambiguous in the tests. 6 As evaluation metric we use the average F1 score computed over all entity types, excluding the non-entity type. ",
                "cite_spans": [
                    {
                        "start": 226,
                        "end": 227,
                        "text": "6",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Setting",
                "sec_num": "5.2"
            },
            {
                "text": "We refer to context as the sequence of tokens before (left context) and after (right context) a candidate mention in a sentence. Different classifiers can be built using different representations of the contexts. For example we can change the window size of the context sequence (i.e., for a window size of 1 we only use the last token before the mention and the first token after the mention). We can treat the left and right contexts independently of each other, we can treat them as a unique combination, or we can use both. We can also choose to use the word form of a token, its PoS tag, a word cluster, or a combination of these. we will use the elementary features that are more predictive and compact: clusters and PoS tags in windows of size at most 2.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Context Representations",
                "sec_num": "5.3"
            },
            {
                "text": "We compare the performance of models trained using the nuclear norm regularizer with models trained using 1 and 2 regularizers. To train each model, we validate the regularization parameter and the number of iterations on development data, trying a wide range of values. The best performing configuration is then used for the comparison. Figure 1 shows results on the development set for different feature sets. We started representing context using cluster labels, as it is the most compact representation obtaining good results in preliminary experiments. We tried several conjunctions: a conjunction of the left and right context, as well as conjunctions of left and right contexts and features of the candidate entity. We also tried all different conjunction combinations of the contexts and the candidate entity features, as well as adding PoS tags to represent contexts. To represent an entity candidate we use standard traits of the spelling of the mention, such as capitalization, ation. Using our richest feature set, the model obtains 76.76 of accuracy in the development, for the task of classifing entities with correct boundaries. If we add features capturing the full entity and its tokens, then the accuracy is 87.63, which is similar to state-of-the-art performance (the best results in literature typically exploit additional gazetteers). Since our evaluation focuses on unknown entities, our features do not include information about the word tokens of entites. the existence of symbols, as well as the number of tokens in the candidate. See Table 3 for the definition of the features describing entity candidates. We observe that for most conjunction settings our regularizer performs better than the 1 and 2 regularizers. Using the best model from each regularizer, we evaluated on the test set. Table  4 shows the test results. For all seed sets, the nuclear norm regularizer obtains the best average F1 performance. This shows that encouraging the max-entropy model to operate on a lowdimensional space is effective. Moreover, Figure  2 shows model performance as a function of the number of dimensions of the intrinsic projection. The model obtains a good performance even if only a few intrinsic dimensions are used. Figure 1f trained with the 10-30 seed, with respect to observations of the associated features in training and development. Non-white conjunctions correspond to non-zero weights: black is for conjunctions seen in both the training and development sets; blue is for those seen in training but not in the development; red indicates that the conjunctions were observed only in the development; yellow is for those not observed in training nor development.",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 338,
                        "end": 346,
                        "text": "Figure 1",
                        "ref_id": null
                    },
                    {
                        "start": 1560,
                        "end": 1567,
                        "text": "Table 3",
                        "ref_id": "TABREF6"
                    },
                    {
                        "start": 1816,
                        "end": 1824,
                        "text": "Table  4",
                        "ref_id": "TABREF7"
                    },
                    {
                        "start": 2049,
                        "end": 2058,
                        "text": "Figure  2",
                        "ref_id": "FIGREF0"
                    },
                    {
                        "start": 2241,
                        "end": 2250,
                        "text": "Figure 1f",
                        "ref_id": null
                    }
                ],
                "eq_spans": [],
                "section": "Comparing Regularizers",
                "sec_num": "5.4"
            },
            {
                "text": "rank model in Figure 1f trained with the 10-30 seed, with respect to observed features in training and development data. Many of the conjunctions of the development set were never observed in the training set. Our regularizer framework is able to propagate weights from the conjunctive features seen in training to unseen conjunctive features that are close to each other in the projected space (these are the yellow and red cells in the matrix). In contrast, 1 and 2 regularization techniques can not put weight on unseen conjunctions.",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 14,
