971:, when used with a logarithmic cooling schedule, has been proven to find the global optimum of any optimization problem. However, such a cooling schedule results in entirely impractical runtimes, and is never used. However, knowing this ideal algorithm exists has led to practical variants that are able to find very good (though not provably optimal) solutions to complex optimization problems.
948:
correctness for each algorithm. Since the proof of correctness is of finite size, it "only" adds a constant and does not affect the asymptotic runtime. However, this constant is so big that the algorithm is entirely impractical. For example, if the shortest proof of correctness of a given algorithm is 1000 bits long, the search will examine at least 2 other potential proofs first.
49:
This alone could be important and often is a great reason for finding such algorithms. For example, if tomorrow there were a discovery that showed there is a factoring algorithm with a huge but provably polynomial time bound, that would change our beliefs about factoring. The algorithm might never be
1123:
Researchers have found an algorithm that achieves the provably best-possible asymptotic performance in terms of time-space tradeoff. But it remains purely theoretical: "Despite the new hash table’s unprecedented efficiency, no one is likely to try building it anytime soon. It’s just too complicated
947:
A single algorithm, "Hutter search", can solve any well-defined problem in an asymptotically optimal time, barring some caveats. It works by searching through all possible algorithms (by runtime), while simultaneously searching through all possible proofs (by length of proof), looking for a proof of
23:
is one with record-breaking theoretical (asymptotic) performance, but which is never used in practice. Typical reasons are that the performance gains only appear for problems that are so large they never occur, or the algorithm's complexity outweighs a relatively small gain in performance. Galactic
959:
theories (as implemented by programs) which perfectly describe previous observations are used to calculate the probability of the next observation, with more weight put on the shorter computable theories. Again, the search over all possible explanations makes this procedure galactic.
180:
are large, it is never used in practice. However, it also shows why galactic algorithms may still be useful. The authors state: "we are hopeful that with further refinements, the algorithm might become practical for numbers with merely billions or trillions of digits."
305:
multiplications. These are galactic – "We nevertheless stress that such improvements are only of theoretical interest, since the huge constants involved in the complexity of fast matrix multiplication usually make these algorithms impractical."
835:
938:
percent. Although no one will ever switch to this algorithm for its very slight worst-case improvement, it is still considered important because "this minuscule improvement breaks through both a theoretical logjam and a psychological one".
853:– i.e., performing one trial decryption for each possible key. For many cryptographic systems, breaks are known, but are still practically infeasible with current technology. One example is the best attack known against 128-bit
385:
big enough to beat existing codes is also completely impractical. These codes, though never used, inspired decades of research into more practical algorithms that today can achieve rates arbitrarily close to channel capacity.
668:
106:
1066:
is huge enough to make the algorithm impractical. An implementation is publicly available and given the experimentally estimated implementation constants, it would only be faster than
46:
An impractical algorithm can still demonstrate that conjectured bounds can be achieved, or that proposed bounds are wrong, and hence advance the theory of algorithms. As Lipton states:
1113:
303:
263:
747:
174:
1764:
936:
540:
223:
882:
1696:"A C++ implementation for an Expected Linear-Time Minimum Spanning Tree Algorithm(Karger-Klein-Tarjan + Hagerup Minimum Spanning Tree Verification as a sub-routine)"
1020:
1331:
884:
operations. Despite being impractical, theoretical breaks can provide insight into vulnerability patterns, and sometimes lead to discovery of exploitable breaks.
738:
1060:
1040:
712:
692:
604:
580:
560:
504:
484:
464:
436:
416:
383:
363:
343:
952:
1575:
Liang, Faming; Cheng, Yichen; Lin, Guang (2014). "Simulated stochastic approximation annealing for global optimization with a square-root cooling schedule".
1778:
Bender, Michael; Farach-Colton, Martin; Kuszmaul, John; Kuszmaul, William; Mingmou, Liu (4 Nov 2021). "On the
Optimal Time/Space Tradeoff for Hash Tables".
