/
fa18_sub_eng.srt
1499 lines (1125 loc) · 21.2 KB
/
fa18_sub_eng.srt
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
1
00:00:00,349 --> 00:00:02,210
Hello and welcome back to
2
00:00:02,220 --> 00:00:03,490
functional analysis.
3
00:00:03,670 --> 00:00:05,289
And as always many, many
4
00:00:05,300 --> 00:00:07,070
thanks to all the nice people
5
00:00:07,079 --> 00:00:08,449
that support me on Steady
6
00:00:08,460 --> 00:00:09,250
or paypal.
7
00:00:10,020 --> 00:00:11,390
We've already reached part
8
00:00:11,399 --> 00:00:12,829
18 in our course.
9
00:00:12,840 --> 00:00:14,090
And today we want to talk
10
00:00:14,100 --> 00:00:16,010
about compact operators.
11
00:00:16,719 --> 00:00:18,440
As the name suggests, this
12
00:00:18,450 --> 00:00:19,579
has something to do with
13
00:00:19,590 --> 00:00:21,389
the compact sets we've already
14
00:00:21,399 --> 00:00:21,899
studied.
15
00:00:22,780 --> 00:00:24,149
Therefore, recall that the
16
00:00:24,159 --> 00:00:25,989
idea of compactness has
17
00:00:26,000 --> 00:00:27,950
been to extend the notion
18
00:00:27,959 --> 00:00:29,469
of finite a little bit,
19
00:00:30,229 --> 00:00:31,709
a similar thing we will now
20
00:00:31,719 --> 00:00:33,220
do for operators
21
00:00:33,849 --> 00:00:34,310
here.
22
00:00:34,319 --> 00:00:36,080
The analogon of finite would
23
00:00:36,090 --> 00:00:37,799
be linear operators between
24
00:00:37,810 --> 00:00:39,139
finite dimensional vector
25
00:00:39,150 --> 00:00:39,799
spaces.
26
00:00:40,340 --> 00:00:41,880
So let's look at an operator
27
00:00:41,889 --> 00:00:43,340
from FN to
28
00:00:43,349 --> 00:00:45,319
FM on both sides,
29
00:00:45,330 --> 00:00:46,889
we can just choose the normal
30
00:00:46,900 --> 00:00:48,459
Euclidean norm, the standard
31
00:00:48,470 --> 00:00:48,880
norm.
32
00:00:49,319 --> 00:00:50,599
However, everything I say
33
00:00:50,610 --> 00:00:51,729
now does not depend on the
34
00:00:51,740 --> 00:00:52,220
norm.
35
00:00:52,229 --> 00:00:53,529
It holds no matter which
36
00:00:53,540 --> 00:00:54,529
norm you choose here.
37
00:00:55,389 --> 00:00:56,409
The first step is not so
38
00:00:56,419 --> 00:00:57,970
hard to show, it tells us
39
00:00:57,979 --> 00:00:59,150
that the linear operator
40
00:00:59,159 --> 00:01:00,590
between finite dimensional
41
00:01:00,599 --> 00:01:02,479
norm spaces is always
42
00:01:02,490 --> 00:01:03,340
continuous.
43
00:01:03,729 --> 00:01:04,888
And of course, you already
44
00:01:04,900 --> 00:01:06,330
know continuous means the
45
00:01:06,339 --> 00:01:07,889
same as bounded for linear
46
00:01:07,900 --> 00:01:08,669
operators.
47
00:01:09,330 --> 00:01:10,809
In other words, the image
48
00:01:10,819 --> 00:01:12,690
of the unit ball under T
49
00:01:12,699 --> 00:01:13,889
is a bounded set.
50
00:01:14,379 --> 00:01:15,839
Now, this is a bounded set
51
00:01:15,849 --> 00:01:17,269
in our finite dimensional
52
00:01:17,279 --> 00:01:18,769
norm space FM.
53
00:01:19,199 --> 00:01:20,610
And there you know this is
54
00:01:20,620 --> 00:01:22,089
one half of the things you
55
00:01:22,099 --> 00:01:23,430
need for being compact.
56
00:01:24,019 --> 00:01:25,510
However, the second ingredient
57
00:01:25,519 --> 00:01:26,989
we just get when we form
58
00:01:27,000 --> 00:01:28,410
the closure of this set,
59
00:01:28,819 --> 00:01:30,389
this means that we just can
60
00:01:30,400 --> 00:01:32,260
use a bar over the set.