                        "end": 23,
                        "text": "Figure 1f",
                        "ref_id": null
                    }
                ],
                "eq_spans": [],
                "section": "Comparing Regularizers",
                "sec_num": "5.4"
            },
            {
                "text": "We have developed a low-rank regularization framework for training max-entropy models in sparse conjunctive feature spaces. Our formulation is based on using tensors to parameterize classifiers. We control the capacity of the model using the nuclear-norm of a matricization of the tensor. Overall, our formulation results in a convex procedure for training model parameters. We have experimented with these techniques in the context of learning entity classifiers. Compared to 1 and 2 penalties, the low-rank model obtains better performance, without the need to manually specify feature conjunctions. In our analysis, we have illustrated how the low-rank approach can assign non-zero weights to conjunctions that were unobserved at training, but are similar to observed conjunctions with respect to the low-dimensional projection of their elements. We have used matricization of a tensor to define its rank, using a fixed transformation of the tensor into a matrix. Future work should explore how to combine efficiently different transformations.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Conclusion",
                "sec_num": "6"
            },
            {
                "text": "There are many ways of defining the rank of a tensor. In this paper we matricize tensors into matrices and use the rank of the resulting matrix. Matricization is also referred to as unfolding.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            },
            {
                "text": "In general, all models in this paper accept real-valued feature functions. But we focus on binary indicator features because in practice these are the standard type of features in NLP classifiers, and the ones we use here. In fact, in this paper we develop feature spaces based on products of elementary feature functions, in which case the resulting representations correspond to conjunctions of the elementary features.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            },
            {
                "text": "In preliminary experiments we tried variations, such as having right prefixes in the columns, and left prefixes, entity tags and classes in the rows. We only observer minor, nonsignificant variations in the results.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            },
            {
                "text": "Also known as the trace norm. tions parameterized by matrices which result lexical embeddings tailored for a particular linguistic relation. Like in our case, the low-dimensional latent projections in these papers are learned implicitly by imposing low-rank constraints on the predictions of the model.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            },
            {
                "text": "In the CoNLL-2003 English training set, only 235 candidates are ambiguous out of 13,441 candidates, i.e. less than 2%. This suggests that in this data the difficulty behind the task is in recognizing and classifying unseen entities, and not in disambiguating known entities in a certain context.6 After removing the ambiguous candidates from the training data, and removing candidates seen in the training from the development and test sets, this is the number of mentions (and number of unique candidates in parenthesis) in the data used in our experiments:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            },
            {
                "text": "regularization, using the 10-30 seed. We use the lexical representation (the word itself) and a word cluster representation of the context tokens and use a window size of one to three. We use two types of features: bag-of-words features (1grams of tokens in the specified window) and ngram features (with n smaller or equal to the window size). The performance of using word clusters is comparable, and sometimes better, to using lexical representations. Moreover, using a longer window, in this case, does not necessarily result in better performance.7 In the rest of the experiments 7 Our learner and feature configuration, using 2 regularization, obtains state-of-the-art results on the standard evalu-",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            },
            {
                "text": "regularization. Each plot corresponds to a different conjunctive feature space with respect to window size (1 or 2), context representation (cluster with/out PoS), using entity features or not, and combining or not full conjunctions with lower-order conjunctions and elementary features.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            }
        ],
        "back_matter": [
            {
                "text": "We thank Gabriele Musillo and the anonymous reviewers for their helpful comments and suggestions. This work has been partially funded by the Spanish Government through the SKATER project (TIN2012-38584-C06-01) and an FPI predoctoral grant for Audi Primadhanty.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Acknowledgements",
                "sec_num": null
            }
        ],
        "bib_entries": {
            "BIBREF0": {
                "ref_id": "b0",
                "title": "A framework for learning predictive structures from multiple tasks and unlabeled data",
                "authors": [
                    {
                        "first": "Rie",