609:
265:
multiplications. This algorithm is not galactic and is used in practice. Further extensions of this, using sophisticated group theory, are the
1449:
1293:
1465:
Anna R. Karlin; Nathan Klein; Shayan Oveis Gharan (September 1, 2020). "A (Slightly) Improved
Approximation Algorithm for Metric TSP".
365:
is chosen large enough, this beats any existing code and can get arbitrarily close to the capacity of the channel. Unfortunately, any
266:
980:
908:
do much better, but could not provably do so.) In 2020, a newer and much more complex algorithm was discovered that can beat this by
1814:
1799:
1160:
1487:
43:
Available computational power may catch up to the crossover point, so that a previously impractical algorithm becomes practical.
1844:
1366:
1834:
1241:
54:
60:
1839:
854:
671:
1747:
40:
An algorithm, even if impractical, may show new techniques that may eventually be used to create practical algorithms.
1357:
893:
113:
1067:
131:
1124:
to construct." and "in practice, constants really matter. In the real world, a factor of 10 is a game ender.”
1073:
1623:
1404:
901:
984:
322:
36:
Even if they are never used in practice, galactic algorithms may still contribute to computer science:
1539:
1409:
1628:
968:
830:{\displaystyle {^{65536}2=\ \atop {\ }}{{\underbrace {2^{2^{\cdot ^{\cdot ^{2}}}}} } \atop 65536}}
109:
272:
232:
1779:
1592:
1506:
1505:
Hutter, Marcus (2002-06-14). "The
Fastest and Shortest Algorithm for All Well-Defined Problems".
1466:
1342:
1299:
1271:
1200:
850:
226:
141:
1758:
1676:
1557:
1445:
1289:
1156:
1148:
911:
509:
192:
135:
1177:
860:
1666:
1633:
1584:
1547:
1437:
1414:
1375:
1281:
1268:
Proceedings of the 53rd Annual IEEE Symposium on
Foundations of Computer Science (FOCS 2012)
1192:
990:
395:
1695:
1219:
717:
1543:
1062:
is the number of nodes of the graph. However, the constant factor that is hidden by the
904:
which produced a path at most 50% longer than the optimum. (Many other algorithms could
1395:
Johnson, David S. (1987). "The NP-completeness column: An ongoing guide (edition 19)".
1144:
1063:
1045:
1025:
697:
677:
589:
583:
565:
545:
489:
469:
449:
421:
401:
368:
348:
328:
314:
177:
25:
1722:
1828:
1638:
1611:
1596:
1418:
466:
is fixed, it can be solved in polynomial time. The running time for testing whether
1317:
1303:
1204:
897:
846:
663:{\displaystyle 2\uparrow \uparrow (2\uparrow \uparrow (2\uparrow \uparrow (h/2)))}
1588:
1441:
443:
439:
112:, considered the most important open problem in computer science and one of the
1380:
1361:
1266:
Le Gall, F. (2012), "Faster algorithms for rectangular matrix multiplication",
1552:
1527:
956:
1680:
1561:
1196:
1671:
1654:
189:
The first improvement over brute-force matrix multiplication (which needs
28:
and Ken Regan, because they will never be used on any data sets on Earth.
1285:
741:
1700:
1436:. Lecture Notes in Computer Science. Vol. 9144. pp. 39–56.
1653:
Karger, David R.; Klein, Philip N.; Tarjan, Robert E. (1995-03-01).
1784:
1655:"A randomized linear-time algorithm to find minimum spanning trees"
1511:
1471:
1153:
People, Problems, and Proofs: Essays from Gödel's Lost Letter: 2010
50:
used, but would certainly shape the future research into factoring.
1276:
345:-bit message, then decoding by finding the closest code word. If
1748:"Tight Cell-Probe Lower Bounds for Dynamic Succinct Dictionaries"
740:
cannot be reasonably computed as the constant is greater than 2
318:
130:
An example of a galactic algorithm is the fastest known way to
325:. It requires assigning a random code word to every possible
849:
jargon, a "break" is any attack faster in expectation than
1815:"Scientists Find Optimal Balance of Data Storage and Time"
1800:"Scientists Find Optimal Balance of Data Storage and Time"
1746:
Tianxiao Li, Jingxun Liang, Huacheng Yu, and Renfei Zhou.