61
00:01:32,989 --> 00:01:34,449
And then we get a set that
62
00:01:34,459 --> 00:01:36,330
is always compact in
63
00:01:36,339 --> 00:01:36,919
FM.
64
00:01:37,629 --> 00:01:38,989
With this essentially, you
65
00:01:39,000 --> 00:01:40,510
now know what a compact
66
00:01:40,519 --> 00:01:41,290
operators.
67
00:01:42,250 --> 00:01:43,470
When we look at the image
68
00:01:43,480 --> 00:01:45,139
of the unit ball and form
69
00:01:45,150 --> 00:01:45,819
the closure.
70
00:01:45,830 --> 00:01:47,309
And this set is compact,
71
00:01:47,319 --> 00:01:49,279
we speak of a compact operator.
72
00:01:49,830 --> 00:01:51,349
Of course, this always holds
73
00:01:51,360 --> 00:01:52,910
in this case, but not in
74
00:01:52,919 --> 00:01:53,500
general.
75
00:01:54,099 --> 00:01:55,510
In order to see that let's
76
00:01:55,519 --> 00:01:56,730
consider the identity
77
00:01:56,739 --> 00:01:57,580
operator.
78
00:01:57,589 --> 00:01:58,330
In LP.
79
00:01:59,110 --> 00:02:00,610
Of course, identity operator
80
00:02:00,620 --> 00:02:02,309
just means we take a sequence
81
00:02:02,319 --> 00:02:03,980
X and send it to itself.
82
00:02:04,449 --> 00:02:05,779
Therefore, calculating the
83
00:02:05,790 --> 00:02:07,099
image of the unit ball is
84
00:02:07,110 --> 00:02:08,660
not so hard, it stays the
85
00:02:08,669 --> 00:02:09,380
unit ball.
86
00:02:10,119 --> 00:02:11,160
Now we are interested in
87
00:02:11,169 --> 00:02:12,259
the closure of the set.
88
00:02:12,270 --> 00:02:13,779
So we use the bars again.
89
00:02:14,490 --> 00:02:15,580
So what you have on the right
90
00:02:15,589 --> 00:02:17,009
hand side is just a closed
91
00:02:17,029 --> 00:02:18,429
unit ball in LP.
92
00:02:19,089 --> 00:02:20,169
And there we know from the
93
00:02:20,179 --> 00:02:21,820
last video it's closed and
94
00:02:21,830 --> 00:02:23,789
bounded but not compact.
95
00:02:24,509 --> 00:02:25,570
Hence, we have something
96
00:02:25,580 --> 00:02:25,809
here.
97
00:02:25,820 --> 00:02:27,350
We would not call a compact
98
00:02:27,360 --> 00:02:29,059
operator because the image
99
00:02:29,070 --> 00:02:30,509
is just too large in this
100
00:02:30,520 --> 00:02:30,929
case.
101
00:02:31,679 --> 00:02:33,229
So you should always remember
102
00:02:33,240 --> 00:02:34,910
compact operators is what
103
00:02:34,919 --> 00:02:36,589
we get when we extend the
104
00:02:36,600 --> 00:02:38,059
finite dimensional operators
105
00:02:38,070 --> 00:02:39,070
here a little bit.
106
00:02:39,949 --> 00:02:41,559
OK, then let's write down
107
00:02:41,570 --> 00:02:42,740
the formal definition.
108
00:02:43,210 --> 00:02:45,139
So we need two norm spaces
109
00:02:45,149 --> 00:02:46,789
and often there will be Banach
110
00:02:46,800 --> 00:02:48,449
spaces then a
111
00:02:48,460 --> 00:02:49,910
bounded linear operator
112
00:02:49,919 --> 00:02:51,500
T from X to
113
00:02:51,509 --> 00:02:52,720
Y is called
114
00:02:52,729 --> 00:02:53,559
compact.
115
00:02:54,050 --> 00:02:55,720
If we have the thing we discussed
116
00:02:55,729 --> 00:02:57,199
before that this set
117
00:02:57,210 --> 00:02:58,880
here is compact.
118
00:02:58,889 --> 00:03:00,699
In Y therefore, in the
119
00:03:00,710 --> 00:03:02,509
case that Y is a finer
120
00:03:02,520 --> 00:03:04,250
dimensional norm space, this
121
00:03:04,259 --> 00:03:05,559
here is nothing special.
122
00:03:06,250 --> 00:03:07,320
However, in the infinite
123
00:03:07,330 --> 00:03:08,860
dimensional case, it really
124
00:03:08,869 --> 00:03:09,309
is.