                        "middle": [],
                        "last": "Kubota",
                        "suffix": ""
                    },
                    {
                        "first": "Ando",
                        "middle": [],
                        "last": "",
                        "suffix": ""
                    },
                    {
                        "first": "Tong",
                        "middle": [],
                        "last": "Zhang",
                        "suffix": ""
                    }
                ],
                "year": 2005,
                "venue": "J. Mach. Learn. Res",
                "volume": "6",
                "issue": "",
                "pages": "1817--1853",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Rie Kubota Ando and Tong Zhang. 2005. A frame- work for learning predictive structures from multi- ple tasks and unlabeled data. J. Mach. Learn. Res., 6:1817-1853, December.",
                "links": null
            },
            "BIBREF1": {
                "ref_id": "b1",
                "title": "Combining labeled and unlabeled data with co-training",
                "authors": [
                    {
                        "first": "Avrim",
                        "middle": [],
                        "last": "Blum",
                        "suffix": ""
                    },
                    {
                        "first": "Tom",
                        "middle": [],
                        "last": "Mitchell",
                        "suffix": ""
                    }
                ],
                "year": 1998,
                "venue": "Proceedings of the Eleventh Annual Conference on Computational Learning Theory, COLT' 98",
                "volume": "",
                "issue": "",
                "pages": "92--100",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Avrim Blum and Tom Mitchell. 1998. Combining labeled and unlabeled data with co-training. In Proceedings of the Eleventh Annual Conference on Computational Learning Theory, COLT' 98, pages 92-100, New York, NY, USA. ACM.",
                "links": null
            },
            "BIBREF2": {
                "ref_id": "b2",
                "title": "Class-based n-gram models of natural language",
                "authors": [
                    {
                        "first": "F",
                        "middle": [],
                        "last": "Peter",
                        "suffix": ""
                    },
                    {
                        "first": "Peter",
                        "middle": [
                            "V"
                        ],
                        "last": "Brown",
                        "suffix": ""
                    },
                    {
                        "first": "Robert",
                        "middle": [
                            "L"
                        ],
                        "last": "Desouza",
                        "suffix": ""
                    },
                    {
                        "first": "Vincent",
                        "middle": [
                            "J"
                        ],
                        "last": "Mercer",
                        "suffix": ""
                    },
                    {
                        "first": "Jenifer",
                        "middle": [
                            "C"
                        ],
                        "last": "Della Pietra",
                        "suffix": ""
                    },
                    {
                        "first": "",
                        "middle": [],
                        "last": "Lai",
                        "suffix": ""
                    }
                ],
                "year": 1992,
                "venue": "Computational Linguistics",
                "volume": "18",
                "issue": "",
                "pages": "467--479",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Peter F. Brown, Peter V. deSouza, Robert L. Mer- cer, Vincent J. Della Pietra, and Jenifer C. Lai. 1992. Class-based n-gram models of natural lan- guage. Computational Linguistics, 18:467-479.",
                "links": null
            },
            "BIBREF3": {
                "ref_id": "b3",
                "title": "Unsupervised models for named entity classification",
                "authors": [
                    {
                        "first": "Michael",
                        "middle": [],
                        "last": "Collins",
                        "suffix": ""
                    },
                    {
                        "first": "Yoram",
                        "middle": [],
                        "last": "Singer",
                        "suffix": ""
                    }
                ],
                "year": 1999,
                "venue": "Joint SIGDAT Conference on Empirical Methods in Natural Language Processing and Very Large Corpora",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Michael Collins and Yoram Singer. 1999. Unsuper- vised models for named entity classification. In Joint SIGDAT Conference on Empirical Methods in Natural Language Processing and Very Large Cor- pora.",
                "links": null
            },
            "BIBREF4": {
                "ref_id": "b4",
                "title": "Efficient online and batch learning using forward backward splitting",
                "authors": [
                    {
                        "first": "John",
                        "middle": [],
                        "last": "Duchi",
                        "suffix": ""
                    },
                    {
                        "first": "Yoram",