1242:"We've found a quicker way to multiply really big numbers"
892:
For several decades, the best known approximation to the
1488:"Computer Scientists Break Traveling Salesperson Record"
955:, which is a formalization of Bayesian inference. All
1076:
1048:
1028:
993:
914:
863:
750:
720:
700:
680:
612:
592:
568:
548:
512:
492:
472:
452:
424:
404:
371:
351:
331:
275:
235:
195:
144:
63:
586:
hides a constant that depends superexponentially on
101:{\displaystyle \Theta {\bigl (}n^{2^{100}}{\bigr )}}
1218:David, Harvey; Hoeven, Joris van der (March 2019).
176:bit operations, but as the constants hidden by the
1107:
1054:
1034:
1014:
930:
876:
829:
732:
706:
686:
662:
598:
574:
554:
534:
498:
478:
458:
430:
410:
377:
357:
337:
297:
257:
217:
168:
108:, although unusable in practice, would settle the
100:
1155:. Heidelberg: Springer Berlin. pp. 109–112.
57:with a large but polynomial time bound, such as
1577:Journal of the American Statistical Association
47:
1362:"The disjoint paths problem in quadratic time"
1723:"Expected Linear-Time Minimum Spanning Trees"
1612:"Simulated annealing: Practice versus theory"
321:that can reach the theoretical capacity of a
93:
69:
8:
1763:: CS1 maint: multiple names: authors list (
1356:Kawarabayashi, Ken-ichi; Kobayashi, Yusuke;
269:and its slightly better successors, needing
53:Similarly, a hypothetical algorithm for the
1332:"Capacity-approaching codes (Chapter 13 of
1220:"Integer multiplication in time O(n log n)"
317:showed a simple but asymptotically optimal
1783:
1670:
1637:
1627:
1551:
1510:
1470:
1408:
1379:
1275:
1139:
1137:
1099:
1075:
1047:
1027:
992:
919:
913:
868:
862:
804:
799:
794:
789:
783:
782:
780:
773:
757:
751:
749:
719:
699:
679:
643:
611:
591:
567:
547:
523:
511:
491:
471:
451:
423:
403:
370:
350:
330:
286:
274:
246:
234:
206:
194:
143:
92:
91:
83:
78:
68:
67:
62:
1432:Biaoshuai Tao & Hongjun Wu (2015).
1235:
1233:
1178:"The status of the P versus NP problem"
1133:
134:, which is based on a 1729-dimensional
1756:
1334:Principles Of Digital Communication II
1108:{\displaystyle m+n>9\cdot 10^{151}}
7:
1149:"David Johnson: Galactic Algorithms"
1616:Mathematical and Computer Modelling
1486:Klarreich, Erica (8 October 2020).
1316:Larry Hardesty (January 19, 2010).
229:: a recursive algorithm that needs
981:expected linear time MST algorithm
781:
752:
64:
14:
1434:Information Security and Privacy
606:. The constant is greater than
1812:William Kuszmaul, as quoted in
1526:Gagliolo, Matteo (2007-11-20).
1367:Journal of Combinatorial Theory
1318:"Explained: The Shannon limit"
1240:Harvey, David (9 April 2019).
1009:
997:
657:
654:
651:
637:
634:
628:
625:
619:
616:
529:
516:
310:Communication channel capacity
292:
279:
267:Coppersmith–Winograd algorithm
252:
239:
212:
199:
163:
148:
55:Boolean satisfiability problem
1:
694:is the number of vertices in
562:is the number of vertices in
1639:10.1016/0895-7177(93)90204-C
1589:10.1080/01621459.2013.872993
1419:10.1016/0196-6774(87)90043-5
1147:; Regan, Kenneth W. (2013).
951:Hutter search is related to
298:{\displaystyle O(n^{2.373})}
258:{\displaystyle O(n^{2.807})}
24:algorithms were so named by
1721:Geiman Thiesen, Francisco.