125
00:03:09,949 --> 00:03:11,309
Hence, I would suggest that
126
00:03:11,320 --> 00:03:12,710
we now look at a common
127
00:03:12,720 --> 00:03:14,559
example, let's
128
00:03:14,570 --> 00:03:16,139
look at an integral operator
129
00:03:16,149 --> 00:03:17,750
defined for the continuous
130
00:03:17,759 --> 00:03:18,389
functions.
131
00:03:18,960 --> 00:03:20,179
So it should take a function
132
00:03:20,190 --> 00:03:22,149
defined from 0 to 1 and
133
00:03:22,160 --> 00:03:23,850
send that to another function
134
00:03:23,860 --> 00:03:24,710
in the same space.
135
00:03:25,360 --> 00:03:26,820
And as often the space of
136
00:03:26,830 --> 00:03:28,139
continuous functions should
137
00:03:28,149 --> 00:03:29,740
carry the supreme norm.
138
00:03:30,509 --> 00:03:32,270
So what we can do is apply
139
00:03:32,279 --> 00:03:33,960
the operator T to a function
140
00:03:33,970 --> 00:03:35,880
F and then we get out a new
141
00:03:35,889 --> 00:03:36,520
function.
142
00:03:36,979 --> 00:03:38,169
Hence, we can look what the
143
00:03:38,179 --> 00:03:39,410
function does at a given
144
00:03:39,419 --> 00:03:41,410
point S where S comes from
145
00:03:41,419 --> 00:03:42,330
the unit interval.
146
00:03:43,029 --> 00:03:44,210
Now the number that comes
147
00:03:44,220 --> 00:03:45,449
out here should be given
148
00:03:45,460 --> 00:03:47,220
by an integral from 0 to
149
00:03:47,229 --> 00:03:49,190
1 where we have the function
150
00:03:49,199 --> 00:03:50,690
F involved, but
151
00:03:50,699 --> 00:03:52,570
also a fixed function
152
00:03:52,580 --> 00:03:53,050
K.
153
00:03:53,550 --> 00:03:54,570
And this function should
154
00:03:54,580 --> 00:03:56,500
get the variable S and also
155
00:03:56,509 --> 00:03:57,550
the integration variable
156
00:03:57,559 --> 00:03:57,949
T.
157
00:03:58,720 --> 00:03:59,949
So we have a function with
158
00:03:59,960 --> 00:04:01,539
two variables and for us
159
00:04:01,550 --> 00:04:03,270
it should be also a continuous
160
00:04:03,279 --> 00:04:03,710
function.
161
00:04:04,520 --> 00:04:06,449
So we have it from C defined
162
00:04:06,460 --> 00:04:08,350
on the unit interval times
163
00:04:08,360 --> 00:04:09,330
the unit interval.
164
00:04:10,100 --> 00:04:11,360
Now, since the function K
165
00:04:11,369 --> 00:04:12,759
goes into the definition
166
00:04:12,770 --> 00:04:14,710
of T I put it into the
167
00:04:14,720 --> 00:04:15,509
index here.
168
00:04:16,160 --> 00:04:17,690
OK, then let's check if
169
00:04:17,700 --> 00:04:19,059
TK is indeed a
170
00:04:19,070 --> 00:04:20,339
compact operator.
171
00:04:21,170 --> 00:04:22,429
An important fact we will
172
00:04:22,440 --> 00:04:24,230
need here is that the function
173
00:04:24,239 --> 00:04:25,609
K is defined on a
174
00:04:25,619 --> 00:04:26,709
compact set.
175
00:04:26,720 --> 00:04:28,709
So it's not just continuous,
176
00:04:28,720 --> 00:04:30,660
it's uniformly continuous
177
00:04:31,420 --> 00:04:32,730
to refresh your memory.
178
00:04:32,739 --> 00:04:34,059
Let's write down what this
179
00:04:34,070 --> 00:04:34,929
exactly means.
180
00:04:35,720 --> 00:04:37,540
For all epsilon crater zero,
181
00:04:37,549 --> 00:04:39,260
there exists a delta crater
182
00:04:39,269 --> 00:04:41,140
zero such that for
183
00:04:41,149 --> 00:04:42,859
all points we put in and
184
00:04:42,869 --> 00:04:44,440
now we need two variables.
185
00:04:45,010 --> 00:04:46,339
And for them, it should hold
186
00:04:46,350 --> 00:04:47,779
if the distance is less than
187
00:04:47,790 --> 00:04:49,420
delta, the distance of the
188
00:04:49,429 --> 00:04:50,820
images should be less than
189
00:04:50,829 --> 00:04:51,589
epsilon.