                        "middle": [],
                        "last": "Singer",
                        "suffix": ""
                    }
                ],
                "year": 2009,
                "venue": "Journal of Machine Learning Research",
                "volume": "10",
                "issue": "",
                "pages": "2899--2934",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "John Duchi and Yoram Singer. 2009. Efficient online and batch learning using forward backward splitting. Journal of Machine Learning Research, 10:2899- 2934.",
                "links": null
            },
            "BIBREF5": {
                "ref_id": "b5",
                "title": "Prototype-driven learning for sequence models",
                "authors": [
                    {
                        "first": "Aria",
                        "middle": [],
                        "last": "Haghighi",
                        "suffix": ""
                    },
                    {
                        "first": "Dan",
                        "middle": [],
                        "last": "Klein",
                        "suffix": ""
                    }
                ],
                "year": 2006,
                "venue": "Proceedings of the Main Conference on Human Language Technology Conference of the North American Chapter of the Association of Computational Linguistics, HLT-NAACL '06",
                "volume": "",
                "issue": "",
                "pages": "320--327",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Aria Haghighi and Dan Klein. 2006. Prototype-driven learning for sequence models. In Proceedings of the Main Conference on Human Language Tech- nology Conference of the North American Chap- ter of the Association of Computational Linguistics, HLT-NAACL '06, pages 320-327, Stroudsburg, PA, USA. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF6": {
                "ref_id": "b6",
                "title": "Low-rank tensors for scoring dependency structures",
                "authors": [
                    {
                        "first": "Tao",
                        "middle": [],
                        "last": "Lei",
                        "suffix": ""
                    },
                    {
                        "first": "Yu",
                        "middle": [],
                        "last": "Xin",
                        "suffix": ""
                    },
                    {
                        "first": "Yuan",
                        "middle": [],
                        "last": "Zhang",
                        "suffix": ""
                    },
                    {
                        "first": "Regina",
                        "middle": [],
                        "last": "Barzilay",
                        "suffix": ""
                    },
                    {
                        "first": "Tommi",
                        "middle": [],
                        "last": "Jaakkola",
                        "suffix": ""
                    }
                ],
                "year": 2014,
                "venue": "Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics",
                "volume": "1",
                "issue": "",
                "pages": "1381--1391",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Tao Lei, Yu Xin, Yuan Zhang, Regina Barzilay, and Tommi Jaakkola. 2014. Low-rank tensors for scor- ing dependency structures. In Proceedings of the 52nd Annual Meeting of the Association for Compu- tational Linguistics (Volume 1: Long Papers), pages 1381-1391, Baltimore, Maryland, June. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF7": {
                "ref_id": "b7",
                "title": "Distributional similarity vs. pu learning for entity set expansion",
                "authors": [
                    {
                        "first": "Xiao-Li",
                        "middle": [],
                        "last": "Li",
                        "suffix": ""
                    },
                    {
                        "first": "Lei",
                        "middle": [],
                        "last": "Zhang",
                        "suffix": ""
                    },
                    {
                        "first": "Bing",
                        "middle": [],
                        "last": "Liu",
                        "suffix": ""
                    },
                    {
                        "first": "See-Kiong",
                        "middle": [],
                        "last": "Ng",
                        "suffix": ""
                    }
                ],
                "year": 2010,
                "venue": "Proceedings of the ACL 2010 Conference Short Papers, ACLShort '10",
                "volume": "",
                "issue": "",
                "pages": "359--364",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Xiao-Li Li, Lei Zhang, Bing Liu, and See-Kiong Ng. 2010. Distributional similarity vs. pu learn- ing for entity set expansion. In Proceedings of the ACL 2010 Conference Short Papers, ACLShort '10, pages 359-364, Stroudsburg, PA, USA. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF8": {
                "ref_id": "b8",
                "title": "Learning Task-specific Bilexical Embeddings",
                "authors": [
                    {
                        "first": "Xavier",
                        "middle": [],
                        "last": "Pranava Swaroop Madhyastha",
                        "suffix": ""
                    },
                    {
                        "first": "Ariadna",
                        "middle": [],
                        "last": "Carreras",