1442:10.1007/978-3-319-19962-7_3
1042:is the number of edges and
16:Classification of algorithm
1861:
1727:franciscothiesen.github.io
1381:10.1016/j.jctb.2011.07.004
894:traveling salesman problem
888:Traveling salesman problem
169:{\displaystyle O(n\log n)}
1553:10.4249/scholarpedia.2575
1185:Communications of the ACM
672:Knuth's up-arrow notation
225:multiplications) was the
114:Millennium Prize Problems
983:is able to discover the
931:{\displaystyle 10^{-34}}
535:{\displaystyle O(n^{2})}
218:{\displaystyle O(n^{3})}
1610:Ingber, Lester (1993).
1197:10.1145/1562164.1562186
877:{\displaystyle 2^{126}}
1845:Analysis of algorithms
1109:
1056:
1036:
1016:
1015:{\displaystyle O(m+n)}
975:Minimum spanning trees
932:
902:Christofides algorithm
878:
831:
734:
708:
688:
664:
600:
576:
556:
536:
500:
480:
460:
446:in general, but where
432:
412:
379:
359:
339:
299:
259:
219:
170:
126:Integer multiplication
102:
52:
1835:Mathematical notation
1672:10.1145/201019.201022
1397:Journal of Algorithms
1110:
1057:
1037:
1017:
985:minimum spanning tree
933:
879:
832:
735:
709:
689:
665:
601:
577:
557:
537:
501:
481:
461:
433:
413:
380:
360:
340:
323:communication channel
300:
260:
220:
185:Matrix multiplication
171:
103:
1694:Thiesen, Francisco.
1286:10.1109/FOCS.2012.80
1270:, pp. 514–523,
1176:Fortnow, L. (2009).
1074:
1070:for graphs in which
1046:
1026:
991:
953:Solomonoff induction
912:
900:was the very simple
861:
841:Cryptographic breaks
748:
718:
698:
678:
610:
590:
566:
546:
510:
490:
470:
450:
422:
402:
369:
349:
329:
273:
233:
193:
142:
132:multiply two numbers
61:
1840:Asymptotic analysis
1544:2007SchpJ...2.2575G
1068:BorĹŻvka's algorithm
969:Simulated annealing
857:, which takes only
744:by 65536, that is,
733:{\displaystyle h=4}
714:. Even the case of
110:P versus NP problem
1659:Journal of the ACM
1528:"Universal search"
1343:MIT OpenCourseWare
1320:. MIT News Office.
1145:Lipton, Richard J.
1105:
1052:
1032:
1012:
928:
874:
827:
819:
730:
704:
684:
660:
635:↑ ↑
626:↑ ↑
617:↑ ↑
596:
572:
552:
532:
496:
476:
456:
428:
408:
375:
355:
335:
295:
255:
227:Strassen algorithm
215:
166:
98:
32:Possible use cases
21:galactic algorithm
1451:978-3-319-19961-0
1295:978-0-7695-4874-6
1055:{\displaystyle n}
1035:{\displaystyle m}
825:
784:
778:
776:
771:
707:{\displaystyle H}
687:{\displaystyle h}
599:{\displaystyle H}
575:{\displaystyle G}
555:{\displaystyle n}
499:{\displaystyle G}
479:{\displaystyle H}
459:{\displaystyle H}
431:{\displaystyle H}
411:{\displaystyle G}
378:{\displaystyle n}
358:{\displaystyle n}
338:{\displaystyle n}
136:Fourier transform
1852:
1819:
1818:
1810:
1804:
1803:
1796:
1790:
1789:
1787:
1775:
1769:
1768:
1762:
1754:
1752:
1743:
1737:
1736:
1734:
1733:
1718:
1712:
1711:
1709:
1708:
1691:
1685:
1684:
1674:
1650:
1644:
1643:
1641:
1631:
1607:
1601:
1600:
1583:(506): 847–863.