190
00:04:52,059 --> 00:04:53,100
On the left hand side, we
191
00:04:53,109 --> 00:04:54,170
measure the distance with
192
00:04:54,179 --> 00:04:55,720
the Euclidean norm in R two.
193
00:04:55,929 --> 00:04:57,100
And on the right hand side,
194
00:04:57,109 --> 00:04:58,399
we measure with the Euclidean
195
00:04:58,410 --> 00:05:00,239
norm in R one, which is the
196
00:05:00,250 --> 00:05:00,940
absolute value.
197
00:05:01,640 --> 00:05:03,600
Why we need the uniform continuity
198
00:05:03,609 --> 00:05:05,299
here we see in a moment.
199
00:05:05,869 --> 00:05:06,859
The first step we have to
200
00:05:06,869 --> 00:05:08,540
do when we see such an integral
201
00:05:08,549 --> 00:05:10,399
operator is to check if
202
00:05:10,410 --> 00:05:12,220
this integral defines
203
00:05:12,230 --> 00:05:14,140
indeed a continuous function.
204
00:05:15,029 --> 00:05:16,709
Simply because otherwise
205
00:05:16,720 --> 00:05:18,029
the operator wouldn't be
206
00:05:18,040 --> 00:05:18,850
well defined.
207
00:05:19,489 --> 00:05:21,309
It really should map continuous
208
00:05:21,320 --> 00:05:22,760
functions to continuous
209
00:05:22,769 --> 00:05:23,440
functions.
210
00:05:23,850 --> 00:05:25,140
Checking the continuity.
211
00:05:25,149 --> 00:05:26,549
Then means we look at the
212
00:05:26,559 --> 00:05:28,029
difference of the images
213
00:05:28,040 --> 00:05:29,459
when we put in different
214
00:05:29,470 --> 00:05:29,989
points.
215
00:05:30,660 --> 00:05:32,059
So this should be small when
216
00:05:32,070 --> 00:05:33,579
the points as one and two
217
00:05:33,589 --> 00:05:34,630
are close together.
218
00:05:35,399 --> 00:05:37,100
Therefore, we first calculate
219
00:05:37,109 --> 00:05:38,290
and then we can look what
220
00:05:38,299 --> 00:05:38,859
happens?
221
00:05:39,750 --> 00:05:40,730
Now, the first thing you
222
00:05:40,739 --> 00:05:41,769
should see here is that we
223
00:05:41,779 --> 00:05:43,220
can put that into one
224
00:05:43,230 --> 00:05:43,809
integral.
225
00:05:44,649 --> 00:05:46,440
So use some parentheses here
226
00:05:46,450 --> 00:05:48,119
and delete this integral
227
00:05:48,130 --> 00:05:48,399
here.
228
00:05:49,299 --> 00:05:50,640
Then of course, we pull in
229
00:05:50,649 --> 00:05:52,239
the absolute value then we
230
00:05:52,250 --> 00:05:53,279
get an inequality.
231
00:05:54,130 --> 00:05:55,589
And with this, we have everything
232
00:05:55,600 --> 00:05:57,369
we need because this one
233
00:05:57,380 --> 00:05:59,000
here is less than a supremum
234
00:05:59,059 --> 00:05:59,989
norm of F.
235
00:06:00,140 --> 00:06:01,899
And this one by the uniform
236
00:06:01,910 --> 00:06:03,859
continuity of K can be as
237
00:06:03,869 --> 00:06:05,149
small as we want.
238
00:06:05,890 --> 00:06:07,390
And exactly this is what
239
00:06:07,399 --> 00:06:08,670
we should formally write
240
00:06:08,679 --> 00:06:10,230
in front of the whole calculation.
241
00:06:10,899 --> 00:06:12,559
So for a given epsilon greater
242
00:06:12,570 --> 00:06:13,700
than zero, we choose the
243
00:06:13,709 --> 00:06:15,309
delta in such a way that
244
00:06:15,320 --> 00:06:16,739
this whole thing holds.
245
00:06:17,390 --> 00:06:18,440
Therefore, we can choose
246
00:06:18,450 --> 00:06:20,000
as one as two from the unit
247
00:06:20,010 --> 00:06:21,799
interval such that the distance
248
00:06:21,809 --> 00:06:22,950
is less than delta
249
00:06:23,510 --> 00:06:25,489
for this as one as two here.
250
00:06:25,500 --> 00:06:27,420
And the same T on both sides,