                        "suffix": ""
                    },
                    {
                        "first": "",
                        "middle": [],
                        "last": "Quattoni",
                        "suffix": ""
                    }
                ],
                "year": 2014,
                "venue": "Proceedings of COLING 2014, the 25th International Conference on Computational Linguistics: Technical Papers",
                "volume": "",
                "issue": "",
                "pages": "161--171",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Pranava Swaroop Madhyastha, Xavier Carreras, and Ariadna Quattoni. 2014. Learning Task-specific Bilexical Embeddings. In Proceedings of COLING 2014, the 25th International Conference on Compu- tational Linguistics: Technical Papers, pages 161- 171, Dublin, Ireland, August. Dublin City Univer- sity and Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF9": {
                "ref_id": "b9",
                "title": "Mallet: A machine learning for language toolkit",
                "authors": [
                    {
                        "first": "Andrew",
                        "middle": [
                            "K"
                        ],
                        "last": "Mccallum",
                        "suffix": ""
                    }
                ],
                "year": 2002,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Andrew K. McCallum. 2002. Mallet: A machine learning for language toolkit.",
                "links": null
            },
            "BIBREF10": {
                "ref_id": "b10",
                "title": "Learning dictionaries for named entity recognition using minimal supervision",
                "authors": [
                    {
                        "first": "Arvind",
                        "middle": [],
                        "last": "Neelakantan",
                        "suffix": ""
                    },
                    {
                        "first": "Michael",
                        "middle": [],
                        "last": "Collins",
                        "suffix": ""
                    }
                ],
                "year": 2014,
                "venue": "Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics",
                "volume": "",
                "issue": "",
                "pages": "452--461",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Arvind Neelakantan and Michael Collins. 2014. Learning dictionaries for named entity recognition using minimal supervision. In Proceedings of the 14th Conference of the European Chapter of the As- sociation for Computational Linguistics, pages 452- 461, Gothenburg, Sweden, April. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF11": {
                "ref_id": "b11",
                "title": "Spectral regularization for max-margin sequence tagging",
                "authors": [
                    {
                        "first": "Ariadna",
                        "middle": [],
                        "last": "Quattoni",
                        "suffix": ""
                    },
                    {
                        "first": "Borja",
                        "middle": [],
                        "last": "Balle",
                        "suffix": ""
                    },
                    {
                        "first": "Xavier",
                        "middle": [],
                        "last": "Carreras",
                        "suffix": ""
                    },
                    {
                        "first": "Amir",
                        "middle": [],
                        "last": "Globerson",
                        "suffix": ""
                    }
                ],
                "year": 2014,
                "venue": "Proceedings of the 31st International Conference on Machine Learning (ICML-14)",
                "volume": "",
                "issue": "",
                "pages": "1710--1718",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Ariadna Quattoni, Borja Balle, Xavier Carreras, and Amir Globerson. 2014. Spectral regularization for max-margin sequence tagging. In Tony Jebara and Eric P. Xing, editors, Proceedings of the 31st Inter- national Conference on Machine Learning (ICML- 14), pages 1710-1718. JMLR Workshop and Con- ference Proceedings.",
                "links": null
            },
            "BIBREF12": {
                "ref_id": "b12",
                "title": "Design challenges and misconceptions in named entity recognition",
                "authors": [
                    {
                        "first": "Lev",
                        "middle": [],
                        "last": "Ratinov",
                        "suffix": ""
                    },
                    {
                        "first": "Dan",
                        "middle": [],
                        "last": "Roth",
                        "suffix": ""
                    }
                ],
                "year": 2009,
                "venue": "Proceedings of the Thirteenth Conference on Computational Natural Language Learning (CoNLL-2009)",
                "volume": "",
                "issue": "",
                "pages": "147--155",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Lev Ratinov and Dan Roth. 2009. Design chal- lenges and misconceptions in named entity recog- nition. In Proceedings of the Thirteenth Confer- ence on Computational Natural Language Learning (CoNLL-2009), pages 147-155, Boulder, Colorado, June. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF13": {
                "ref_id": "b13",
                "title": "Relation extraction with matrix factorization and universal schemas",