1572:
1566:
1565:
1555:
1523:
1517:
1516:
1514:
1502:
1496:
1495:
1483:
1477:
1476:
1474:
1462:
1456:
1455:
1429:
1423:
1422:
1412:
1392:
1386:
1385:
1383:
1353:
1347:
1346:
1340:
1328:
1322:
1321:
1313:
1307:
1306:
1279:
1263:
1257:
1256:
1254:
1252:
1246:The Conversation
1237:
1228:
1227:
1215:
1209:
1208:
1182:
1173:
1167:
1166:
1141:
1114:
1112:
1111:
1106:
1104:
1103:
1061:
1059:
1058:
1053:
1041:
1039:
1038:
1033:
1021:
1019:
1018:
1013:
937:
935:
934:
929:
927:
926:
883:
881:
880:
875:
873:
872:
836:
834:
833:
828:
826:
821:
820:
815:
814:
813:
812:
811:
810:
809:
808:
779:
777:
774:
772:
769:
762:
761:
739:
737:
736:
731:
713:
711:
710:
705:
693:
691:
690:
685:
669:
667:
666:
661:
647:
605:
603:
602:
597:
581:
579:
578:
573:
561:
559:
558:
553:
541:
539:
538:
533:
528:
527:
506:in this case is
505:
503:
502:
497:
485:
483:
482:
477:
465:
463:
462:
457:
437:
435:
434:
429:
417:
415:
414:
409:
398:whether a graph
384:
382:
381:
376:
364:
362:
361:
356:
344:
342:
341:
336:
304:
302:
301:
296:
291:
290:
264:
262:
261:
256:
251:
250:
224:
222:
221:
216:
211:
210:
175:
173:
172:
167:
107:
105:
104:
99:
97:
96:
90:
89:
88:
87:
73:
72:
1860:
1859:
1855:
1854:
1853:
1851:
1850:
1849:
1825:
1824:
1823:
1822:
1813:
1811:
1807:
1798:
1797:
1793:
1777:
1776:
1772:
1755:
1750:
1745:
1744:
1740:
1731:
1729:
1720:
1719:
1715:
1706:
1704:
1693:
1692:
1688:
1652:
1651:
1647:
1609:
1608:
1604:
1574:
1573:
1569:
1525:
1524:
1520:
1504:
1503:
1499:
1492:Quanta Magazine
1485:
1484:
1480:
1464:
1463:
1459:
1452:
1431:
1430:
1426:
1410:10.1.1.114.3864
1394:
1393:
1389:
1355:
1354:
1350:
1338:
1330:
1329:
1325:
1315:
1314:
1310:
1296:
1265:
1264:
1260:
1250:
1248:
1239:
1238:
1231:
1226:. hal-02070778.
1217:
1216:
1212:
1180:
1175:
1174:
1170:
1163:
1143:
1142:
1135:
1130:
1121:
1095:
1072:
1071:
1044:
1043:
1024:
1023:
989:
988:
977:
966:
945:
915:
910:
909:
890:
864:
859:
858:
843:
800:
795:
790:
785:
754:
753:
746:
745:
716:
715:
696:
695:
676:
675:
608:
607:
588:
587:
564:
563:
544:
543:
519:
508:
507:
488:
487:
468:
467:
448:
447:
420:
419:
400:
399:
394:The problem of
392:
367:
366:
347:
346:
327:
326:
312:
282:
271:
270:
242:
231:
230:
202:
191:
190:
187:
140:
139:
128:
123:
79:
74:
59:
58:
34:
17:
12:
11:
5:
1858:
1856:
1848:
1847:
1842:
1837:
1827:
1826:
1821:
1820:
1805:
1791:
1770:
1738:
1713:
1686:
1665:(2): 321–328.
1645:
1629:10.1.1.15.1046
1602:
1567:
1518:
1497:
1478:
1457:
1450:
1424:
1403:(2): 285–303.
1387:
1374:(2): 424–435.
1348:
1323:
1308:
1294:
1258:
1229:
1210:
1168:
1161:
1132:
1131:
1129:
1126:
1120:
1117:
1102:
1098:
1094:
1091:
1088:
1085:
1082:
1079:
1064:Big O notation
1051:
1031:
1011:
1008:
1005:
1002:
999:
996:
987:of a graph in
976:
973:
965:
962:
944:
941:
925:
922:
918:
889:
886:
871:
867:
842:
839:
824:
818:
807:
803:
798:
793:
788:
768:
765:
760:
756:
729:
726:
723:
703:
683:
659:
656:
653:
650:
646:
642:
639:
636:
633:
630:
627:
624:
621:
618:
615:
595:
584:big O notation
571:
551:
531:
526:
522:
518:
515:
495:
486:is a minor of
475:
455:
427:
407:
391:
388:
374:
354:
334:
315:Claude Shannon
311:
308:
294:
289:
285:
281:
278:
254:
249:
245:
241:
238:
214:
209:
205:
201:
198:
186:
183:
178:big O notation
165:
162:
159:
156:
153:
150:
147:
127:
124:
122:
119:
118:
117:
95:
86:
82:
77:
71:
66:
44:
41:
33:
30:
26:Richard Lipton
15:
13:
10:
9:
6:
4:
3:
2:
1857:
1846:
1843:
1841:
1838:
1836:
1833:
1832:
1830:
1816:
1809:
1806:
1801:
1795:
1792:
1786:
1781:
1774:
1771:
1766:
1760:
1749:
1742:
1739:
1728:
1724:
1717:
1714:
1703:
1702:
1697:
1690:
1687:
1682:
1678:
1673:
1668:
1664:
1660:
1656:
1649:
1646:
1640:
1635:
1630:
1625:
1622:(11): 29–57.
1621:
1617:
1613:
1606:
1603:
1598:
1594:
1590:
1586:
1582:
1578:
1571:
1568:
1563:
1559:
1554:
1549:
1545:
1541:
1537:
1533:
1529:
1522:
1519:
1513:
1508:
1501:
1498:
1493:
1489:
1482:
1479:
1473:
1468:
1461:
1458:
1453:
1447:
1443:
1439:
1435:
1428:
1425:
1420:
1416:
1411:
1406:
1402:
1398:
1391:
1388:
1382:
1377:
1373:
1369:
1368:
1363:
1359:
1352:
1349:
1344:
1337:
1335:
1327:
1324:
1319:
1312:
1309:
1305:
1301:
1297:
1291:
1287:
1283:
1278:
1273:
1269:
1262:
1259:
1247:
1243:
1236:
1234:
1230:
1225:
1221:
1214:
1211:
1206:
1202:
1198:
1194:
1190:
1186:
1179:
1172:
1169:
1164:
1162:9783642414220
1158:
1154:
1150:
1146:
1140:
1138:
1134:
1127:
1125:
1118:
1116:
1100:
1096:
1092:
1089:
1086:
1083:
1080:
1077:
1069:
1065:
1049:
1029:
1006:
1003:
1000:
994:
986:
982:
974:
972:
970:
963:
961:
958:
954:
949:
943:Hutter search
942:
940:
923:
920:
916:
907:
903:
899:
895:
887:
885:
869:
865:
856:
852:
848:
840:
838:
822:
816:
805:
801:
796:
791:
786:
766:
763:
758:
755:
743:
727:
724:
721:
701:
681:
673:
648:
644:
640:
631:
622:
613:
593:
585:
569:
549:
524:
520:
513:
493:
473:
453:
445:
441:
425:
405:
397:
389:
387:
372:
352:
332:
324:
320:
316:
309:
307:
287:
283:
276:
268:
247:
243:
236:
228:
207:
203:
196:
184:
182:
179:
160:
157:
154:
151:
145:
137:
133:
125:
120:
115:
111:
84:
80:
75:
56:
51:
45:
42:
39:
38:
37:
31:
29:
27:
22:
1808:
1794:
1773:
1741:
1730:. Retrieved
1726:
1716:
1705:. Retrieved
1699:
1689:
1662:
1658:
1648:
1619:
1615:
1605:
1580:
1576:
1570:
1538:(11): 2575.