                "authors": [
                    {
                        "first": "Sebastian",
                        "middle": [],
                        "last": "Riedel",
                        "suffix": ""
                    },
                    {
                        "first": "Limin",
                        "middle": [],
                        "last": "Yao",
                        "suffix": ""
                    },
                    {
                        "first": "Andrew",
                        "middle": [],
                        "last": "Mccallum",
                        "suffix": ""
                    },
                    {
                        "first": "Benjamin",
                        "middle": [
                            "M"
                        ],
                        "last": "Marlin",
                        "suffix": ""
                    }
                ],
                "year": 2013,
                "venue": "Proceedings of the 2013 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies",
                "volume": "",
                "issue": "",
                "pages": "74--84",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Sebastian Riedel, Limin Yao, Andrew McCallum, and Benjamin M. Marlin. 2013. Relation extraction with matrix factorization and universal schemas. In Proceedings of the 2013 Conference of the North American Chapter of the Association for Computa- tional Linguistics: Human Language Technologies, pages 74-84, Atlanta, Georgia, June. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF14": {
                "ref_id": "b14",
                "title": "Towards Combined Matrix and Tensor Factorization for Universal Schema Relation Extraction",
                "authors": [
                    {
                        "first": "Sameer",
                        "middle": [],
                        "last": "Singh",
                        "suffix": ""
                    },
                    {
                        "first": "Tim",
                        "middle": [],
                        "last": "Rockt\u00e4schel",
                        "suffix": ""
                    },
                    {
                        "first": "Sebastian",
                        "middle": [],
                        "last": "Riedel",
                        "suffix": ""
                    }
                ],
                "year": 2015,
                "venue": "NAACL Workshop on Vector Space Modeling for NLP (VSM)",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Sameer Singh, Tim Rockt\u00e4schel, and Sebastian Riedel. 2015. Towards Combined Matrix and Tensor Fac- torization for Universal Schema Relation Extraction. In NAACL Workshop on Vector Space Modeling for NLP (VSM).",
                "links": null
            },
            "BIBREF15": {
                "ref_id": "b15",
                "title": "Rank, tracenorm and max-norm",
                "authors": [
                    {
                        "first": "Nathan",
                        "middle": [],
                        "last": "Srebro",
                        "suffix": ""
                    },
                    {
                        "first": "Adi",
                        "middle": [],
                        "last": "Shraibman",
                        "suffix": ""
                    }
                ],
                "year": 2005,
                "venue": "Learning Theory",
                "volume": "",
                "issue": "",
                "pages": "545--560",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Nathan Srebro and Adi Shraibman. 2005. Rank, trace- norm and max-norm. In Learning Theory, pages 545-560. Springer Berlin Heidelberg.",
                "links": null
            },
            "BIBREF16": {
                "ref_id": "b16",
                "title": "Maximum-margin matrix factorization",
                "authors": [
                    {
                        "first": "Nathan",
                        "middle": [],
                        "last": "Srebro",
                        "suffix": ""
                    },
                    {
                        "first": "Jason",
                        "middle": [],
                        "last": "Rennie",
                        "suffix": ""
                    },
                    {
                        "first": "Tommi",
                        "middle": [
                            "S"
                        ],
                        "last": "Jaakkola",
                        "suffix": ""
                    }
                ],
                "year": 2004,
                "venue": "Advances in neural information processing systems",
                "volume": "",
                "issue": "",
                "pages": "1329--1336",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Nathan Srebro, Jason Rennie, and Tommi S Jaakkola. 2004. Maximum-margin matrix factorization. In Advances in neural information processing systems, pages 1329-1336.",
                "links": null
            },
            "BIBREF17": {
                "ref_id": "b17",
                "title": "Introduction to the CoNLL-2003 Shared Task: Language-Independent Named Entity Recognition",
                "authors": [
                    {
                        "first": "Erik",
                        "middle": [
                            "F"
                        ],
                        "last": "Tjong",
                        "suffix": ""
                    },
                    {
                        "first": "Kim",
                        "middle": [],
                        "last": "Sang",
                        "suffix": ""
                    },
                    {
                        "first": "Fien",
                        "middle": [],
                        "last": "De Meulder",
                        "suffix": ""
                    }
                ],
                "year": 2003,
                "venue": "Proceedings of the Seventh Conference on Natural Language Learning at HLT-NAACL 2003",
                "volume": "",
                "issue": "",
                "pages": "142--147",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Erik F. Tjong Kim Sang and Fien De Meulder. 2003. Introduction to the CoNLL-2003 Shared Task: Language-Independent Named Entity Recog- nition. In Proceedings of the Seventh Conference on Natural Language Learning at HLT-NAACL 2003, pages 142-147.",
                "links": null
            },
            "BIBREF18": {
                "ref_id": "b18",
                "title": "Universal schema for entity type prediction",
                "authors": [
                    {
                        "first": "Limin",
                        "middle": [],
                        "last": "Yao",
                        "suffix": ""
                    },
                    {
                        "first": "Sebastian",
                        "middle": [],
                        "last": "Riedel",
                        "suffix": ""
                    },
                    {
                        "first": "Andrew",
                        "middle": [],
                        "last": "Mccallum",
                        "suffix": ""
                    }
                ],
                "year": 2013,
                "venue": "Proceedings of the 2013 Workshop on Automated Knowledge Base Construction, AKBC '13",
                "volume": "",
                "issue": "",
                "pages": "79--84",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Limin Yao, Sebastian Riedel, and Andrew McCallum. 2013. Universal schema for entity type prediction. In Proceedings of the 2013 Workshop on Automated Knowledge Base Construction, AKBC '13, pages 79-84, New York, NY, USA. ACM.",
                "links": null
            }
        },
        "ref_entries": {
            "FIGREF0": {
                "num": null,
                "text": "Avg. F1 on development for increasing dimensions, using the low-rank model inFigure 1etrained with all seeds.",
                "uris": null,
                "type_str": "figure"
            },
            "FIGREF1": {
                "num": null,
                "text": "shows the parameter matrix of the low-Full parameter matrix of the low-rank model. The ticks in x-axis indicate the space for different entity types, while the ticks in y-axis indicate the space for different prefix context representations.a ll \u2212 c a p 1 = 1 a ll \u2212 c a p 1 = 0 a ll \u2212 c a p 2 = 1 a ll \u2212 c a p 2 = 0 a ll \u2212 lo w = 1 a ll \u2212 lo w = n b \u2212 to k = 1 n b \u2212 to k = 2 n b \u2212 to k > 2 c a p = 1 c a p = 0 d u m m y prefix=NNP (b) The subblock for PER entity type and PoS representation of the prefixes. The ticks in x-axis indicate the space of the entity features used, while the tick in y-axis indicates an example of a frequently observed prefix for this entity type.",
                "uris": null,
                "type_str": "figure"
            },
            "FIGREF2": {
                "num": null,
                "text": "Parameter matrix of the low-rank model in",
                "uris": null,
                "type_str": "figure"
            },
            "TABREF1": {
                "html": null,
                "type_str": "table",
                "text": "",
                "content": "<table/>",
                "num": null
            },
            "TABREF3": {
                "html": null,
                "type_str": "table",
                "text": "Average-F1 of classification of unseen entity candidates on development data, using the 10-30 training seed and 2 regularization, for different conjunctive spaces (elementary only, full conjunctions, all). Bag-of-words elementary features contain all clusters/PoS in separate windows to the left and to the right of the candidate. N-grams elementary features contain all n-grams of clusters/PoS in separate left and right windows (e.g. for size 3 it includes unigrams, bigrams and trigrams on each side).",
                "content": "<table/>",
                "num": null
            },
            "TABREF5": {
                "html": null,
                "type_str": "table",
                "text": "",
                "content": "<table><tr><td>compares different context represen-</td></tr><tr><td>tations and their performance in classifying un-</td></tr><tr><td>seen candidates using maximum-entropy classi-</td></tr><tr><td>fiers trained with Mallet (McCallum, 2002) with</td></tr></table>",
                "num": null
            },
            "TABREF6": {
                "html": null,
                "type_str": "table",
                "text": "",
                "content": "<table><tr><td>cap2 are from</td></tr></table>",
                "num": null
            },
            "TABREF7": {
                "html": null,
                "type_str": "table",
                "text": "Results on the test for models trained with different sizes of the seed, using the parameters and features that obtain the best evaluation results the development set. NN refers to nuclear norm regularization, L1 and L2 refer to 1 and 2 regularization. Only test entities unseen at training are considered. Avg. F1 is over PER, LOC, ORG and MISC, excluding O.",
                "content": "<table/>",
                "num": null
            }
        }
    }
}