1535:
1532:Scholarpedia
1531:
1521:
1500:
1491:
1481:
1460:
1433:
1427:
1400:
1396:
1390:
1371:
1370:. Series B.
1365:
1351:
1333:
1326:
1311:
1267:
1261:
1249:. Retrieved
1245:
1223:
1213:
1191:(9): 78–86.
1188:
1184:
1171:
1152:
1122:
978:
967:
964:Optimization
950:
946:
905:
898:metric space
891:
847:cryptography
844:
393:
313:
188:
129:
48:
35:
20:
18:
1358:Reed, Bruce
1119:Hash tables
851:brute force
444:NP-complete
138:. It needs
1829:Categories
1785:2111.00602
1732:2022-11-13
1707:2022-11-19
1512:cs/0206022
1472:2007.01409
1128:References
957:computable
390:Sub-graphs
1681:0004-5411
1624:CiteSeerX
1597:123410795
1562:1941-6016
1405:CiteSeerX
1277:1204.1111
1093:⋅
921:−
817:⏟
802:⋅
797:⋅
418:contains
158:
65:Θ
1759:cite web
1360:(2012).
1022:, where
742:tetrated
674:, where
582:and the
542:, where
396:deciding
121:Examples
1540:Bibcode
1345:. 2005.
1304:2410545
1251:9 March
1205:5969255
906:usually
1701:GitHub
1679:
1626:
1595:
1560:
1448:
1407:
1302:
1292:
1203:
1159:
775:
770:
1780:arXiv
1751:(PDF)
1593:S2CID
1507:arXiv
1467:arXiv
1339:(PDF)
1300:S2CID
1272:arXiv
1201:S2CID
1181:(PDF)
896:in a
823:65536
759:65536
440:minor
438:as a
288:2.373
248:2.807
1765:link
1677:ISSN
1558:ISSN
1446:ISBN
1290:ISBN
1253:2023
1157:ISBN
1087:>
979:The
319:code
1667:doi
1634:doi
1585:doi
1581:109
1548:doi
1438:doi
1415:doi
1376:doi
1372:102
1282:doi
1224:HAL
1193:doi
1101:151
870:126
855:AES
845:In
670:in
442:is
155:log
85:100
1831::
1761:}}
1757:{{
1725:.
1698:.
1675:.
1663:42
1661:.
1657:.
1632:.
1620:18
1618:.
1614:.
1591:.
1579:.
1556:.
1546:.
1534:.
1530:.
1490:.
1444:.
1413:.
1399:.
1364:.
1341:.
1336:)"
1298:,
1288:,
1280:,
1244:.
1232:^
1222:.
1199:.
1189:52
1187:.
1183:.
1151:.
1136:^
1115:.
1097:10
924:34
917:10
837:.
19:A
1817:.
1802:.
1788:.
1782::
1767:)
1753:.
1735:.
1710:.
1683:.
1669::
1642:.
1636::
1599:.
1587::
1564:.
1550::
1542::
1536:2
1515:.
1509::
1494:.
1475:.
1469::
1454:.
1440::
1421:.
1417::
1401:8
1384:.
1378::
1284::
1274::
1255:.
1207:.
1195::
1165:.
1090:9
1084:n
1081:+
1078:m
1050:n
1030:m
1010:)
1007:n
1004:+
1001:m
998:(
995:O
866:2
806:2
792:2
787:2
767:=
764:2
728:4
725:=
722:h
702:H
682:h
658:)
655:)
652:)
649:2
645:/
641:h
638:(
632:2
629:(
623:2
620:(
614:2
594:H
570:G
550:n
530:)
525:2
521:n
517:(
514:O
494:G
474:H
454:H
426:H
406:G
373:n
353:n
333:n
293:)
284:n
280:(
277:O
253:)
244:n
240:(
237:O
213:)
208:3
204:n
200:(
197:O
164:)
161:n
152:n
149:(
146:O
116:.
94:)
81:2
76:n
70:(
Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.